GATA KAMSKY

CHESS GAMER

Volume 1: Awakening (1989–1996)

by
Gata Kamsky

Thinkers Publishing 2019

www.thinkerspublishing.com

Editor in Chief
Romain Edouard

Consulting Editor
Daniël Vanheirzeele

Proofreading
Ian Marks

Graphic Artist
Philippe Tonnard

Cover design
Mieke Mertens

Typesetting
i-Press ‹www.i-press.pl›

First edition 2019 by Thinkers Publishing

Gata Kamsky — Chess Gamer: Volume 1

Copyright © 2019 Gata Kamsky

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any
Editor in Chief
Romain Edouard

Consulting Editor
Daniël Vanheirzeele

Proofreading
Ian Marks

Graphic Artist
Philippe Tonnard

Cover design
Mieke Mertens

Typesetting
i-Press ‹www.i-press.pl›

First edition 2019 by Thinkers Publishing

Gata Kamsky — Chess Gamer: Volume 1

Copyright © 2019 Gata Kamsky

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system
or transmitted in any form or by any means, electronic, mechanical, photocopying, recording
or otherwise, without the prior written permission from the publisher.

ISBN 978-94-9251-028-0
D/2018/13730/10

All sales or enquiries should be directed to Thinkers Publishing, 9850 Landegem, Belgium.

e-mail: [email protected]
website: www.thinkerspublishing.com
 TABLE OF CONTENTS

Title page
Key to symbols
Introduction

Game 1. Gata Kamsky —Krunoslav Hulak

Game 2. Gata Kamsky — Mikhail Tal
Game 3. Gata Kamsky — Vasily Ivanchuk
Game 4. Anatoly Karpov — Gata Kamsky
Game 5. Gata Kamsky — Joël Lautier
Game 6. Gata Kamsky — Garry Kasparov
Game 7. Gata Kamsky — John Fedorowicz
Game 8. Alexander Beliavsky — Gata Kamsky
Game 9. Gata Kamsky — Alexei Shirov
Game 10. Gata Kamsky — Garry Kasparov
Game 11. Gata Kamsky — Vladimir Kramnik
Game 12. Gata Kamsky — Johann Hjartarson
Game 13. Gata Kamsky — Vladimir Kramnik
Game 14. Vladimir Kramnik — Gata Kamsky
Game 15. Gata Kamsky — Viswanathan Anand
Game 16. Viswanathan Anand — Gata Kamsky
Game 17. Gata Kamsky — Valery Salov
Game 18. Gata Kamsky — Anatoly Karpov
Game 19. Gata Kamsky — Curt Hansen
Game 20. Gata Kamsky — Viswanathan Anand
Game 21. Gata Kamsky — Anatoly Karpov
Game 22. Gata Kamsky — Anatoly Karpov

1
 KEY TO SYMBOLS

! a good move
? a weak move
!! an excellent move
?? a blunder
!? an interesting move
?! a dubious move
 only move
= equality
 unclear position
 with compensation for the sacrificed material
 White stands slightly better
 Black stands slightly better
 White has a serious advantage
 Black has a serious advantage
+- White has a decisive advantage
-+ Black has a decisive advantage
 with an attack
 with initiative
 with counterplay
 with the idea of
 better is
 worse is
N novelty
+ check
# mate

2
This book is dedicated to my loving wife,
Vera Nebolsina,
whose support and encouragement
were among the key factors
that allowed me to complete the work.

G. K.

3
 INTRODUCTION

Dear Reader,

This work was originally envisioned as a single-volume collection of my most memorable

games, annotated by me, à la Bobby Fischer's My 60 Memorable Games. However the
more I delved into the past, the more things started to rise up from the recesses of my
memory, which, along with deeper analysis and more detailed introductions to the games,
made it more like an anthol-ogy of chess stories. Thankfully, my gracious editors decided to
split the work into two volumes.

What you are going to find in this first volume is a selection of my most memo-rable battles on
the chessboard during the first of two different periods. It cov-ers the time from my arrival in
the USA as an up-and-coming young talent in early 1989, acquiring the freedom to
play in any open tournament in the world and quickly gaining precious experience to
grow into a challenger for the World Championship in 1996. This period ended with my early
retirement from the game to pursue other goals in life. The second volume will cover my
experiences after returning to the professional chess arena in 2004, having completed
my college and higher education, with the goal of fighting for the world title again. I came
very close to achieving this; winning the World Cup in 2007 placed me amongst the top
three players in the world. Such a success-ful return from retirement had never been
attempted or achieved before in chess history.

I have included just a couple of games after 2013 since I felt that they were important to
show the reader how, despite the ever-changing nature of chess, some aspects of it remain
eternal.

One of the beliefs I have held for most of my life is that in chess there is only one goal
worth pursuing this profession for - that of becoming World Cham-pion. Everyone knows
and remembers who the World Champions were; very few remember those who did not
succeed. Having said that, I also believe that chess represents a great way for people of
different nations to communicate and to settle disputes, as well as to forge friendships.
Some of my lifelong friendships, which I value greatly, came about thanks to chess.

4
Chess is known to help a person develop who he or she is, to discover his or her gift and to
present an opportunity to show that gift in a peaceful way. The other aspects of chess are well
known - art, science, sport, war strategy, his-tory, psychology (very much) and others - all of
which make it a unique and great ‘game’.

In this book, I have attempted to share my perception of chess and the as-pects
mentioned above through my annotated games and style, which was influenced mostly
by the chess books and biographies of great players that I read when I was a child.
Steinitz, Lasker, Capablanca, Alekhine, Botvinnik, Petrosian and others were my teachers to
a great extent.

I must apologize in advance if some of my comments will be found offensive, but because
chess has been a great part of my life, I do not try to be fair or objective, as each game,
or more correctly, story, is tied to my emotions in a strong bond that reminds me of what
was happening at that particular mo-ment.

This book is intended for fans of the royal game with a wide range of skill. There are
deep strategic battles, tactical storms, devious traps, life-affecting errors in time trouble
and everything else!

My sincerest hope is that each reader will find something memorable or useful that will echo in his
or her mind and, perhaps, provide something that will improve his or her level of chess
understanding.

Above all, the short game-stories in this book are meant to inspire the reader to search for
and discover the secrets of the game and its beauty.

I must say a big thanks to the great publishing and editorial team at Thinkers Publishing for
their support and patience in working with me, in particular Messrs Romain Edouard and
Daniël Vanheirzeele, as well as the many others involved in this project, without whom this
effort would not have been pos-sible.

Gata Kamsky
five-time US Chess Champion
October 2018

5
 GAME 1
Gata Kamsky
Krunoslav Hulak
New York 1989
This was my first round game in the event that was a turning point in my life. I had just arrived in the US, at the
invitation of the American Chess Foundation, accompanied by my coach and father. Little did I know that I would
never return to the country known as the USSR. My opponent had just come fresh from winning another GMA event
(the Grand Master Association, recently created by Garry Kasparov in his quest to challenge FIDE’s supremacy over
the chess world). There is a lot that could be said about the circumstances surrounding the tournament. Regarding
the chess itself, I can say, with a great deal of amusement (now!), that the ‘top level’ Commodore computer that my
father had bought for an enormous amount of money in Russia from a certain Mr. Lyapunov (who claimed to be a
helper of GM Beliavsky) did not work in the USA because of the difference in the electrical current. When my
father and I went to a local electrical store to ask how to get the computer to work, the guy looked at us like we were
aliens and told us that the computer belonged in a museum!! He tried the power adaptor, but it still didn’t work, so
for the whole tournament I was on my own, with lots and lots of heavy chess theory books and notes full of my
games and analysis. This probably seems incredible to any young person who now has access to all the information
thanks to his or her laptop, yet such were the days in the past.

1.d4 Nf6 2.Nf3 e6 3.Bg5

The Torre Attack. It and the London System, 3.Bf4, were trusty weapons which were to serve me well for many
years, even against the strongest opposition. It is named after the Mexican grandmaster who was active in the 1920s,
Carlos Torre.

3...h6 4.Bh4 b6 5.Nbd2 Bb7 6.e3 Be7

Black doesn’t commit his central pawns to any configuration just yet, waiting for me to declare my intentions first.
Some of the typical ideas for Black include the ...g7–g5 pawn push, followed by ...Nf6–h5, chasing my dark-squared
bishop, since its exchange, in general, is considered good for Black.

7.h3

Preparing an escape square on h2 for my dark-squared bishop.

7
7...c5 8.c3 Nc6 9.Bd3 cxd4

The first critical position of the game. Black decides to simplify the structure in the center, as taking on d4 with the
c3-pawn would allow ...Nc6–b4, attacking my d3-bishop. After that, a logical follow-up by ...Nf6–e4 or ...Bb7–a6
would make sense, depending on the retreat path of my bishop. As a result, Black would be able to simplify the
position to almost certain equality.
Therefore my next move is almost mandatory.

10.exd4

10.cxd4 Nb4 11.Be2 Ne4 (11...Rc8 seems to be a loss of time as White simply chases the b4-knight and gains more
control over the center. 12.0-0 0-0 13.a3 Nbd5 14.Bg3²) 12.Bxe7 Qxe7 13.a3 Nxd2 14.Qxd2 Nd5 15.0-0 0-0 16.Bd3
Nf6 17.Qe2 d6 18.e4 Rac8=

10...Nh5

It appears that White wasted a tempo with h3. The resulting position seems to be dynamic: both sides have mobile
pawns in the center, and the position is semi-closed, allowing the possibility of opposite-side castling.
During the game I thought it was really smart of Black to exchange my dark-squared bishop, but many years later,
looking at this position again, I think that my bishop on h4 is very awkward and hinders the pawn expansion. Of
course during the game I suspected that my opponent did not expect me to even consider castling long at this point.
However one of the advantages of youth is in bravely storming positions with such optimism that even the wisdom
of age sometimes has no choice but to concur. At least, sometimes...
a) 10...0-0 is a logical alternative.

8
11.Bg3! leaves White with a small advantage.
[11.g4?! A bit too early. White’s forces are still unprepared for a general assault on the kingside. 11...Qc7! Black
wants to play ...Bd5, while Bd3 will always be answered by ...Bd6. (11...e5 12.Bxf6 Bxf6 13.d5 Ne7 14.d6!

leads to unpleasant positions for Black: 14...Ng6 — or 14...Nc6 15.Ne4 Re8 16.Rg1 Bg5 17.Nfxg5 hxg5 18.Qe2² —
15.Be4 Bxe4 16.Nxe4 b5 17.Qb3!? a5 18.0-0-0 Rb8 19.h4 with a huge attack, although Black may hold if he can
defend like a computer:

9
19...a4 20.Qc2 Bxh4 21.Rdg1 b4 22.Nxh4 Nxh4 23.g5 Ng6™ 24.gxh6 bxc3 25.bxc3 Qb6 26.hxg7 Kxg7 27.Ng3

27...Rg8! 28.Nf5+ Kf8 29.Rh7 a3 30.Nh6 Rg7 31.Rxg6 Qb2+ 32.Qxb2 axb2+ 33.Kb1 Rxg6 34.Rh8+ Kg7 35.Rxb8
Rxh6 36.a4 Rxd6 37.a5 Kf6 38.Kxb2 Ke7 39.Rb5 f6 40.c4 Kd8 41.Kc3 Rd1=) 12.Bxf6 (12.g5 hxg5 13.Bxg5
Capturing with the knight would have allowed ...Nxd4, attacking the h1-rook . 13...Nd5 14.Rg1 f5!?∞) 12...Bxf6
13.g5 (13.h4 Bd8 14.g5 h5∞) 13...hxg5 14.Ne4 Qf4 and White’s attack doesn’t succeed:

10
15.Rg1 Nb4 16.Nxf6+ gxf6 17.Be2 Bxf3 18.Bxf3 Nc2+ 19.Qxc2 Qxf3 20.h4 Kg7 21.hxg5 fxg5 22.Rxg5+ Kf6
23.Rg1 Rg8µ]
11...d5 (11...d6 is also possible, but a bit passive, allowing White his space advantage) 12.Qe2 Bd6 13.Bxd6 Qxd6
14.Ne5 Qc7 15.f4 Ne8

16.g4! with the initiative. White will probably play Rg1 next, then castle long. (16.Qe3 Nd6 17.0-0-0 b5 18.Kb1 Nc4
gives counterplay)
b) 10...Qc7!? is extremely interesting as well, with the idea of 11.Qe2 (11.Bg3 Bd6 Exploiting the fact that the pawn
is no longer on h2 and can’t support the g3-bishop. 12.Bxd6 Qxd6 13.Qe2 Ne7 14.Ne5 Qc7 15.0-0-0 Ned5ƒ)
11...Nd5∞

11.Bxe7 Nxe7

11
A good move, as Black opens up his b7-bishop and prepares to transfer his e7-knight via d5 to a more useful square
on f4.

12.g3

Prophylaxis. White has to take the f4-square under control.

12...Qc7 13.Qe2 0-0

Black finally thinks it is a good time to castle, since if White castles short, Black’s b7-bishop will become a monster.
In hindsight, perhaps it was better for Black to bring his h5-knight back into the center and wait for the white king to
make a decision on castling.

14.0-0-0

Plans have been declared — it’s war! Pawns will be racing against opposing kings and defenses will be tested. I had
12
a pretty good idea that if I managed to hide my king in the a1-corner and put the rook on c1, then Black’s ...b5–b4
pawn break could be sidestepped by my c3–c4 pawn push and White should be safe, something which can’t be said
about Black’s king. At the same time, White’s d3-bishop is going to be a crucial component in both defense and
offense.

14...Nf6 15.Rhg1 Nfd5

This move retains the option of playing ...f5, however in positions like this, every tempo is important. An immediate
...b5 was preferable, as taking this pawn would open the b-file for the black heavy pieces and White would be tied to
pure defense.
15...b5!

16.Kb1
[16.Bxb5 Rab8 17.Kb1 Bxf3 18.Nxf3 Qa5 forces White’s b5-bishop to move, since an attempt to protect it with c3–
c4 even loses to Black’s ...a7–a6 resource. 19.Bc4 Qxc3 20.Rc1 Qb4

13
21.Bd3 (21.g4 A reckless ‘attacking’ move like this cannot succeed as most of White’s pieces are tied up defending
the king. 21...Ng6 22.g5 hxg5 23.Rxg5 Nf4 24.Qc2 Rfc8 and Black’s initiative on the queenside threatens to become
decisive) 21...Rfc8 Perhaps a bit timid, but exchanging a pair of rooks will forever eliminate White’s ambitions on
the kingside and make the game more a question of whether White can hold on to draw. 22.Rxc8+ Rxc8 23.Rc1 Rb8
24.Qd2 Qb7 25.Ne5 a5 26.Ka1 Nf5!∞ This position seems roughly balanced, thanks to the reduced material,
however there is plenty of play left. Black maintains a slight advantage thanks to White’s weak isolated d4-pawn.]
16...Rab8 17.Ne5

Black can’t push ...b4 that easily as White will just answer c3–c4. 17...d6
[17...a6?! This move serves a double function, firstly to protect the b5pawn and secondly to secure an outpost on the
d5-square for the black knights. The move’s main drawback is the waste of valuable time. 18.f4 Nfd5 19.Rc1
(19.Rgf1?! f6 20.Ng4 Nxc3+ 21.bxc3 Nd5 22.Rc1 Nxc3+ 23.Kb2 Rbc8 24.Qf2 Qa5 25.Bb1 Bd5 26.Ne3 Qb4+
27.Ka1 Na4 28.Nb3 Bxb3 29.axb3 Rxc1 30.Rxc1 Qa3+ 31.Qa2 Qxc1 with a large material advantage.) 19...d6
20.Ng4 Nf5 21.Qf2 a5! with a complex battle ahead.]
18.Ng4 Nfd5 19.Rc1 Bc6 20.f4 a5

14
21.Ka1!∞ This cautious move, sending White’s king further into the corner away from the X-ray of the black b8-
rook, is probably the most solid continuation. The game is sharp and unpredictable, though it does seem a bit easier
to play with White, as Black’s h6-pawn gives White a potential target for the attack on the kingside.
a) 21.Nf3 Qb7 22.Rgf1 Nf5ƒ 23.Qf2 b4 24.c4

24...Nc3+! 25.Ka1 Be4! 26.Qd2 Bxf3 27.bxc3 Be4!µ The simplest. White is overextended and exposed on both
flanks with numerous pawn weaknesses as a result. Black has a very clear advantage.
b) 21.Ne3?! b4! With this thematic move Black blows open White’s queenside fortifications and forces White to
switch to defense. 22.c4 Nc3+ 23.bxc3 bxc3+ 24.Ka1 cxd2 25.Qxd2 Bd7!

26.g4?! (26.Qc3?! Rfc8 27.g4 d5µ) 26...d5!µ The point: Black threatens to capture White’s key f4-pawn and, at the
same time, threatens to destroy White’s connected central pawns.

16.Kb1 Rab8 17.Rc1

15
White sets up his defensive perimeter on the queenside before starting offensive action on the kingside.

17...Bc6?!

This move seems to be Black’s first inaccuracy.

Black plans to use his b-pawn to ram White’s defenses on the queenside, however to support this plan he is forced to
place his light-squared bishop in an awkward position. In addition, after a potential Nf3–e5, White will have an extra
option of exchanging the black c6-bishop, creating lasting pressure on the light squares across the whole board.
a) 17...b5 is a bit too late now, as White has covered his vulnerable c3-pawn with Rc1, and while Black will
definitely have a measure of compensation for the pawn, it’s far from certain that this extra pawn will not decide the
game in the later stages. 18.Bxb5!? (18.Ne5?!) 18...Nf6 19.Bd3 a5 20.Ne5 d6 21.Nec4 a4 22.f4²
b) 17...d6 Perhaps something simple like this move, denying White’s f3-knight an obvious developmental square,
was best, as a knight on e5 would be ideal in helping White launch his kingside attack. 18.Ka1! (18.Nh4?! seems
terrible: 18...b5 19.f4 Bc6 20.Rgf1 b4 21.c4 Nc3+ 22.bxc3 bxc3+ 23.Nb3 a5 24.Rxc3 a4 25.f5 exf5 26.Nxf5 Nxf5
27.Bxf5 axb3 28.axb3 Qa5 and Black has a strong initiative due to the open white king) 18...Bc6

16
19.Nh4 (19.Bb1 b5 20.Qd3 g6 21.h4 Nf6 22.g4 Ned5 23.g5 Nf4 24.Qf1 hxg5 25.hxg5 N6h5 26.Rh1 Kg7 27.Qe1
Rh8!= Black succeeds in establishing an almost unbreakable fortress on the kingside, leading to a temporary stasis in
the position.) 19...b5 20.f4 a5 21.Rgf1 White readies himself for the attack, which will be difficult to defend against.

For example 21...Rfe8 22.g4 Qd7 23.Qf2 (23.g5 Nf5 24.Qf2 Nde3 25.Rg1 hxg5 26.Rxg5 f6 27.Rgg1‚ ) 23...f6
24.Rce1 (24.g5 hxg5 25.fxg5 f5 26.Ng2 a4 27.a3 Qb7 28.Rg1!‚ and White is ready to push his h-pawn as a battering
ram against the black kingside fortifications.) 24...Rbd8 25.Ne4 Rf8 26.Ng3‚ It is obvious that Black has fully
switched to defense on the kingside.

18.Ne5 b5 19.f4

17
19...a5

a) 19...b4 This idea no longer works because White has access to the c5-square for his d2-knight. 20.c4 Nc3+
21.bxc3 bxc3+ 22.Nb3 a5 23.Ka1

23...a4 (23...Ba8 24.Rxc3 d6 25.Nf3 a4 26.Nbd2) 24.Nc5! Rb2 (24...d6 25.Na6+– This double attack decides the
game.) 25.Qe1 d6 26.Nxc6 Nxc6 27.Qxc3+–
b) 19...d6 An attempt to get rid of White’s best piece makes sense, but during the game my opponent probably
didn’t like the fact that I would suddenly transpose into an endgame where bishop and knight and a more
harmonious pawn structure would give me a certain small but lasting advantage. For example 20.Nxc6 Qxc6

18
21.Qe4 (21.Rgf1!? keeping the queens on is an alternative. 21...Nf5 22.Qf3 Nfe3 23.Rf2 Nc4 24.Ka1 Nxd2 25.Rxd2
Rfe8 26.g4 Ne7 27.Qf2 d5 28.Rdc2 Qd6 29.h4 Rec8 30.a3 a5 31.Kb1! White prepares for Black’s eventual ...b5–b4
break. 31...b4 32.cxb4 axb4 33.Rxc8+ Nxc8 34.a4 b3!= with a balanced game) 21...Nf6 22.Qxc6 Nxc6 23.a3 a5
24.g4 Nd5 25.Rgf1 Rfc8 26.Ne4 Rd8

27.f5 This is a thematic attempt by White to break Black’s pawn formation in the center. (27.Rf2!? f5 28.Ng3 g6∞
leads to a complex game where Black’s slightly weakened pawn formation is compensated for by an incredibly
strong d5-knight.) 27...e5 28.dxe5 dxe5 (28...Nxe5 29.Be2 Ne3 30.Rf4 d5 31.Nd2 N3c4 32.Nf3 f6∞) 29.Bxb5 Rxb5
30.c4 Rb6 31.cxd5 Rxd5

19
32.Rc5 (32.f6 Rd4 33.Rfe1 gxf6 34.Nxf6+ Kg7 35.Ne4 f6=; 32.Rfd1 Nd4 33.Rc8+ Kh7 34.Ka2 Rdb5 35.Rd2 Rb8!
36.Rc7 R5b7 37.Rxb7 Rxb7 38.Rd3 h5! with an approximately equal game, though White’s queenside pawn
majority has strong potential to turn the tables in his favor) 32...Rxc5 (32...Rd4 33.Nc3!²) 33.Nxc5 Nd4 34.Rc1 f6
35.Ne4 Rb8 36.Ka2 h5∞

20.Ka1

White’s king is finally secure and safeguarded against possible black ideas with ...b4 and ...Nc3. At this point, I was
optimistic about winning, as White has a clear plan for his attack on the kingside.

20...d6

At long last Black decides to give up hope of an attack and focuses on the center and defense.

21.Nxc6 Qxc6 22.Rgf1

20
22.Qf3 This is an interesting alternative suggested by the computer, with the idea of keeping the rook on g1, since
it’s going to be needed there for the g4–g5 push, opening the g-file. However, it seems that Black has sufficient
defensive resources after the following sequence: 22...b4 23.c4 Nf6 24.Qe2 d5

25.g4! A superior move, since the alternative below suddenly gives Black a new target, White’s central d4-pawn.
[25.c5 b3 26.a3 Qa4 27.Nf3 Nc6 28.Rcd1

28...Kh8 (28...Nb4 29.g4ƒ) 29.g4! (29.Bb1?! Na7! Forcing White’s b1-bishop to return to the d3-square to prevent
Black’s knight from hitting White’s a3-pawn with the coming ...Na7-b5.) 29...Nxd4 30.Nxd4 Qxd4 31.g5 Ne4
32.gxh6 g6 33.Bxe4 Qxe4 34.Qd2 Kh7 35.h4∞ White has good attacking potential.]
25...dxc4 26.Nxc4 Nfd5

21
White clearly has a more dangerous initiative, however it is not easy to choose the right continuation, as the
following options show. 27.Qe4 (27.Qf2? This slow move even allows Black to recapture the initiative. 27...a4 28.f5
a3 29.b3 Qc7!³ Suddenly Black has managed to expose a number of weak dark squares in the white camp. White
must be extremely careful not to end up with a much worse game.) 27...Nf6 28.Qe5 Seems best. 28...Qd5 29.g5
hxg5 30.Rxg5 Qxe5 31.fxe5 Nfd5 32.Nxa5 g6∞ And in this exciting endgame, Black seems to at least hold his own.

22...Qd7

The game reaches the point where both sides have completed their preparations and open conf lict must start. The
complexity of the position arises from the fact that there are numerous ways of placing the minor pieces for both
sides. With the game move, Black wants to play ...b4, and therefore removes his queen from the X-ray of White’s
c1-rook.
22...Nf5

22
23.Qe4 Probably strongest, forcing Black to further weaken his kingside pawn structure, thus creating an additional
target for White.
[23.Qf2 Nfe3 24.Rfe1 Nc4 25.Ne4 f5 26.Nd2 Rf6 27.Bf1

Black’s defenses hold for now, but he has to remain on constant alert, as any careless move could spell trouble, for
example 27...a4!? (27...Nxd2?! 28.Qxd2 a4 29.Bg2 Qd7 30.Bxd5 exd5 31.Qg2 Qf7 32.a3² fixes Black with an
inferior structure, which could become a decisive factor in the pawn endgame. 27...Ndb6? 28.g4!± results in an
immediate opening of the position which favors White, due to his very strong bishop.) 28.a3 Nc7 29.Bg2 Qb6²]
23...g6 (23...Nde7? 24.g4+–) 24.Rg1 Kh8 25.g4 Nfe7 26.h4‚

23.Ne4

White centralizes the knight and provokes Black into active defensive measures like ...f5, which could later be used
for a ...g4-break.
23.g4 is also possible immediately, but after ...f5 White has to decide how to defend his f4-pawn. 23...f5 24.Rce1
23
Ng6 25.Qh2 Rbe8 26.Bb1 Nge7 27.g5 h5∞ Black successfully keeps the kingside locked.

23...Rb6

23...Nc6 24.f5 exf5

25.g4!! Suddenly White’s attack is almost decisive. 25...f4 26.g5 hxg5 27.Qh5 g6 (27...Qf5 28.Rg1 Qg6 29.Qf3+–)
28.Qh6+– Black’s kingside defenses fall.

24.Bb1!

This move creates the threat of setting up a Bb1/Qd3 mating battery, therefore Black is finally forced to play ...f7–f5,
which opens another front on the kingside for the white forces.

24...f5 25.Nd2 Nf6

Played with the idea of creating an outpost on the e4-square by pushing ...d6–d5 next, followed by planting the f6-
24
knight on the newly created e4-square. This would close off White’s b1-bishop as well, but such a plan gives White
more possibilities than he is deprived of. Still, White’s position is preferable because of his attacking potential.
25...Rf6 was another alternative, attempting to hold the knight on d5, putting pressure on White’s f4-pawn, but as
can be seen from the following lines, passive defense doesn’t guarantee Black a quiet life.
a) 25...Rf6!? 26.g4 Ng6 27.Qh2

The following lines demonstrate the dangers for Black: 27...fxg4 (27...Rb8 28.Rce1 Kh8 29.Qg3 Nge7 30.Re2²)
28.hxg4 Ngxf4 29.g5 hxg5 30.Qh7+ Kf7 31.Ne4 Rf5 32.Rg1 Qe7 33.Rce1 Rb8 34.Ng3 Rf6 35.Nh5 Nxh5 36.Qxh5+
g6 37.Qxg5 Rg8 38.a4 bxa4 39.c4 Nf4 (39...Nb4 40.Ref1+–) 40.Rg4 Nh5 41.c5 a3 42.bxa3 Qd7 43.Qe3 Qc6
44.Be4±.
b) 25...Rbb8 26.Rce1 Nc7 27.g4 Rbe8 28.g5

White is ready for this push. 28...hxg5 29.fxg5 g6 30.h4 Kg7 31.h5 gxh5 32.Qxh5 Rh8 33.Qf3 Qc6 34.Qf2 Rh3
35.Be4 Qd7 36.Bg2 Rh2 37.Ne4 Ned5 38.g6 fxe4 39.Qg3 Rxg2 40.Qxg2 Nf6 41.Rxf6 Kxf6 42.Qf2+ Kg7 43.Rh1
Kg8 44.Rh7 Re7

25
45.Rh8+!!+– A pretty f inish to White’s attack.

26.g4‚ 26...d5 27.Nf3 Ne4 28.Ne5

The pawn structure has changed and now seems to be a good time to stand back and take a look at the position.
Black has achieved his objective of planting his knight on the e4-square, while White has transferred his knight to an
improved location, the e5-square, made possible by Black’s ...d6–d5 push. The overall transformation seems to favor
White, as he now also has a clear plan of doubling on the g-file. Meanwhile any potential ...b4 break can be met with
c4, when Black can’t really recapture on c4 with his d5-pawn as it would weaken his central e4knight. Black
probably has to try to eliminate White’s strong central e5-knight with ...Ne7–c6 before White breaks through on the
kingside.

28...Qe8 29.Rg1 b4?

A mistake after which Black’s position becomes critical. As mentioned before, Black’s only chance was to eliminate
White’s most active attacking piece, the e5-knight.
26
29...Nc6 30.gxf5 Nxe5 ( 30...exf5 31.Rg2 Kh7 32.Rcg1 Rb7 33.h4 Rf6 34.h5

Black manages to establish the last line of defense. White’s advantage is clear, and it’s only a matter of time before
he breaks through and starts cleaning up Black’s weak pawns in the center and on the queenside.
34...Kh8 35.Rg6 Nxe5 36.dxe5 Rxg6 37.Rxg6 Qc8 38.Qd3 Qc5 39.Ra6 Ra7 40.Rxa7 Qxa7 41.Qxd5 Qg1 42.Qd8+
Kh7 43.Qd7+–) 31.dxe5 exf5 32.Qd3 Qe6 33.h4

33...b4 (33...Rd8 34.h5 Qf7 35.Qd4 Re6 36.Rg2 Qc7 37.Rcg1 Re7 38.Rg6 Kh8 39.Bd3±) 34.c4 dxc4 35.Rxc4±
White’s advantage is almost decisive thanks to his strong light-squared bishop and control of the open files.

30.c4 b3 31.a3+–

27
It is now clear that Black’s attack on the queenside has failed, while White’s initiative is getting stronger with every
move. The advantage is decisive.

31...dxc4 32.Nxc4

The computer demonstrates a quicker path to a win, but at the time of the game I was only fourteen, and since this
was my first game in a big international event against a strong grandmaster, I did not want to take any unnecessary
risks.
32.gxf5! exf5 33.Qxc4+ Kh7 34.Qc7 Rbf6 35.Qa7 Nd6 36.Rc7 Re6 37.Qc5 Ref6 38.d5 Ng6 39.Nd7 Qd8 40.Nxf8+
Qxf8 41.Qxa5+– with a winning material advantage.

32...Rc6 33.gxf5 exf5 34.Nxa5 Rxc1 35.Rxc1

White eliminates Black’s queenside pawns, while maintaining the pressure on the kingside. I believed the game was
easily won and started to play carelessly.

28
35...Nd5 36.Qf3 Qb8 37.Bxe4?!

37.Rc5! Nxf4 38.Re5! Ng6 39.Rxf5 Nd2 40.Rxf8+ Qxf8 41.Qd3 Nxb1 42.Kxb1 Nf4 43.Qxb3+ Kh8 44.Qf3 Qf5+
45.Ka1+– would have made converting White’s material advantage easier.

37...fxe4 38.Qxe4 Nxf4 39.Nc6 Qb5 40.d5 Qd3 41.Ne7+ Kh8

42.Qxd3?!

Not yet the big mistake, however the idea that follows is flawed as it allows Black a miracle escape. Quite nervous at
this point, I wanted to transpose into the endgame phase, where I believed in my technique. The correct, and
winning, move was 42.Ng6, though it required deep calculation.
42.Ng6+! Kg8 43.Qxd3 Nxd3 44.Nxf8 Nxc1 45.Ne6 Kf7 46.a4 Nd3 47.a5 Nb4 48.Nc7! g5 49.a6 Nxa6 50.Nxa6 h5
51.Nc5+–

42...Nxd3

29
43.Rd1??

It was not too late to return to the winning line after 43.Ng6. Now, however, quite suddenly, things are not clear at
all, due to White’s weak back rank and Black’s pawn advantage on the kingside.

43...Rf6?

Black returns the favor by losing a crucial tempo. Much stronger was 43...Kh7!!

which the computer evaluates as equal. I feel I must demonstrate the following lines to show how incredibly
complex this endgame is, as if the middlegame with its opposite wing attacks wasn’t enough. 44.Nc6
a) 44.d6 Nf2! (44...Ne5 45.Nd5 Nd7 46.Kb1± If White manages to get his king out of the box, he’s very close to
winning as he has connected passed pawns on the queenside.) 45.Rf1 Rf6! 46.Nc8 (46.d7? Rd6µ) 46...Rf4! 47.Nb6
Rd4 48.d7 Nxh3 49.a4 g5!

30
and in the resulting pawn race, Black is just in time. 50.Kb1 h5! Black’s pawns must travel together. (50...g4??
51.Kc1 g3 52.Rf7+!! Kg6 53.Rf3+–) 51.Rc1! (51.Kc1 Nf4 Keeping White’s king boxed in. 51.a5 g4 52.a6 g3 53.a7
g2 54.Rc1 g1=Q 55.a8=Q Qg6+ 56.Ka1 Qxb6–+, winning for Black.) 51...g4 52.Rc8 Rd1+ 53.Rc1 Rd4= Leading to
a repetition as the black pawns are extremely dangerous.
b) 44.h4 Nf2 45.Rg1 Rf3 It is vital for the black rook to get behind White’s d-pawn and keep White’s weak back
rank in check.

46.Kb1 Ne4 47.Nc6 Rd3 48.Nb4 Rh3 49.Kc1 Rxh4 50.Rd1 Nd6 51.Rd3 g5 52.Rxb3 g4 53.Rg3 Rh1+ 54.Kc2 h5
55.a4 Ra1 56.b3 Nf5 57.d6 Nxd6 58.Kb2 Re1 and again Black has managed to turn the game in his favor.
44...Nf2 (44...Nc5? 45.d6 Rf3 46.a4+–) 45.Rf1 Rf3!

31
46.Nb4 (46.a4 Rxh3 47.a5 Ne4 48.a6? Nd2!!µ A study-like maneuver that exposes the vulnerability of the white
king. 49.Rd1 Rh4µ) 46...Rg3!! An amazing computer move. Other moves are worse as either the white king or a
pawn starts to move forward.
a) 46...Rf4 47.Kb1 Rd4 48.Kc1 Nxh3 49.Rd1 Rc4+ (49...Rxd1+ 50.Kxd1 Nf4 51.d6 Kg6 52.a4 Ne6 53.a5 Nc5
54.a6+–) 50.Kd2 Nf2 51.Re1 h5 52.Ke3 Ng4+ 53.Kd3 Rf4 54.Kc3 Nf6 55.d6 Rf5 56.Rd1 Nd7 57.a4 h4 58.Nd5 h3
59.a5 h2 60.Nb6+–
b) 46...Rxh3 47.d6 Rh4 48.Nc6, winning for White.
47.Kb1 (47.a4 Rxh3 48.a5 Rh4 49.Nc6 Ne4! and Black is better: 50.Rd1 Nd2 51.Nb4 Rxb4 52.Rxd2 Ra4+ 53.Kb1
Rxa5 54.Rd3µ) 47...Ne4 48.Kc1 Rxh3 49.Rd1 Nd6 50.Nc6

50...h5 (50...Rh4=) 51.a4 h4 52.a5 Nc4! 53.d6 (53.a6? Rh2µ) 53...Rh2 54.Nd4 Nxa5 55.d7 Nb7 56.Rd3 h3 57.Rxb3
Nd8 58.Rg3 Kg8 59.b4 Rh1+ 60.Kd2 Kf7 61.Nf3 h2 62.Rh3

32
Finally the outcome of this complex endgame is a draw.

44.a4!

White has the advantage again, though Black has certain defensive resources.

44...g5?

Black makes another big mistake, after which the game cannot be saved. He had to stop the advance of White’s a4-
pawn.
44...Ne5 A very tricky move, the idea of which is to keep this knight close to the c4-square. 45.d6 (45.Kb1? Rf2
demonstrates the idea behind Black’s previous move, viz. to play ...Ne5–c4, attacking White’s b2-pawn, with strong
mating threats against the white king.) 45...Kh7!! Another incredible move. Black prepares to transfer his rook to the
fourth rank where it will be able to stop White’s a-pawn and exert potential pressure on White’s b2-pawn as well.

33
(45...Rf4?? 46.Ng6+! This def lection idea wins the game for White. 46...Nxg6 47.d7+–) 46.Nc8! (46.a5 Rf4 47.Kb1
Rf2 48.Nd5 Nc4 49.Nb4 Nxa5 50.d7 Nb7 51.Nc6 Kg6 52.d8=Q Nxd8 53.Nxd8 Rh2 and Black has very good
drawing chances.) 46...Nd7 Not a move Black is happy to play, but he seems to have drawing chances. (46...Nf7
47.Kb1 Rf4 48.a5 Ra4 49.Rd3 Rxa5 50.Rxb3 Rd5 51.Rb7 Nd8 52.Rd7±) 47.Kb1!

47...Rf8 (47...Rf5 48.Ne7 Ra5 49.Rd4 Rg5 50.Kc1+– and after the white king comes out of the box, White is
winning.) 48.Ne7 Nb6 49.Nc6! (49.a5? Nc4! with counterplay) 49...Nd7 (49...Nc4? 50.d7 Rf2 51.Ka1 Rxb2
52.Nb4!+–) 50.a5 Ra8 51.Rc1 Ra6 52.Rc3 Kg6 53.Kc1 Kf6 54.Kd2

34
54...g5 ( 54...Ke6 55.Nd8+ Kxd6 56.Nb7+ Ke5 57.Re3+ Kf4 58.Rxb3± and White keeps his winning chances alive)
55.Ke3 Ke6 56.Nd4+ Kxd6 57.Nxb3 Ra8 58.Nd2 Kd5 59.Nc4 Rb8 with very good drawing chances.

45.a5 h5 46.Nc6+–

Now Black cannot save the game anymore.

46...Nc5 47.d6 Rf4

47...Rf7 48.a6 Nxa6 49.d7+–

48.Kb1 1–0

The pawns cannot be stopped. What made this game especially memorable was that GM Gelfand, who also played
in the tournament, came up to me and congratulated me on my victory in words that I will not mention here.

35
 GAME 2
Gata Kamsky
Mikhail Tal
New York 1990
This game was selected for its pure nostalgic value. I consider myself very lucky to have been born at the romantic
time in chess history before the advent of the computer era. Many of the all-time greats and ex- World Champions
are gone now, but at their peak, their games captivated the chess world. My opponent was one such player. Mix a
brilliant sacrificial style at the board with an exotic lifestyle away from it and you have an enduring legendary World
Champion whose games inspired and taught countless developing chess kids. I was, of course, one of those kids who
admired Tal’s games. Personally, I met Mr. Tal only a few times in my life, the earliest I can remember being his
brief visit to the Pioneers’ Palace in St. Petersburg (Leningrad at that time) when I was just becoming a first category
player. It was rumored that Mr. Tal was staying at a certain location in the city and people were welcome to come
and play some blitz games with him. I remember visiting the place and watching him play several games, my
shyness defeating my desire to play blitz. Another factor was the knowledge, painstakingly imparted by trainers to
the kids, that blitz is bad for a young person’s chess development!! Now I was meeting Mr. Tal for the second time,
in the USA at this New York Open. The New York Opens were very special events, organized by Mr. Cuchi for
many years, and later discontinued, unfortunately, but during the time they were held, many elite grandmasters flew
to the Big Apple to try their luck in one of the biggest chess opens in the world. Fate managed to pair me with one of
my childhood idols and I desperately wanted not to lose in terrible fashion!! It was really sad, though, to see how ill
my opponent looked. Even though the game certainly has its special moments, it is the memory of playing against
Mr. Tal that ensures its selection for the book.

1.d4 d6

A psychological offensive!

Everyone knew that I didn’t study and therefore didn’t know much theory, and mainly played queen’s pawn
openings as White. The move in the game delays the development of Black’s king’s knight to discourage the
development of White’s queen’s bishop!

2.Nf3 Nf6 3.c4

I wanted to play bold chess, worthy of my illustrious opponent, and had therefore actually prepared the line in the
game until move nine. Taken together with the subsequent plan, which was over-the-board inspiration, it was quite
36
in keeping with how I played QP openings, with long castling and a direct attack on the black king. My opponent
stayed true to his favorite opening, the King’s Indian Defense, and so the game was set to be unpredictable, but fun!

3...g6 4.Nc3 Bg7 5.e4 0-0 6.Be2 e5 7.Be3

This line versus the King’s Indian was quite popular in those days. Only many years later did Kramnik’s use of one
of his KID-killer weapons, the Bayonet Attack, in his match versus Kasparov finally cause the KID to lose a lot of
its adherents. The development of White’s c1-bishop to the e3-square is one of the key ideas in this opening,
however White has a huge selection of plans at his disposal and each line can lead to positions that may have a
completely different strategic twist. For example, one such line was employed by Karpov to win game seven of the
World Championship match versus Kasparov in New York in 1990. This was actually the only game I attended, as
my benefactor and sponsor, Mr. James Cayne of the Bear & Sterns stock exchange company, was present, and I felt
obliged to provide him with my personal expert commentary, along with a prediction of Karpov’s victory in that
game, even though US fans were rooting for their hero, Mr. Kasparov. Back to the present game. The idea behind
developing White’s queen’s bishop to e3 before castling is to lure Black into a mistake and play the ‘main’ standard
move 7...Nc6, after which White doesn’t castle, but plays 8.d5 Ne7 9.Nd2, when he has developed his bishop to an
ideal post and regrouped in a way that is impossible in the old main line with 7.0-0. Of course, I couldn’t hope that
such a beginner’s trick would work against my opponent.

7...exd4

My opponent indicates his intention to deviate from the standard closed center, where Black goes for a kingside
attack against White’s monarch. Now the focus of the game shifts to the center. Perhaps one of the factors for such a
game plan was my opponent’s desire to use his superior experience in playing these types of positions. The main
idea for Black now is to go for a timely ...c7– c6 and ...d6–d5 break, and if that doesn’t work, dance around White’s
rigid central pawn formation, poking with ...a7–a5 and ...h7–h5, with a complicated strategic battle. However I had
something different in mind when preparing this line.

8.Nxd4 Re8 9.Qc2!?

37
Now, in the computer age, I have found quite a few games predating this move, however at the time, no such
databases containing almost the entire history of chess were known, so the surprise effect was quite valuable. Even
now there are not that many games in the database, especially with the long castling plan I employed here. I
remember that my opponent went into a long think at this point.

9...Qe7

The main move these days is 9...c6, however let’s leave delving into the theoretical dispute to the specialists, and
look at the position as practical players. My preparation contained the basic idea of castling long and pushing pawns,
without analyzing much. In the Sicilian Dragon, such an approach is considered to be the norm, however in this
position, the presence of a white pawn on c4 is a major structural difference. It’s not easy to say whether it’s an asset
to White, preventing Black’s queenside pawn push, or a target.

10.f3 c6

A standard move for these positions. Black prepares his ...d6–d5 break. There was another, more aggressive,
approach, the immediate 10...c5, which would have led to completely different types of structures.
10...c5!?

38
Black exploits the fact that the usual escape squares for White’s d4-knight are all taken, and it has to retreat to the
awkward b3-square. At the same time, Black leaves himself with a permanently weak d6pawn (which will be
difficult to attack though), in compensation for which he gains a tempo for development and potential access to the
d4-square. In addition, since White’s b3-knight prevents White from supporting the c4-pawn with b2–b3, White will
have to spend more tempi relocating his b3-knight to a better pasture, thus giving Black valuable time to create
dynamic piece counterplay. 11.Nb3 Nc6 12.0-0 Be6 13.Rad1

13...Ne5!? (13...Rad8 This quiet developing move leads the game into a more strategic phase, where both sides have
chances to outplay their opponent.) 14.Nb5 (14.Nd2 Nc6! This knight takes aim at the d4-square as soon as the
white guardian of the square leaves. If White refuses the repetition, then he must be ready to part with his pair of
bishops. 15.Qc1 Nd4 16.Rf2 Nxe2+ 17.Nxe2 a6!

39
This move starts the standard queenside pawn advance for Black, with the idea that such action will open some lines
and diagonals for Black’s rooks and bishops. 18.b3 b5 19.Nf4 Nd7 20.Nf1 Ne5 21.Nxe6 fxe6!³ This seems to give
Black the superior position, since he has better control over the center, and the glaring hole in the white center, the
d4-square, is going to play a major role for Black’s minor pieces.) 14...Red8 15.f4

15...Nc6 (15...Neg4?! Such aggression is too early. The black pieces are not well placed to support it. 16.Bc1±
White manages to keep his strong center protected and his pieces coordinated, while Black will have problems with
his distant g4-knight.) 16.e5!
a) 16.Nxd6 Nb4! (16...Rxd6?! 17.Bxc5 Rxd1 18.Bxe7 Rxf1+ 19.Kxf1 Nxe7 20.Nc5 Bc8 21.e5 Nd7 22.Ne4!
This move allows the rerouting of White’s knight to the very powerful d6-square, where it will hinder Black greatly
while preparing the push of White’s queenside pawn majority. Objectively White has the superior chances here.

40
22...Nf8 23.Nd6 Nf5 24.Qe4 Nxd6 25.exd6± White’s active queen and two pawns are superior to Black’s two minor
pieces and rook, since Black has clear problems activating his queenside pieces, which are tied to the defense of the
b7-pawn.) 17.Qb1 Rxd6 18.Bxc5 Rxd1 19.Rxd1 Qc7!

This standard tactical trick, utilizing the open location of White’s king on g1, holds the black position together.
20.Bd6 (20.Bxb4 Qxf4∞ Black gets certain compensation due to his active pieces on the kingside and White’s weak
central e4-pawn.) 20...Qb6+ 21.Bc5 Qc7 Leading to a draw by perpetual, since neither side can really deviate
without drastically worsening his position.
b) 16.h3? This move is an example of how tricky such dynamic positions are to play for both colors. 16...Nb4
17.Qb1 Bf5–+
16...Bf5 17.Qc1 Ne8 18.exd6 Nxd6 19.Bxc5 Qxe2 20.Nxd6 Qxb2 21.Qxb2 Bxb2 22.Nxb7 Rxd1 23.Rxd1 Be6
24.Nd6 Rd8 25.Kf2 Bg7 26.Rc1 Bb2 27.Rc2 Bg7 28.Nb5

41
28...Rc8² We have reached a type of pawn-down endgame where Black usually survives, thanks to his superior pair
of bishops.

11.g4!?

White prevents Black’s central breakthrough and prepares simply to push his pawns on the kingside. Black must
come up with a counter or his king will be annihilated.

11...a6?!

Black can ill-afford to play such slow moves. Either 11...c5 or 11...Na6 would have started active counterplay.
a) 11...c5! 12.Nb3 (12.Nf5? This knight sacrifice idea is standard in similar KID structures, but it doesn’t work here
for tactical reasons. 12...gxf5 13.gxf5 Bxf5–+ simply leaves Black a piece up. With his next moves, Black puts this
bishop on g6, in lieu of the pawn that just went missing in action.) 12...Nc6

42
13.0-0 White has to go back to his original plan of short castling. The position is still very complex as Black cannot
really exploit White’s g4-pawn immediately. The game might develop as follows:
a1) 13.h4? This is just too provocative. White needs to complete his development first. 13...Bxg4 14.fxg4 Nb4
15.Qd2 Nxe4 16.Nxe4 Qxe4 17.Kf2 Nc2 18.Bf4 Nxa1 19.Rxa1

19...a5!? Immediately exploiting the temporar y disorganization of White’s pieces. Black needs to get access to the
d4-square for his g7-bishop when he’ll just dominate White’s pieces, although 19...Be5, removing White’s Bf4
defender, is not bad either. In both cases, Black has full material equality and the initiative. (19...Be5 20.Bh6 Re6
21.Rf1 Rf6+ 22.Kg1 Rxf1+ 23.Kxf1 Qh1+ 24.Kf2 Qxh4+–+) 20.a4 Be5 21.Bg5 Ra6!

43
Black manages to find a way to bring his inactive rook into the game with devastating effect.
22.Rc1 Qc6! 23.Rf1 Rb6!–+ Not allowing the transfer of White’s e2-bishop to the d5-outpost.
a2) 13.g5?! Nh5 A tempting option for Black as it stops any potential white attack, however this allows White to
transfer the focus of the game back to the center. (13...Nb4!? The idea of sacrificing on e4 is another typical idea in
such structures, especially when White’s king is still in the center, however this time it is not as effective. 14.Qd2
Nxe4 15.Nxe4 Bf5 16.Bd3 b5 17.0-0 Rad8!

with mindboggling complications, for example 18.Rad1 Nxd3 19.Qxd3 d5 20.Bxc5 dxe4 21.Qxd8 Qxd8 22.Rxd8
Rxd8 23.cxb5 Bh3 24.Re1 Rd5 25.Rxe4 Rxg5+ 26.Kf2 Be6 27.a4 Rh5 28.h3 Bxb2 29.Re3²) 14.Nd5 Qd8 15.0-0-0
a5 16.a4 Be6 17.Kb1 b6!

44
This little pawn move allows Black to protect the c5-pawn before the future invasion starts with ...Nb4. White’s king
is open and he must go on to the defensive. Black has a very comfortable advantage and an easy game plan to
follow.
a3) 13.0-0-0?! a5³ It becomes obvious that Black’s attacking chances on the queenside are greater than White’s on
the kingside.
13...h5 14.g5 Nd7 15.Nd5 Qd8

16.f4!? An ambitious plan, but Black should have enough defensive resources.
[16.Rad1 a5 17.a4 Nb6 18.Nxb6 Qxb6 19.Rxd6 Qc7 20.Rd5 (20.Rd2 b6 21.f4 Nb4 22.Qb1 Bd7 23.Nc1 Bxa4
24.Nd3 Bc6 25.e5 Rad8 26.Nxb4 Rxd2 27.Bxd2 cxb4³) 20...Bh3 21.Rfd1 Nb4 22.Qd2 b6! 23.Bf4 Qc6!° with strong
compensation for the sacrificed pawn. White’s b3-knight is truly terrible, serving no function and hindering his own
pieces and pawns.]
16...a5 17.f5 Nb4 18.Nxb4 axb4 19.f6 Bf8 20.Nd2 Ne5 21.Nf3 Nc6 22.Nh4 Bh3 23.Rf4 Ne5 24.b3 Ra6 25.Qd1 Qd7
Black has successfully managed to prevent any white sacrifices on h5 to get at his king. The position is in a bit of an
45
impasse as Black cannot break through on the queenside and White cannot break through on the kingside, while
each is stuck with multiple pawn weaknesses. 26.Qc2 (26.Bxh5?? gxh5 27.Qxh5 Bg4–+) 26...Bg4 27.Bf1 Rea8

28.Bc1³ A great defensive move that takes the important a3-square away from the black rook. (28.Qd2? Ra3 29.Qc2
b5 30.cxb5 Be6 31.Nf3 Bxb3 32.Qd2 Nxf3+ 33.Rxf3 Rxa2 and Black breaks through to win.)
b) 11...Na6!?

A strong developing move that intends to exploit the awkward position of the white queen on c2. 12.g5
b1) 12.a3 This is too slow. 12...Nc7! Black is ready for the ...d6–d5 breakthrough. (12...Bxg4!? 13.fxg4 Nxe4
14.Nxe4 Qxe4 15.Qxe4 Rxe4 16.Nc2 Rae8 17.Kd2 Rxe3 18.Nxe3 Bh6 19.Kc2 Rxe3 20.Rae1 Nc5 21.Bd1 Rh3
22.Re7 Bg5 23.Rc7 Bf4 24.b4 Na4‚ ) 13.g5 Nh5 14.0-0-0 d5³ and Black takes over the initiative.
b2) 12.0-0-0 Nb4

46
13.Qb3 (13.Qd2 d5 14.g5 Nd7 15.a3 dxe4! 16.Nxe4 Na6 White’s opening gamble with g4 has not paid off. Black is
simply better, especially the monster ‘Gufeld bishop’ on g7. White must now attend to the defense of his open king.)
13...a5! This is stronger, as Black prepares for the ...d6–d5 break and also a possible ...Nf6–d7–c5 maneuver,
creating strong counterchances. (13...d5!? is a possible attempt to break open the position, however White has
enough pieces in the center to counter it. 14.g5 dxc4 15.Bxc4 Nfd5 16.Nxd5 cxd5 17.Bxd5 Nxd5 18.Qxd5 Bh3
19.Rhg1 Red8 20.Qb5² Black has some compensation for the sacrificed pawn, but it’s not clear if it will be
sufficient.)
12...Nh5 13.0-0-0 Nc5

This is one of the standard locations for the black knight. With this move, Black prevents White’s f3–f4 push and
further expansion in the center. In addition, he prepares to hit White’s main strength, the e4-pawn, with the thematic
...f7–f5 break. 14.Rhe1 a5 It is quite possible for Black to play more slowly, now that White’s kingside attack has
been stopped. Typical KID minor piece play takes place now. (14...f5 15.gxf6 Qxf6 16.Kb1 with a messy
middlegame ahead) 15.f4 (15.Kb1 Be5 16.h4 f5 17.gxf6 Qxf6 18.Rg1 a4 19.Bg5 Qg7 20.Be3 Nf4 21.a3∞ Both sides
keep chances of success.) 15...Nxe4 16.Bxh5 Nxc3 17.Qxc3 gxh5 18.f5
47
18...Be5
[18...Qd8 19.Qd3! A stronger choice for White, keeping the f5-pawn in place and restricting Black’s c8-bishop for
the moment. The position is close to equal. But if 19.f6 Bf8µ Black’s f8-bishop doesn’t look like a happy camper,
however Black does enjoy the pair of bishops, a temporary extra pawn and the possibility of invading White’s camp
via the light squares. Black seems to be in control of the game.]
19.Qc2 a4! 20.Nf3 Ra5!!

What a rook lift!! Now all the black pieces are coordinating. White retains close to equal chances since he hasn’t
made any big mistakes, but he’ll have to be careful not to slip into a worse position. 21.Kb1 Bxf5 22.Qxf5 Bc3
23.Qc2 Bxe1 24.Rxe1 Qe4 25.Bd2 Qxc2+ 26.Kxc2 Rxe1 27.Nxe1 Rf5 28.Nd3

48
28...Kg7 (28...Rf3 29.Bf4 d5 30.cxd5 cxd5 31.Kc3=) 29.Bf4 (29.Nf4 h4³ Black is pressing quite hard.) 29...Kg6
30.Bxd6 Kxg5 31.b3 axb3+ 32.axb3 Kg4 33.Nc5! b5 34.Nd3 and White should hold a draw.

12.g5?!

I was afraid that Black would get to my king first, so decided to prevent any possible ...d6–d5 break. However this
allows Black to shut down White’s further kingside advance by playing 12...Nh5. Stronger, of course, was the brave
12.0-0-0, which I saw during the game, but thought too risky. Strange psychology for one who had just played 11.g4
a move earlier!!
12.0-0-0 b5

13.Bf4! (13.g5 Now that Black has committed himself to the ...b5 push, this move makes sense, though Black is
quite OK here. 13...Nh5; 13.Rhg1?! with the idea of 13...b4 14.Na4 c5 15.Nf5? gxf5 16.gxf5 Bxf5µ 17.Bf4 Bg6
18.Bxd6 Qb7 19.Nxc5 Qb6 still doesn’t work for White.) The text is best, as White simply gets out of all the tactics
involving this bishop on the e-file. The game becomes difficult for Black as Nf5 sacrifices finally become
dangerous, for example 13...Bb7, with the idea of finally developing the b8-knight, leads to massive complications:
49
a) 13...d5? 14.Bxb8±
b) 13...Nfd7 14.h4 (14.Rhg1 Ne5 15.Kb1 Bb7 16.Qd2!

The key move in such positions. White not only prepares his potential Bh6 invasion, but also leaves the c2-square
available for his d4-knight. 16...Nbd7 17.Bh6 Bxh6 18.Qxh6 b4 19.Nf5 Qf8 20.Qxf8+ Nxf8 21.Na4 gxf5 22.gxf5+
Kh8 23.Rxd6 Red8 24.c5!ƒ This little pawn push is the reason why White can fight for an advantage a piece down in
the endgame. Black still has to be accurate.) 14...Nb6 (14...Ne5 15.Qd2 Nbd7 16.h5 Bb7 17.Nc2!? Nb6 18.cxb5
axb5 19.Qxd6 Qxd6 20.Rxd6 Bf8 21.Rd2 b4 22.Nd1 Rxa2 23.Kb1 Raa8 24.Nde3 c5 25.b3!? Bg7 26.h6 Bh8
27.Rhd1²

In this very messy position it is easy to lose track of who is doing what and who is better. However it does seem that
White is still pressing for the advantage, due to his superior pawn structure. Black’s defense looks difficult.) 15.h5
Be5! Another standard defensive idea. Black attempts to regain control over the dark squares in the center.

50
16.Bxe5!? (16.Qd2 b4 17.Nb1 c5 18.Nb3 a5 19.Bxe5 dxe5 20.hxg6 fxg6 21.Qd6 Qxd6 22.Rxd6 N6d7 23.a4! Bb7
24.Rhd1 Re7 25.g5 Kf7

26.R1d5! A beautiful positional sacrifice in the style of Petrosian. Black equalizes, but some accuracy is required.
26...Bc6 27.N1d2 Re6 28.Rxe6 Kxe6 29.Nxc5+ Nxc5 30.Rxc5 Kd6 31.Nb3 Nd7 — 31...Bxa4? 32.Rd5+ followed
by 33.Nc5 would be a bad idea for Black — 32.Rxa5 Rxa5 33.Nxa5 Bxa4 34.b3 Bc6 35.c5+ Kxc5 36.Nxc6 Kxc6
37.Kd2 Nc5 38.Bc4 Kd6 39.Ke3 Nd7=) 16...dxe5 17.Nb3 White’s idea is simple. He wants to prevent Black’s pawn
expansion on the queenside by offering the c4-pawn as a sacrifice. If Black accepts, he’ll be stuck with doubled and
very weak c-pawns, something which would be extremely dangerous in the potential endgame.

51
17...Be6 18.Na5 Qb4 (18...Ra7 19.Qd2 Rd7 20.Qe3 Nxc4 21.Bxc4 Bxc4 22.Rxd7 Nxd7 23.Rd1! Be6 24.Nxc6 Qc5
25.Qxc5 Nxc5 26.b4 Nd7 27.h6! and the endgame is suddenly lost for Black.

White’s expansion on the kingside didn’t result in a mating attack, but created a strong potential passer on the
opposite color to Black’s bishop. This is what chess is all about: some form of energy inherent to each piece keeps
converting into something else.) 19.Nb7 Ra7 20.Nd6 Rf8 21.a3 Qc5 22.Qd2 Nxc4 23.Bxc4 Bxc4 24.Qh6 Be6
25.Kb1 Rd7

52
26.Nf5!+– The final decisive knight move which ends its tour on a high note.
14.Nf5 gxf5 15.Bxd6 Qd8

16.e5! Very strong! White attempts to close off Black’s key defender, the g7-bishop.
[16.gxf5 allows Black to eliminate the d6-bishop and thereby most of the danger to his position: 16...Nd5! The only
move. 17.Bxb8 (17.c5 Ne3 18.Qd2 Nxd1 19.Bxd1 Nd7 20.Rg1 Kh8 21.f4 Qf6 22.e5 Qxf5 23.Bc2 Qh3 24.Ne4 Bh6
25.Kb1 Rg8–+ and Black wins) 17...Qg5+ 18.f4 Nxf4 19.Bxf4 Qxf4+ 20.Kb1 Rad8

53
Black needs to exchange heav y pieces to eliminate the danger to his open king. 21.Bh5! Regrouping to activate the
e2-bishop. 21...c5 22.Qg2 Bxe4+ 23.Nxe4 Rxd1+ 24.Rxd1 Qxe4+ 25.Qxe4 Rxe4 26.Rd7 Bd4 27.Bxf7+ Kf8 28.Bd5

28...Rf4 29.Rf7+ Ke8 30.cxb5 axb5 31.Rxh7 Rf2 32.b4 c4 33.Bc6+ Kf8 34.Bxb5 c3 35.Rc7 Rb2+ 36.Kc1 Rxa2=
Finally, after all the complications, the game has transposed into an equal endgame.]
16...fxg4 17.Rhg1 Nbd7

54
18.f4! Re6 19.c5 (19.exf6!? Qxf6 20.c5 makes sense too) 19...Nd5 20.Nxd5 cxd5 21.Bxg4 Rg6 22.Kb1 Bc6 23.Qf2
(23.f5 Rxd6 24.exd6 Ne5³) 23...Kh8 24.h4 and despite his extra minor piece, Black is completely tied up thanks to
White’s powerful d6-bishop which cements his advantage.
Back to 12.g5.

12...Nfd7?!

The provocation continues. Black puts too much trust in the solidity of his position by playing such backward
moves, but such was the spirit of my opponent. Preferring to face danger and resourceful play in inferior situations
made his games a must-study for people. One of the reasons I chose this line was in memory of another childhood
idol, my compatriot, little-known outside Russia, the legendary Tatar IM, Nezhmetdinov, who was one of the few
people to have a positive score against Tal. Nezhmetdinov played a similar brand of chess, so when these two met, it
was fire versus fire, the board virtually crackling with explosive energy. I did not pretend to try to imitate
Nezhmetdinov’s style, but approached it in my favorite Lasker way, trying to play attacking chess against a player
who is a known attacker himself.
55
12...Nh5! This move leads to a lockdown on the kingside, but the knight will be stuck on the side of the board with
no easy path to a good square, with the result that Black will have to play in the center and on the queenside without
his h5-knight. Both Black’s attacking chances on the queenside and White’s on the kingside are likely to be
diminished and the game will probably continue with slow positional maneuvering. 13.0-0-0 Nd7 14.Qd2 Nc5

15.Nb3! This move is yet another typical way for White to fight Black’s best piece, the c5-knight, by attempting to
exchange it.
[15.Rhg1 Bh3! Preventing White’s h2-h4, which would support his g5-pawn. 16.Kb1 Rad8 17.Bf1 Bc8 18.h4?
(18.Bf2 Be5 19.Be3 Bg7=) 18...d5µ]
15...Be5 16.Nxc5 dxc5 17.Na4 Bd4 18.h4

18...Be6! (18...Rd8 19.Nxc5 Ng3 20.Rhe1 Nf5 21.Bxd4 Nxd4 22.Qa5±) 19.Bxd4 cxd4 20.Qxd4 Ng3 21.Rhe1 Rad8
22.Qc5 Qc7 23.Bd3 b5 24.Nc3 Qf4+ 25.Kb1 Qxf3 26.Qxc6 bxc4 27.Bc2

56
With an unbalanced game where both sides have chances. 27...Qf4 28.Qxa6 Qxh4 29.Rxd8 Rxd8 30.Qb6 Re8
(30...Qxg5? 31.Rg1 wins material) 31.Qe3 Nh5 32.Rf1 This move allows White to take control of the critical f4-
square. 32...Qh2 33.Qc5!= Preventing ...Qe5.

13.h4

Necessary, otherwise White’s previous play makes no sense.

13...b5?!

It was time for Black to develop.

13...c5! 14.Nb3 (14.Nf5 Bxc3+! 15.Qxc3 gxf5–+) 14...Nc6 15.0-0-0
[15.Nd5 Qd8 16.h5 a5 17.hxg6 fxg6!?

57
18.a3 (18.a4 Nb6! 19.Qd2 Nxd5 20.cxd5 Nb4 21.Kf2 Rf8 22.Kg2 Rf7 23.Rh4 Bd7 24.Rah1 Bh8 25.f4 Qe7 26.Bf3³
with yet another unbalanced position, where Black’s chances may well be better, as he has the superior minor
pieces.) 18...a4 19.Nd2 Nb6!³ (19...Nd4 isn’t as good: 20.Bxd4 Bxd4 21.f4 Nb6 22.Nxb6 Qxb6 23.0-0-0∞)]
15...a5

16.Nd2!? White leaves his knight on c3 to prevent ...a5–a4 and also retains the option of playing Nb5, covering the
d4-square.
a) 16.Nb5 a4 17.Nd2 Nf8 Black is ready to send his second knight to a perfect outpost on the d4-square. (17...a3!?
18.bxa3 Nd4 19.Bxd4 cxd4 20.Nb3 Nc5µ)
b) 16.a3 a4 17.Nd2 Nd4 18.Bxd4 cxd4 19.Nd5 Qd8 20.f4 Nc5 21.h5 b5 22.Kb1 Bb7 and Black is fighting for an
advantage, since his g7-bishop has no counterpart to oppose it.
16...Nb4
a) 16...Nd4 17.Bxd4 cxd4 18.Nd5 Qd8 19.Nb3 Nb6 20.Nxd4 Nxd5 21.cxd5 Qb6 22.Bb5 Re7 23.Qa4!

58
23...Bg4!! A very pretty idea to avoid the lockdown. 24.Rhe1 Rc8+ 25.Kb1 Bh5! Black’s position is suddenly
looking good, as White’s kingside pawns can become easy targets for his dark-squared bishop. White can keep the
delicate balance with accurate defense, but cannot hope for more.
b) 16...Nb6 17.Qb3 Nb4 18.Kb1 Be6 19.h5 a4 20.Qa3 d5 21.cxd5 Bxd5 22.hxg6 hxg6 23.Rc1 Be6 24.f4! Red8
25.Ka1 Rac8∞ The position remains very complex.
17.Qa4! Finding a safe harbor on a3, where it will also prevent the further opening of White’s king.
a) 17.Qb1 Nb6 18.a3 (18.Bf4 Bd7 19.a3 — if 19.Nf1?! then 19...d5! — 19...Ba4! Black must keep his b4-knight in
place to keep White passive. Accepting Black’s sacrifice would be dangerous as all the black pieces would come for
the white king. 20.axb4 axb4‚) 18...d5!!

This position is hard to evaluate. It seems that Black has very good chances of his queenside and central
counterattack working. 19.cxd5 (19.Nb5 Na4µ) 19...N4xd5 20.Nxd5 Nxd5 21.Bf2 (21.Nc4 Nxe3 22.Nxe3 Be6µ)
21...Nf4µ
b) 17.Nd5?! Qd8 18.Qb3 a4 19.Qa3 Nxd5 20.cxd5 b5µ with a devastating attack.
59
17...Nb6 18.Qa3 Be6 19.h5

19...d5! 20.cxd5 Bxd5 21.Nxd5 N6xd5 Black has finally opened the position successfully, although White maintains
equality.

14.h5?

Consistent, but not best. For some reason I was afraid to get into tactics, trusting my opponent’s ability too much.
14.cxb5!

14...axb5
[14...Bxd4 Such an anti-positional move can hardly be recommended in view of how badly Black’s king is now
weakened, but since this is the computer’s first line of defense, we’ll look at it briefly. 15.Bxd4 axb5 16.h5 Na6
17.0-0-0 Nac5 18.Kb1

60
18...Ne6 (18...b4? The pawn must remain on b5, to deny White’s e2-bishop the c4-square. 19.hxg6 fxg6 20.Bc4+ is
immediate curtains for Black) 19.Be3 b4 (19...Ba6 20.b4!) 20.Na4 Ba6 21.Bxa6 Rxa6 22.b3 c5 23.hxg6 fxg6
24.Qc4 Rea8 25.Qd5!

Preventing ...Nb6. 25...Kf8 26.Rh2! Nc7 27.Qc4 Ne5 28.Qe2 Kg8 29.f4 Nf7 30.Bc1!± Again, White is utilizing the
idea we have seen previously. His dark-squared bishop is heading to the b2-square, where it will both protect his
king and monitor a major diagonal. Advantage to White.]
15.Nxc6 Nxc6 16.Nd5!

61
This gives White a comfortable advantage. 16...Qe5 17.Qxc6 Qxb2 18.Kf2 and Black has very limited counterplay.

14...b4

15.hxg6?

Another mistake. White should not have allowed Black to retake with the f-pawn.

15...hxg6?!

Returning the favor.

15...fxg6! 16.Na4 c5 17.Nb3 a5 18.0-0-0 Bb7 19.Bf4 Be5 Black is simply much better, as all White’s pieces are
lacking in coordination 20.Be3 Bc6 21.Nd2 Na6 22.f4 Bd4 23.Bg1 Rab8 24.b3 Bxg1 25.Rdxg1 Nc7 26.Bg4 Ne6
27.Bxe6+ Qxe6 28.Re1 Bxa4 29.bxa4 Rf8³ and only Black is playing for an advantage, while White is faced with
the daunting task of protecting his numerous weaknesses and naked king.
62
16.Na4

16...d5?

This move is perhaps my opponent’s only real mistake, and it costs him the game. After the only alternative, 16....c5,
Black would have had good defensive resources.
16...c5 17.Nb3 a5! (But not 17...Nc6? 18.0-0-0 a5 19.Qd2±) 18.0-0-0 Bb7 19.Kb1 Ra6 20.Bf4 Ne5 21.Rh4 White’s
only chance to create something is to triple along the h-file.

21...Bc6 22.Nd2 (22.Rdh1 Bxa4 23.Bd1 f6 24.Qg2 Nbd7–+ simply leaves Black with an extra minor piece)
22...Nbd7 23.Rdh1 Nf8 24.Be3 Ned7 25.Bd1! Ne6 26.Nf1!

63
Finally White manages to achieve his goal. Now the position becomes quite messy, but Black has really good
chances of defense, as White’s a4-knight is too great a liability. 26...Nxg5 27.Qh2 f6! 28.Ka1 Nf7 29.Rg4 Nf8
30.Rgg1 Qd7 31.Rxg6 Nxg6

32.Qh7+ Kf8 33.Ng3! Re5 34.Qxg6 Bxa4 35.Rh7 Ng5 36.Bxa4 Qf7 37.Qxf7+ Kxf7 38.Rh1 Nxf3 39.Bf4 Nd4
40.Bxe5 fxe5 41.Kb1 Ra8³
Black has succeeded in defending against White’s attack, keeping a minor material advantage, which should,
however, lead to a draw.

17.0-0-0?!

17.Nb3! dxe4 18.f4!+–

64
Normally this sort of positional pawn sacrifice is something I would play nowadays without hesitation. White takes
control of the entire center, and with the black pieces denied any squares for development, White’s attack plays
itself.
However during my teenage years I still had that huge dose of respect for my opponent that would often intimidate
me into making less aggressive decisions on the board.

17...c5 18.Nb3 d4 19.Bf4 Nc6

20.Rh4!

White’s plan is clear. Tripling the heavy pieces on the h-file is a huge threat to the vulnerable black king.

20...Nce5 21.Rdh1?

Too hasty, and a serious mistake.

65
21.Kb1!?

Prophylaxis against potential checks on the dark squares and taking on g5 with check was necessary. Besides, White
frees the c1-square to facilitate relocating the b3-knight.
21...Bb7
[21...a5? would be too slow now: 22.Nc1 (22.Rdh1 d3 23.Bxd3 Nxf3 24.R4h3 Nxg5µ) 22...Ra6 23.Rdh1 and Black
cannot avoid White’s heavy pieces tripling on the h-file with decisive threats to the black king.]
22.Nc1! (22.Na5?! d3 23.Rxd3 Nxd3 24.Bxd3 Ne5 25.Be2 Nc6∞) 22...Bc6 23.Rdh1 Qe6 24.Rh7±

Black is still in trouble, as White’s center is now unassailable. 24...Kf8 25.Nd3 Nxd3 26.Bxd3 Bxa4 27.Qxa4 Ne5
28.Qd1 Nxd3 (28...Nxc4 29.Qc1!+–) 29.Qxd3 Rad8 30.Qd2 d3

66
The only move, but now Black’s queenside pawns are ripe for White’s e3-bishop. (30...Qxc4 31.Rxg7 Kxg7
32.Qh2+– Rh8 33.Be5++–) 31.Be3± Qc6 32.Qh2 Be5 33.Bf4 Bd4 34.Rd1 Qe6 35.b3 Bc3 36.Bd2 Bd4

37.f4!! White gets there first! 37...a5 (37...Qxe4 38.Qh6+ Kg8 39.Rh1+–) 38.f5 gxf5 39.g6 Qxg6 40.exf5 Qxf5
41.Bh6+ Ke7 42.Qc7+ Rd7 43.Re1+ Kf6 44.Bg7+ Kg6 45.Qg3+ Qg5 (45...Kxh7 46.Rh1+) 46.Rh6+ Kxg7
47.Qxg5++–

21...d3! 22.Bxd3 Nxf3 23.Rh7!?

67
The only move, however I only counted on 23...Nxg5 and 23...Be5.

23...Be5?

And here is Black’s final mistake. He could have achieved the superior position by developing his queenside.
a) 23...Nxg5 24.Rxg7+ Kxg7 25.Qh2+–
b) 23...Nde5! 24.Nb6 This move is probably the reason for Black discarding ...Nde5, as his d7-knight was protecting
the b6 square. However... 24...Bb7!!µ

Black’s minor pieces are much stronger than White’s clumsy rooks in this position. Black is much better. 25.Nxa8
Bxa8 26.Be2 Nxg5 27.Bxg5 Qxg5+ 28.Qd2 Qxd2+ 29.Nxd2 a5 30.R7h2 a4 Black gets a very comfortable
advantage, aided by his killer bishops and gorgeous knight on e5. It’s probably a technical win against any human.

24.Bxe5 Qxe5

24...Qxg5+ will not save Black either: 25.Kb1 Qxe5 26.Qg2 Nd4
68
27.Nbxc5! The dark squares, which were Black’s advantage during the game, turn out to be his weakness as the
tactics work in White’s favor: 27...Nxc5 28.Qf2 Be6 29.Nxc5 Re7 30.Rh8+ Qxh8 31.Rxh8+ Kxh8 32.Qf6+.

25.Qg2!+–

The only move, but a winning one. White saves his g5-pawn and Black cannot protect his queenside pawns and king
at the same time.

25...Nd4

26.Rh8+?!

The safe way to victory, but there was a flashier, Tal-like, alternative which would have been worthy of this game:
26.Nb6! Nxb6 27.Qg3!! Bh3 28.R1xh3 Nxb3+ 29.axb3 Qd4 30.Qf2! Qe5 31.Qxf7#.

26...Qxh8 27.Rxh8+ Kxh8 28.Nbxc5 Nxc5 29.Nxc5 Ne6

69
30.Na4!!

The only move, but being a very technical player in those days, it was easy for me to see.
White has to transfer his knight to d5, to connect both c- and e-pawns, block the d-file and prepare the advance of
the extra pawns.

30...Bb7 31.Nb6 Rad8 32.Nd5 Kg7 33.Qf2 Bxd5

33...Nxg5 34.Qf6+ Kh6 35.Qf4 Bxd5 36.cxd5 f6

37.e5!! Now this is a move that is worthy of my opponent’s attention. 37...Rxe5 (37...Rxd5 38.exf6!) 38.Qxf6+–

34.Qf6+ Kg8 35.cxd5 Nc5 36.Kd2+–

70
Black’s pieces are paralyzed and he can only watch as White improves his position.

36...a5 37.Qd4 Nd7

37...Rc8 38.e5 b3 39.a3 Red8 40.d6 Ne6 41.Qe4 White needs to put his bishop on the c4-square to win the game.

41...Rc5 (41...Nxg5 42.Qg4 Ne6 43.Bxg6+–) 42.Bc4 Ng7 (42...Re8 43.d7 Rd8 44.Bxe6 fxe6 45.Qxg6++–) 43.Kd3
Rdc8 44.Bxb3+–

38.Bb5 Rf8 39.e5

71
The rest needs no comment.

39...Nb8 40.d6 Nd7 41.Kd3 Rfe8 42.Bxd7 Rxd7 43.Ke4 Rdd8 44.Kd5 Rb8 45.d7

Black resigned.

1–0

Not a very spectacular game, but a good example of Lasker’s psychological approach. My opponent hated positions
with fixed structures, and therefore tried to avoid them by not playing ...c5. This damaged his position, for in the end
he still had to play it, but in a much worse scenario. My gamble could have backfired several times though, had
Black been more inclined to punish me for my hasty moves.

72
 GAME 3
Gata Kamsky
Vassily Ivanchuk
Tilburg 1990
This game, and the opening set-up I found over the board for White against Bird’s Variation of the Spanish, had a
big influence on this particular line, as it was quite popular at the time. White’s plan with the pawn push f4–f5 is
very unpleasant for Black and caused the line to virtually disappear over the next few years. Tilburg was one of the
top events for the world elite and winning it together with GM Ivanchuk not only brought me the final GM norm
needed to obtain the GM title, but also recognition that my high rating was not a fluke.

1.e4 e5 2.Nf3 Nc6 3.Bb5 Nd4 4.Nxd4 exd4 5.0-0 Bc5 6.d3 c6

7.Ba4

Previously the other retreat was played more often, leading to doubled d-pawns in the center for Black:
7.Bc4 d5 8.exd5 cxd5 9.Bb3 Ne7 with a comfortable game.

7...Ne7

7...Qh4 This move is sometimes played to prevent White’s queen from developing in similar fashion to the kingside.
8.Nd2

73
8...Ne7 (8...d6 9.c3 The threat of a double attack on d4 forces Black to give up his important central pawn, yielding
the center and hence the advantage to White. 9...dxc3 10.bxc3 Ne7 11.d4 Bb6 12.Nf3 Qh5 13.Ba3 Bc7 14.e5 dxe5
15.Qb1!‚

15...a5 16.Re1 Ng6 17.Bd1 Qf5 18.Bc2 Qe6 19.Qb2 Bd6 20.Bxg6 hxg6 21.Bxd6 Qxd6 22.Rxe5+± and despite the
reduced material, White’s initiative is very strong) 9.c3 dxc3 10.bxc3 d5

74
11.d4 (11.Nf3 is also possible, but it’s best for White to wait with Nd2-f3. As the well-known saying goes, ‘the
threat is stronger than its execution’. 11...Qh5 12.exd5 b5 13.Bc2 cxd5 14.Bf4 0-0 15.d4 Bb6 16.Bd6 Re8 17.Re1
Be6 18.Qb1± White manages to attack Black’s weak b5-pawn.) 11...Bd6 (11...Bb6?! 12.Ba3! and Black can forget
about castling kingside. 12...dxe4 13.Nxe4‚ ) 12.e5 Bc7 13.Ba3! (13.Bc2?! 0-0 14.Ba3 Re8 15.Rb1 b5 provides
some measure of counterplay for Black) 13...Ng6

14.f4! Perhaps the strongest.

[14.g3?! Qg4 (14...Qh3 15.c4 dxc4 16.Qf3! Qd7 17.Rad1 Rb8 18.Nxc4 b5 19.d5 bxa4 20.e6 fxe6 21.dxc6 Qf7

75
22.Rd8+!! Bxd8 23.Nd6++–) 15.f3 Qh3 16.c4!

Perhaps the most forcing line! 16...dxc4 (16...Be6 17.Rb1‚) 17.Nxc4 b5 18.d5! h5 19.Qe2 h4 (19...bxa4 20.Nd6+
Kd8 21.Nxf7+ Ke8 22.Nxh8 Nxh8 23.Qc4 Bb6+ 24.Kh1 c5 25.e6 White’s attack plays itself ) 20.Nd6+ Bxd6
21.exd6+ Kd8 22.dxc6 bxa4 23.c7+ Kd7 24.Qb5++– with mate to follow.]
14...Bf5 (14...Nxf4?? 15.Qf3+–) 15.Rf2!!

76
White prepares to transfer his d2-knight to the e3-square, where it will attack Black’s f5-bishop which is blocking
White’s pawn advance on the kingside. 15...0-0-0 16.Bc2!± White’s space advantage threatens to become decisive.
16...Bxc2 17.Qxc2 Nf8 18.Nf3 Qh5 19.Rb1 g6 20.Qa4+–

8.f4!

I was very proud to come up with this whole plan during the game, and it was only later that I found out that former
World Champion GM Boris Spassky had played this line as White much earlier.
Black is faced with the unpleasant choice of allowing f4–f5, blocking the c8-bishop and at the same time creating
opportunities for a major attack with Qh5, Nd2, Rf3–h3 etc., or playing the move in the game.

8...f5?

8...d5 became the latest move around 2015, but still there are not many enthusiasts willing to play this position as
Black. 9.f5 a5 10.a3 0-0
77
11.Bf4 (11.Nd2 This is another alternative. 11...f6 12.Rf4 Bd7 13.Rh4 Be8 14.Qg4 Bf7 15.Nf3‚ White amasses the
attacking pieces near Black’s king. Surely this is not what Black wants on move 15 in a solid Spanish game!?)
11...dxe4 12.dxe4 b6 13.Kh1 Ba6 14.Rf2!

14...Bd6
[14...Ra7 15.Nd2 d3 16.Rf3 dxc2 17.Bxc2 Rd7 (17...f6 18.Nb3 Rd7 19.Qb1 Bd6 20.Nd4 Bxf4 21.Bb3+ Kh8
22.Ne6±) 18.Rh3 h6 19.Bxh6‚ Hopefully this line illustrates how dangerous the black position is after the opening.]
15.e5 Bc7 16.f6 gxf6 17.exf6 Ng6

78
18.Bxc7 Qxc7 19.Nd2 Rfe8 20.Qh5 Qd6 21.Bxc6 Re5 22.Qh6 Qf8 23.Qh3 Rd8 24.Nf3

24...Bc8 (24...Re3? 25.Qh5 h6 26.Bd5+–) 25.Qg3 Re6 26.Bb5 Rxf6 27.Nxd4 Rxf2 28.Qxf2 Qc5 29.c3±
White manages to keep his extra pawn and better pawn structure.

9.Bb3 d5

79
Forced, but now the black king becomes stranded in the center and White can make sure of the e5- and g5-squares
for his knight. After just nine moves, White has a very serious advantage.

10.exd5 Nxd5?!

This move allows White the possibility of later exchanging this knight.
Perhaps it was better for Black to recapture on d5 with the pawn.
10...cxd5 11.Qh5+ g6 12.Qh6 Kf7 13.Nd2 Ng8

14.Qh3
[14.Qg5!? Clearly a computer move! It comes with the very clear idea of transposing into a superior endgame.
However human thinking would perhaps be not to leave the queens on board, as it would allow White to keep his
attacking chances. 14...Qxg5?! 15.fxg5 Ne7 16.Ne4

80
16...Bb4 (16...Bb6 17.Nd6+ Kg7 18.Re1 Nc6 19.Ne8+ Kf8 20.Bf4!+– Suddenly, even without queens, the black
king finds itself in a mating net.) 17.Bf4 Be6 18.Nd6+ Bxd6 19.Bxd6

19...Nc6 (19...Rhd8 20.Bc5 Nc6 21.Rf4± White wins the black d4-pawn.) 20.Rae1 Rad8 21.Bf4 Rhe8 22.g4 (22.Re2
a6 23.Rfe1 Re7 24.a3 b5 25.Kf2 Rdd7 26.Kf1 and White still has to prove his advantage.) 22...Re7 23.Bg3

81
23...Bc8 (23...Na5 24.gxf5 gxf5 25.Ba4 a6 26.Rxe6 Rxe6 27.Bc7 Rg8 28.Bxa5 Rxg5+ 29.Kf2 and White is close to
winning this endgame.) 24.Rxe7+ Kxe7 25.Re1+ Kf7 26.gxf5 Bxf5 27.Kf2 Be6 28.Ke2 Ke7 29.Kd2 Kd7 30.Rf1 a6
31.a4 Rg8 32.Rf6 Ke7 33.Bc7 Rg7 34.Bb6 Kd6 Black is barely holding his position, with some hopes of surviving
the ordeal.]
14...Nf6 15.Nf3 Kg7 16.Bd2 Qd6

17.Rfe1 (17.Rae1 a5 18.a3 a4 19.Ba2 Re8 20.Rxe8 Nxe8 21.Re1 Nf6 22.Qh4 Bd7 Again White has the superior
position, but he still has to prove that it is enough for a win.) 17...a5
a) 17...Bd7 18.Re5 Rhe8 19.Rae1 b5 20.a3 a5 21.Qg3 a4 22.Ba2 Rab8 23.Qf2 Ng4 24.Qh4 Nf6 25.h3

82
25...b4 This breakthrough by Black is not that great, because he exchanges his only defender of the weak and
doubled d4-pawn. On the other hand, Black is doomed to passivity as White has numerous plans at his disposal,
including the g2–g4 break. (25...Rbc8 26.g4 fxg4 27.hxg4 Bxg4 28.Rxe8 Rxe8 29.Rxe8 Bxf3 30.Rc8 h5 31.f5 Bg4
32.Bg5 Nd7 33.Qh2! forcing the game into an easily won endgame) 26.axb4 Bxb4 27.Bxb4 Qxb4 28.Rb1 Qc5
29.Qf2²
b) 17...Ne4!? With the idea of keeping White’s rook from the e5square, which, if allowed, would enable White to
develop his pieces to optimal squares in preparation for the attack. 18.dxe4 fxe4 (18...dxe4? 19.Ng5 wins of course)
19.f5

The only move to lead to the win, but a good one. 19...Bxf5 (19...exf3 20.Qh6++–) 20.Qh6+ Kg8 21.Bf4 d3+
22.Kh1

83
White is winning, for example 22...Qd8 23.Bc7 Bf8 24.Qf4 Qd7 25.Rxe4 Bxe4 26.Qxe4 Qxc7 27.Ng5 Qe7
28.Qxd5+ Kg7 29.Qd4+ Kh6 30.Qh4+ Kg7 31.Rf1+–.
18.a3 a4 19.Ba2 Re8 20.Re5 Be6 21.Rae1 Bf7 22.Qh4

22...Qb6 (22...Ng4? 23.Ng5+–) 23.Bc1 Bd6 24.h3! The black d6-bishop, which covers all the weak dark squares, is
worth much more than White’s e5-rook. 24...Rac8 25.R1e2 h6 26.Rxe8 Rxe8 27.Rxe8 Bxe8 28.Qf2 Bc5 29.b4 axb3
30.Bxb3 Qd6 31.Nxd4 Bxa3 32.Ne6+ Kh7 33.Bxa3 Qxa3 34.Qd4 Qe7 35.Qe5

84
White has reached an ideal position where Black has weaknesses on b7 and d5 and a weak king. Strategically, Black
is lost.

11.Re1+

After this move Black loses castling rights, and more importantly has to spend some tempi to organize his pieces,
which gives White precious time to develop his pieces to ideal squares, especially the b1-knight coming to the e5-
outpost.

11...Kf8?!

More resilient was 11...Kf7, but Vassily was extremely unhappy with his position and started to blitz out his moves,
which was his custom when he was in a much worse position. Somehow I allowed myself to become affected by his
fast play and started blitzing out moves myself, which led perhaps to a unique situation in an elite event with people
on other boards still playing early moves based on theory, and Vasya and I blitzing!

85
12.Qh5!?

a) 12.Bxd5

A stronger move, according to the computer. The idea is to create discoordination among the black pieces and
develop with tempo, but I think my move isn’t worse and, in fact, White should keep his superior forces on the
board since the black king is so exposed. 12...Qxd5 (12...cxd5 13.Qf3 Kf7 14.Nd2 Qd6 15.Nb3 Bb6 16.Qf2 Bd7
17.Nxd4 Rhe8 18.Be3 Kg8 19.a4 Rac8 20.c3 h6 21.a5 Bc7 22.Nf3 with a healthy extra pawn) 13.Re5 Qd6 14.Nd2
Be6 15.Qe1 Kf7 16.Nf3 h6 17.Nh4 Rhe8 18.Bd2 Bb6 19.Qg3 Bd7 20.Nxf5 Qf6 21.Nh4 Bc7 and White still has to
prove that his extra pawn is more powerful than the two black bishops.
b) 12.Nd2!?± The latest Stockfish engine now gives this move as its top choice. The earlier analysis was done with
Stockfish 8 which preferred 12.Bxd5.

12...g6 13.Qh6+

86
13...Kg8

13...Kf7 14.Nd2 Bf8 15.Qh3 Bb4 16.a3 Ba5 17.Bxd5+ Qxd5 18.b4 Bc7 19.Nf3 Bd7 20.Bb2 Bb6 21.Re5 Qd6
22.Rae1 h6 23.Re7+ Qxe7 24.Rxe7+ Kxe7 25.Qh4+ Kf7 26.Ne5+ Ke6 27.Nxg6 Rh7 28.a4 a6 29.Ne5 Ba7 30.Qf2
with a technical win.

14.Nd2 Bf8 15.Qh3 Bg7 16.Nf3 h6 17.Ne5+–

White is clearly strategically winning as Black has too many pawn weaknesses and much worse minor pieces
compared to White.

17...Qf6 18.Bxd5+ cxd5 19.b3

The most practical decision, to attack Black’s weak d4-pawn.

This move also helps White to develop his c1-bishop to its best square on d4.

19...Kh7 20.Bb2 Qb6

87
21.Qf3?

A mistake caused by blitzing, missing an easier continuation: 21.Nf7! Rg8 22.Re7 h5 23.Rae1+– with a mating
attack.

21...Re8 22.Qf2?!

Another inaccuracy.
There is nothing wrong with capturing the d5-pawn.
22.Qxd5 Rd8 23.Qf7! Qf6 24.Qc7!

24...Qd6 (24...Be6 25.Qxb7+–) 25.Qxd6 Rxd6 26.Nc4 Re6 27.Rxe6 Bxe6 28.Re1 Bd5 29.Re7 Kg8 30.Rd7 Bc6
31.Rd6+– White captures a second extra pawn, making this an easy technical win.

22...Bxe5 23.Rxe5 Rxe5 24.Bxd4

88
24...Re2!

The move that I missed, which gives Black certain hopes of a draw in the resulting opposite-colored bishops
endgame.

25.Bxb6 Rxf2 26.Bxf2 Kg7 27.a4 Be6 28.a5 a6 29.Bd4+ Kf7

White finally takes the black d4-pawn, but Black has managed to transpose into an opposite-colored bishops
endgame. Thankfully, the presence of rooks on the board keeps White’s large advantage; the only question is
whether it is winning.

30.Kf2 Rc8

Black had kept up his blitz play, which was part of the reason why he embarked on the following dubious plan of
activating his king. It is not clear whether White can make much progress if Black just stands guard.

31.Ra2!? g5 32.Be5 Kg6 33.Ke3

89
The king goes to the center to support the c2–c4 break, trying to create a protected passed d-pawn.

33...Kh5 34.Kd2?

Inconsequent. There was nothing wrong with Kd4.

34.Kd4! Kg4 35.c3! with the idea of meeting ...gxf4 with Rf2. However, how can one see such finesses while
blitzing? Here is a lesson for younger players: don’t fall for the trick of fast play. Use your time; it is given to you
for a reason. 35...Kh5 36.Re2 Rd8 37.Bc7 Re8 38.Kc5 Bd7

39.Re5! Best, as can be seen from the notes below. White has to play this endgame carefully to avoid Black’s
devious traps.
[39.Rxe8 Bxe8 40.Kxd5 (40.d4 Kg6 41.Kxd5 Bf7+ 42.Ke5 Bxb3 transposes) 40...Bf7+ (40...Bc6+? 41.Ke6 Kg4
42.d4 Bxg2 43.d5 gxf4 44.d6 Bc6

90
45.d7 Bxd7+ 46.Kxd7 f3 47.Bb6 Kh3 48.Kc7 f4 49.Kxb7 Kg2 50.b4 f2 51.Bxf2 Kxf2 52.b5 f3 53.bxa6 Kg1 54.a7
f2 55.a8=Q f1=Q

with a winning game, according to the computer’s queen endgame evaluation. However, White still has a lot work
ahead of him. 56.Qg8+ Kxh2 57.c4 Qb1+ 58.Ka7 Qa1 59.Qd5 Probably best. A centralized queen is always good in
positions like this.) 41.Ke5 Bxb3 42.d4

91
42...Kg6 (42...gxf4 43.d5+–) 43.d5 It seems like White is winning, as the threat of d6 followed by Ke6 forces Black
to give up another pawn. However, the drawing power of the opposite-colored bishops has yet to say its last word.
43...Kf7 (43...gxf4 44.d6+–) 44.fxg5 hxg5

45.Bd8!! The most precise continuation. White makes sure to retain two different distant passed pawns. (45.Kxf5? A
really bad mistake as Black now manages to stop White’s connected pawns on the kingside. 45...Bxd5! After this
move Black can draw this endgame. 46.g4 Bc4 47.Kxg5 Be2 48.h4 Bd1 49.h5 Be2 50.Be5 Bf3 51.Bd4 Be2 52.Be5
Bf3 53.Kf4 Be2 54.c4 Bd1 55.Bb2 Be2 56.c5 Bd1 57.Be5 Be2 58.Kg5 Bd1 59.Bb2 Be2 60.Kh4 Ke6 61.g5 Kf5

92
and we are back to this key position. Black holds the fortress. 62.g6 Bc4 63.g7 Bg8 64.Bd4 Bh7 65.Be3 Bg8 66.Bd2
Kf6 67.h6 Kg6 68.Kg4 Kh7 69.Kf5 Bd5 70.Kf6 Kg8 with a draw) 45...g4 46.Bg5! Bc4 47.h4!+– After this move,
it’s an easy win as Black cannot handle both White’s passed dand h-pawns.]
39...Bc6 ( 39...Rc8 40.Kb6 Bc6 41.Rxf5+–) 40.d4

40...gxf4 (40...Kg6 41.g3 Kf6 42.Kd6 Kg6 43.c4 Rxe5 44.Kxe5 gxf4 45.gxf4 dxc4 46.bxc4 Ba4 47.c5 Bb3 48.d5
Kf7 49.Kd6 Bc2 50.Bb6 Be4 51.c6 bxc6 52.dxc6 Ke8 53.c7 Bb7 54.Ke6 Bc8+ 55.Kf6 Black loses his h-pawn and
the game.) 41.Rxe8 Bxe8 42.Bxf4 Kg6 43.c4 dxc4 (43...Bc6 44.cxd5 Bb5 45.Kd6+–) 44.bxc4 Ba4 45.Kb6 Bb3
46.d5 Kf7 47.Kc7 Ke8 48.d6 Ba4 49.Kxb7 Again with an easy win.

34...Rg8 35.g3 Kg4 36.Ke3 Kh3 37.c3

I didn’t want to risk playing c4 before the time control and the mad blitzing had stopped. Thankfully, Vassily
managed to out-trick himself.

93
37...d4+?

A mistake. Black should have shuffled back and forth, trying to trap me in one of the positions analyzed above
leading to a draw with two extra passed g- and h-pawns in the opposite-colored bishops endgame without rooks.
Now White gets what he wanted.

38.Kxd4 Rd8+ 39.Ke3 Bxb3 40.Rb2 Bd5 41.c4 Bg2

This is what Black was hoping for — his king can finally attack the h2-pawn. However, it suddenly finds itself in a
mating net.
41...Bc6 42.d4 finally achieves the ideal structure of a passed d-pawn supported by the powerful e5-bishop and
centralized king.

42.Rb6!

94
42...Bf1

42...gxf4+ 43.Bxf4 Rh8 44.d4 Kxh2 45.g4+ Kh3 46.gxf5 White has too many passed pawns.

43.Rxh6+?

43.Kf2 would have won immediately: 43...Bxd3 44.Rxh6+ Kg4 45.h3#.

43...Kg2 44.d4 g4

44...gxf4+ was not really an alternative due to 45.Kxf4 Bxc4 46.h4 when the advance of White’s h-pawn is
unstoppable.

45.c5 Bc4 46.Rf6

1–0

Finally Vassily had had enough and extended his hand. It was during this tournament that he became close friends
with my father and me and I’m honored to still call him my friend so many, many years later. It was fitting that
while I finished the first half of the event in clear first place, Vassily ‘won’ the second half and caught me at the end,
so we shared first and second places. The picture taken at the closing ceremony shows it all. I was sixteen years old.

95
 GAME 4
Anatoly Karpov
Gata Kamsky
Tilburg 1991
This tournament was a follow-up to my successful event of the previous year. The 1991 edition was improved with
the participation of both Karpov and Kasparov. While I didn’t manage to get into the top places, I did hold Kasparov
to a 1-1 tie, via two draws, and managed to beat Karpov 1,5–0,5. The following game was key to my minimatch
victory.

1.d4 Nf6 2.c4 g6

The Grünfeld became one of my main openings in those days, and while I wasn’t studying theory very hard yet, I
did have some lines well prepared. Before the game, I felt that Karpov might decide to play something very
positional, and even go for an endgame type of position.

3.Nc3 d5 4.cxd5 Nxd5 5.e4 Nxc3 6.bxc3 Bg7 7.Be3 c5 8.Qd2 Qa5 9.Rb1

This move was really popular in those days, because White believed that provoking Black to play ...a6 weakened
Black’s defensive set-up based on the pawn formation of a7, b6 and c5. With the pawn on a6, the black pawn on b6
would be vulnerable, and White could potentially threaten the possible capture of Black’s c5-pawn as well.

9...a6 10.Rc1

And here it is. White believes in his endgame skills. I was quite happy with the position and felt that Black should
equalize comfortably here. The modern approach to this variation for Black involves heavy analysis following
10...0-0 with ideas like ...Rd8, ...Nc6, ...e6, or the break with ...f5 and Black doesn’t have the option of bringing his
b8-knight to the a6-square. But modern theory gives me a headache; it requires too much memorization, which I feel
is a big problem with contemporary chess.

10...cxd4 11.cxd4 Qxd2+ 12.Kxd2 0-0 13.Nf3 e6

An important move. Black prevents White’s d4–d5, when the weakness of the b6-square might become real, and
Black’s potential rook transfer to the d8-square could be met by Be3–b6. For the above-mentioned reasons, securing
the development of Black’s b8-knight becomes a priority.
96
14.Bc4 Nc6 15.Rhd1 Bd7 16.Ke1

Both sides have completed their development and now is the moment to assess the position with its possible
transitions. White is planning to go for d4–d5, and since his f3-knight is not doing much, there’s a possibility of
transferring it to the c5-square sometime in the future, but at the moment he is waiting for Black to declare the
position of his rooks.

16...Na5!

An important moment. With his last move Black chases White’s c4-bishop away from the good square where it was
supporting the key d4–d5 pawn break. Without the c4-bishop, the central pawn break becomes toothless, and since
the pawn structure is a classic queenside 2 vs. 1 for Black vs. white’s extra central pawn, Black is quite happy to
exchange rooks. Textbook chess rules still apply to top chess, in this case, a distant passed pawn is (in most cases)
superior to a central one.

17.Be2 Rfc8

Black, of course, contests the only open file, the c-file.

18.Bd2

A defensive sort of move. Its one advantage is that White maintains the tension, which would have gone if he had
proceeded with the planned d5-break. For example
18.d5

97
18...b5 A move in the spirit of the Grünfeld Defense. Black simply puts his a5-knight on the c4-square, forcing
White to give up one of his bishops, resulting in a position where Black is at least equal thanks to his strong bishop
pair. White is now forced to declare his intentions in the center.
[18...exd5?! would be a tactical mistake: 19.Rxd5! Rxc1+ 20.Bxc1± White wins a pawn. 20...Bb5 21.Bxb5 axb5
22.Rxb5 Nc4

23.Rc5! Not an obvious move, but the only one to keep White’s advantage. (23.Rxb7?! isn’t satisfactory, e.g.
23...Nd6 24.Rc7 Nxe4 25.Rc2 Nc3 26.Bd2 Nxa2=

98
While it looks like Black’s a2-knight is stranded, there’s no way for White to trap it, because Black’s g7-bishop
controls the vital b2- and c3-squares. Black’s rook will come to either the b2- or b3-square to help the knight get out
by playing ...Na2–c3. 27.Rc4 h6 28.Kd1 Rb8 29.Ne1 Rb1+ 30.Ke2 Rb3=) 23...b5 24.e5! h6 25.Rxb5 Rxa2 26.Rc5±
White retains strong winning chances thanks to his extra pawn and strong central presence.]
19.d6!
[19.Bb6?! Nc4 20.Bxc4 Rxc4 21.Rxc4 bxc4

22.Bd4 With the idea of exchanging the dark-squared bishops and trying to capture Black’s c4-pawn, since the white
king is so close to the scene of the action. (22.d6 Bb2!³) 22...exd5 23.Bxg7 Kxg7 24.Rxd5 (24.exd5 c3³ Here Black
may start thinking about playing for a win.) 24...Bc6 25.Rd4 Re8 26.Rxc4 Bxe4 and White now has to start thinking
about defense.]
19...Rc6 (19...Nc4?! 20.Bxc4 bxc4 21.Bd4 f6 22.e5 fxe5 23.Nxe5 Bxe5 24.Bxe5²)

99
20.e5
[20.Rxc6 Nxc6 21.Rc1 Rc8 22.Bb6! (22.Bd1 f6 23.a4 Bf8 24.axb5 axb5=; 22.Bf4 f6!³ An important, and at the same
time, thematic move for black in the Grünfeld, to prevent White from establishing firm control over the center with
strong support for his far advanced d6-pawn.) 22...f5∞ with a complicated struggle ahead.]
20...Rac8 21.Bg5 h6=

Once again the position is very complex and hard to evaluate, since White’s central pawn duo on d6 and e5 can be
both a great strength and a potential weakness.

18...Nc6

100
19.Bc3

19.d5 would allow the classic Grünfeld knight maneuver after 19...exd5 20.exd5 Ne7 21.d6 Nd5!

Heading for its dream square on c3, where it will be protected by the g7-bishop and take aim at a wide range of
critical squares in the white camp. 22.Ng5 Nc3 23.Bxc3 Bxc3+ 24.Rxc3 Rxc3 25.Ne4 Rc2 26.Nf6+ Kg7 27.Nxd7
Re8 28.Rd2 Rc1+ 29.Rd1 Rc2 30.Rd2 with a draw.

19...Na7!

A very important move. During the game I felt that White had wasted precious time by shuffling his pieces, so
decided to play for an advantage. This move could not have been played without White’s 9.Rb1!!

20.Bd2

A psychological ploy!!! Karpov realizes that he’s not better and offers the chance to repeat moves.
101
20...Nb5³

21.e5!?

A very interesting move. It seems like White has deliberately created a weakness on d4 and given Black the perfect
square on d5. However, on closer inspection, the c5- and d6-squares are potential targets for White’s f3-knight, and
if he manages to exploit them, he won’t be worse at all.

21...Bc6?!

A tempting move, however one of the recurring themes in this book, and something that I want to impart to the
reader, is the rule of improving your least-functioning piece.
In this case, Black’s g7-bishop has just been blocked and is not doing anything. His c8-rook is opposing its
counterpart on c1. In addition, Black’s b5-knight is a menacing piece, all by itself, but with many possible jumps.
So, after all the above reasoning, it was logical for Black to improve his g7-bishop to a better location!
21...Bf8!

102
The bonus of this move is the new threat to take over the c-file with ...Bf8-a3.
22.a4
[22.Rxc8 Rxc8 23.Rb1 Bc6 24.Rc1 Rd8 25.Ba5 (25.Rc4 Bd5 26.Rc2 Bxf3–+) 25...Rb8 26.Bd2 Ba3 27.Rb1 Be4
28.Rb3 Bd5 29.Rb1 Rc8 30.Bxb5 axb5 31.Rxb5 Rc2 32.Bh6 Rxa2 33.Nd2 Be7

Black has a huge positional advantage, since White has no access to any open line to threaten Black’s king, while
Black has a strong pair of bishops, an active rook and strong passed b-pawn. 34.Rb1 Ra4 35.Rc1 Bc6 36.Be3
(36.Nc4 Bxg2µ) 36...Kg7!?µ]
22...Na3 23.a5 Ba4 24.Rxc8 Rxc8 25.Rc1 Rxc1+ 26.Bxc1 Bb4+ 27.Bd2 Nc2+ 28.Kf1 Bxd2 29.Nxd2 Nxd4 30.Bf3
Bb5+µ

103
31.Kg1 (31.Ke1 Nxf3+ 32.Nxf3 Bc6 Pinning White’s knight to the defense of the g2-pawn, with a winning
endgame.) 31...Bc6 32.Bxc6 (32.Bd1 Bd5–+) 32...Nxc6 33.Nc4 Kf8 34.f4 Ke7 35.Kf2 Kd7

36.Ke2 (36.Ke3 Ne7 37.Nd6 Nf5+–+) 36...Nd8 37.Kd3 Kc6 38.Kd4 Kb5–+
The black king has successfully joined the attack on White’s a5-pawn. Black is winning.

22.a4!

104
Unexpected, and strong! White gambles on the psychological fact that Black’s position looks so good, he’ll waste a
lot of time choosing the strongest continuation!

22...Na7?!

Played so that Black can achieve his dream set-up of ...Bd5 and ...Nc6. However, while it looks good, Black has no
strong threats, so White’s psychological gamble paid off!
22...Nc7! would have been much stronger, however I didn’t see the amazing follow-up. 23.Ra1 (23.a5? Ba4 and
Black wins the exchange.) 23...a5!!µ A fantastic subtle move, which I had not seen. Black fixes White’s a4-pawn
and provides additional control of the b4-square for his pieces.

23.Ra1 Bd5 24.Kf1!

A typical Karpovian move. As a former student of ex-World Champion GM Petrosian, he obviously learned all
about prophylaxis a long time ago. As GM Petrosian was one of my favorite players as well, I understood and
admired the technique.
105
24...Nc6 25.a5 Rc7

Quiet, but Black is in no rush and can afford to strengthen his position. The black rooks can now double on either
the c- or d-file. Still, there was nothing wrong with the immediate 25...Bf8!³, improving the least useful piece, the
bystanding bishop.

26.Be3 Bf8

Finally bringing my dark-squared bishop back into the game! Now this piece can potentially join in the attack on
White’s a5-pawn.

27.Ne1!

White follows through with his own plan to improve his f3-knight which wasn’t doing much. There is also an
indirect tactical trick that allows him to defend his a5-pawn.

27...Bb4?

Missing a tactical defense by White.

27...Nb4! was much stronger, of course. Black takes control of the c-file and ties down White’s e1-knight which has
to protect the c2-square. 28.Nd3 (28.Rdc1?! Rxc1 29.Rxc1 Nc6µ and White cannot keep his a5-pawn.) 28...Nc2
29.Rab1 (29.Rac1 Rac8µ) 29...Rc3

106
30.Nf4! (30.Bd2? Ra3µ) 30...Bc6 31.Bd3 Rd8 32.Bd2 Rxd3 33.Nxd3 Bb5 34.Be3 Nxd4 35.Bxd4 Rxd4 36.Ke2 Ra4³
Black has sufficient material compensation for the missing exchange and thanks to his strong pair of bishops, he
retains hopes of an advantage.

28.Rdc1!

Suddenly almost all of Black’s advantage has evaporated. His b4-bishop is going to be hit next move by Ne1–d3,
heading for the c5-square, giving White an extra tempo to complete his plan. I was quite stupefied at this point
because of my blunder on the previous move.

28...Bxe1!

After taking some time to pull myself together, I was quite proud to find this move, which seemingly gives White
the pair of bishops and seriously weakens Black’s kingside. However, Black gets rid of White’s strong knight and is
able to limit the power of White’s bishop pair, while the weak dark squares on the kingside are going to be taken
care of by the black king heading to the g7-square, followed by ...h7–h6.
107
28...Rac8 29.Nd3 Bxa5? (29...Be7 30.Nf4! Taking out Black’s centerpiece on d5 and with it any chances of winning
the game.) 30.Rxc6 Rxc6 31.Rxa5±

29.Rxe1 Nb4ƒ

29...Rac8

A logical alternative, which, however, is too passive. 30.Rec1 Ne7 (30...Nb4 31.Rxc7 Rxc7 32.Bh6! The key move,
after which Black will be tied up and unable to improve his king. 32...Nc2 33.Rc1 Rc6 34.Rd1 Na3 35.f3 Nb5
36.Ke1 Rc2 37.Bxb5 axb5 38.Rc1 Opposite-colored bishops mean an easy draw.) 31.Bd3 Bc4 32.Ke2 Bxd3+
33.Kxd3 Rxc1 34.Rxc1 Nc6 35.Ke4 Rd8 36.Rc5 Rd7 37.h4 Kg7 38.f3 Ne7 39.g4 Nd5 40.Bd2

Black has a nice knight on d5, but White’s activity compensates for it. The natural 40...Rc7 doesn’t win: 41.Rxc7
Nxc7 42.Kd3 Nd5 43.Kc4 Kf8 44.Kc5 Ke7 45.h5 Kd7 46.hxg6 hxg6 47.Bh6 Kc7 48.Bg5

108
With a draw, as 48...b6+?, the only way for Black to attempt to make further progress, backfires spectacularly:
49.axb6+ Nxb6 50.Bd8+ Kxd8 51.Kxb6 Kd7

52.g5!! The key winning move! White successfully elbows out Black’s king to win the pawn endgame. (Instead,
52.Kxa6?? would lose for White after 52...Kc6 53.g5 Kd5 54.Kb5 Kxd4 55.Kc6 Kxe5–+.)

30.Rac1 Nc2!

109
Judging that Black needs both rooks to gain at least some minuscule advantage.

31.Bg5?!

The first inaccuracy by my opponent in this game. Perhaps the psychological effect of having to defend passively
with the white pieces was a factor. The bishop is placed here to prevent Black’s c7-rook from moving, since the
bishop will be relocated to e7 and from there to c5, which is an ideal square for it, but achieving this transfer is like
the ‘American dream’ — many hope for it, but few obtain it!! At this point, however, I realized that my opponent
was still playing for a win and was mighty glad to see the game go on!!
31.Red1 Kg7 ( 31...Rac8 32.Bh6! Rc6 33.Rb1 R8c7 34.f3= Black has absolutely nothing, in fact White has potential
to improve his position further, given the very weak dark squares around Black’s king. 34...Na3 35.Rb2 Nc4
36.Bxc4 Rxc4 37.Bg5=) 32.Bd3 Rac8 33.Rxc2 Rxc2 34.Bxc2 Rxc2 35.Bg5=

31...Rac8 32.Red1 Nb4!?

Something of a psychological ploy. Black suddenly agrees to the exchange of a pair of rooks, however the rooks
110
remaining on the board are not so easy to exchange, and Black hopes to get the rook further along the c-file before
placing his knight back on the ideal c6-square, obtaining slight but unpleasant pressure on White’s weak a5- and d4-
pawns.

33.Rxc7 Rxc7 34.Rb1

Continuing the psychological battle. White would be very comfortable after the careless retreat to c6.

34...Nc2! 35.Bd8

35.Rc1 Rc6 36.Rd1 Kf8!

A great move that prevents the transfer of White’s g5-bishop to the c5-square via e7! 37.Bd8 Rc3 38.Bb6 Kg7
39.Bc5 Rb3 Both sides have achieved what they wanted. White has managed to put his darksquared bishop on an
ideal square, Black to activate his rook without allowing its white counterpart any play in return.

35...Rc8 36.Bb6 Be4

111
Karpov has always been a notorious time trouble player, leaving himself precious few minutes in any position. That
partly explains my last move. I also wanted to prevent White from playing Bd3, trying to get an opposite-colored
bishops endgame, which would have been an easy draw.

37.Rd1™

37.Rb2? Nxd4µ

37...Rc3!

Just in time before White’s b6-bishop closes the c-file, eliminating any hope of victory for Black!

38.Bc5 Kg7 39.Kg1 Rb3!

Black makes sure that White’s d1rook doesn’t get any good squares.

40.f3 Bc6

112
The time control has been reached and it’s time to evaluate the position. It’s not clear what progress has been made
by either side, so confusing has the game been, but I felt confident that I had obtained at least a small advantage,
thanks to my active rook and c2-knight sowing confusion in White’s camp and limiting his options. One of the plans
that I wanted to carry out long term was to prepare a slow pawn advance on the kingside, supported by the king. The
computer of course laughs at all these notions and claims that the position is equal, but imagine the old classic time
control with its limited time for thinking, the lack of a clear plan and the need to be on constant alert for tricks by
Black and psychologically it becomes very uncomfortable.

41.Kf2 Nb4!

Black goes back to his ideal set-up, since White cannot exchange his knight and be in time to defend his weaknesses.

42.Rd2?!

And now a second inaccuracy. It was necessary to prevent Black’s plan.

42.Bc4! Rb2+ 43.Kg3 and Black cannot improve his position much further. 43...g5 44.Re1 h5 45.Re2 Rb1 46.Be7
Kg6 47.Rd2=

113
The point of all White’s rook maneuvering. He has managed to dislodge Black’s rook from the second rank where it
was hitting all the the white pawns on the kingside. Now everything is protected and Black should be careful about
expanding further as he risks the fury of White’s bishops which are potentially playing across the whole board.

42...Bd5µ

Now Black is simply much better.

43.h4 Nc6³ 44.Bb6 h5

a) 44...Ra3 is very interesting of course, but I was afraid of giving White even a tiny amount of activity. The
computer actually agrees with me, as Black has to play very accurately to avoid immediate draws. Thus the idea of
Black attacking and capturing White’s a5-pawn, for all practical purposes, is really impractical after all! 45.Rc2 Ra4
(45...Nxa5? 46.Rc8‚ ) 46.Ke3 h6 47.g4 Ra1 48.Kf2 Nxa5 49.Rc8 Nc6 50.Bc5 Rb1! 51.Bf8+ Kh7

114
52.Rc7 Nxd4 53.Rxf7+ Kg8 54.Rf6 Rb2 55.Ke3 Nxe2 56.Bxh6= and suddenly it’s a draw!!
b) 44...h6!? was more cautious, retaining the option of playing ...g5. I was struck with a much more ambitious idea,
which happened in the game.

45.Rc2 Ra3

46.g4?

A serious inaccuracy, but with time going by so fast and faced with possible ...Nxa5 ideas, White decided to force
the issue. Another chess rule of defense says that exchanges favor the defender, as any endgames with 4 v 3 on the
same flank prove.
a) 46.Rc5 Ra4µ
b) 46.Kg3 Simply waiting was necessary. If Black takes the a5- or d4-pawns, he risks getting mated. For example
46...Ra4 47.Bd3 Nxd4 48.Rc8 Nc6 49.Bc2
115
49...Ra3 (49...Rc4 50.Bd3 Rc3 51.Bd8! and suddenly White is winning) 50.Rc7 Nxa5 51.Bc5 Ra2 52.Bb1 Rb2
53.Be7+– These lines demonstrate how activity is much more important than material advantages.

46...Ra1

Played in the spirit of the position, threatening to hit the newly-created potential weakness, White’s h4-pawn.
46...g5 47.hxg5 (47.gxh5 gxh4³) 47...hxg4 48.fxg4 Kg6µ

47.Rc3?!

After this inaccuracy it will be diffcult for White to save the game. Perhaps he missed Black’s intermediate move.
47.gxh5 gxh5

48.Rc3 The right move order! (48.f4 Rh1 49.Kg3 Be4 50.Rd2 Ne7 51.Bxh5 Nf5+ 52.Kf2 Rxh4 53.Be2 Rxf4+ and
White still has serious defensive chances, as exchanges follow) 48...Rh1 (48...Kh6 49.Kg3 Ra4 50.Bd1 Ra2 51.Bc2
Nxa5 52.Rc8 Nc4 53.Bd8 Rxc2 54.Bf6=) 49.Bf1 Rxh4 50.Kg3 Rh1 51.Kg2 Rh4 52.Kg3=
116
47...Rh1

48.Kg3

48.gxh5 might have been better. 48...Rh2+!! The best chance. 49.Ke3 gxh5 50.Rc1! Ne7 (50...Rxh4 51.Rg1+ Kf8
52.Bc5+ Ne7 53.Bd3 White has finally managed to active his rook, resulting in good compensation for the pawn and
an equal position.) 51.Rg1+ Kf8 52.Bc5 Ke8

53.Bxe7!
[53.Bd3 Rh3 54.Rg7 Bxf3 55.Kd2 Rh2+ 56.Ke3 (56.Kc1 Rxh4 57.Rh7 Nc6 58.Rh8+ Kd7 59.Rh7 Rf4 60.Kd2 Bg4µ
and Black has everything under control, while enjoying two extra pawns.) 56...Bc6µ]
53...Kxe7 54.Bd3 Rxh4 55.Rb1! With the idea of finally eliminating Black’s key piece, the d5-bishop. Once that’s
done Black will have a weak pawn on b7, which, together with the reduced material and possibility of White’s rook
hitting the b7- or f7-pawns, will give him excellent chances of a draw.
117
48...g5!!

The extra exclamation mark is for imagination. At this point, White was short of time again, so this move was
extremely unpleasant, since Black suddenly creates a passed pawn that won’t be so easy to deal with.

49.hxg5 h4+ 50.Kg2

50.Kf2 of course loses on the spot to 50...Rh2+ 51.Ke1 h3 52.f4 Rxe2+ 53.Kxe2 h2–+

50...Re1 51.Rc2?!

a) 51.Kf2?! h3!–+
b) 51.Bc4! The computer finds an anti-human resource which seems to work, although White doesn’t solve all his
problems if Black is accurate. 51...Bxc4 52.Rxc4

52...Re2+!
118
[52...Kg6 53.Kh3 Re3 (53...Kxg5? 54.f4+!) 54.Kxh4 Rxf3 55.Rc1 Kg7 56.Rh1 Ne7

57.g6!! Kxg6 58.Rh3= An amazing position. White seems to hold with precise defense. Of course, this whole line is
beyond a human as White purposefully leaves himself with a bad bishop and king on h4 in danger of getting mated.
58...Rxh3+ 59.Kxh3 Nc6 60.Kg3 Kg5 61.Kf3=]
53.Kh3 Re3! 54.Kg2 Rd3µ Black keeps winning chances.

51...Kg6!

The point. The black king finally joins the battle and White is in big trouble as suddenly his b6-bishop is outside the
war zone.

52.Kf2 Ra1 53.f4?

A human move, protecting the pawn, but it opens up Black’s central d5-bishop, which becomes a decisive factor.
53.Rb2!
119
This was White’s last chance, preparing to transfer his way ward b6-bishop back into the game, where it is needed.
53...Rc1! The only move that keeps Black’s advantage. In this very complex position White might have sufficient
resources for a draw once his b6-bishop gets back into play.
a) 53...Kxg5 54.Bc5 Rxa5 55.Rxb7 Ra2 56.Ke3 Rxe2+ 57.Kxe2 Nxd4+ 58.Bxd4 Bxb7 59.Be3+ Kg6 An important
position, but White manages to draw because Black’s bishop is of the ‘wrong color’ for the a-pawn.

White trades pawns and gives up his bishop to reach a theoretical draw: 60.Bb6 h3 61.Kf2 h2 62.Kg2 Bxf3+
63.Kxh2 Bxg4 64.Kg3 Kf5 65.Bc7 Bh5 66.Kf2 Ke4 67.Ke1 a5 68.Kd2 a4 69.Kc3 Kf5 70.Kb4 Bd1 71.Ka3 f6
72.exf6 Kxf6 73.Kb2 e5 74.Bxe5+ Kxe5 75.Ka1=.
b) 53...Ra3

120
54.Rb1!! (54.Bc5 Rc3!µ) 54...Kxg5 55.Rh1!= Another amazing defensive resource. Black cannot make further
progress and the game is equal.

53...Ra3!–+

White’s position is becoming critical.

54.Rd2 Rg3!!

A key move, transferring the black rook back to the kingside, cutting off White’s king from Black’s passed h-pawn.
It is still not too late to make a mistake, but after the game move, Black is finally winning as the h-pawn is moving
fast.
54...h3? is tempting, but wrong. 55.Kg1! Suddenly White is fine, and in case of 55...Rg3+ 56.Kh2 Rg2+ 57.Kxh3 he
is even better:

121
57...Kg7 58.f5 Bf3 59.Bxf3 Rxd2 60.f6+ Kg8 61.g6 fxg6 62.Kh4±.

55.Bc5

Creating last-minute mating threats with the help of the Bd6–f8 transfer.

55...Kg7!

Prophylaxis until the last move! Patience!

55...Rg2+ 56.Ke1 Kg7!–+ The winning idea. (56...h3? 57.Bd3+ Kg7 58.Rxg2 hxg2 59.Kf2 Nxa5 with advantage,
but probably not enough to win).

56.Rd3

This allows a quick end, however other moves were no better:

a) 56.Bf1 Rxg4–+
122
b) 56.f5 Rg2+ 57.Ke1 h3 58.f6+ Kg8 59.Rd3 h2 60.Rh3 Rg1+ 61.Kf2 h1=Q–+

56...Rg2+ 57.Kf1 Rxe2!–+

A final small tactic. White loses a piece and it only remains to calculate how to deal with White’s mass of pawns.

58.Kxe2 Bc4–+ 59.d5 exd5 60.Kd2 Bxd3 61.Kxd3 d4 62.Ke2 h3 63.Kf3 d3

63...h2 was a faster win, but at this point I saw the final position and decided go with this line.

64.Be3

64.Bg1 Nd4+!–+

64...Nd4+?!

a) 64...Nxa5 65.f5 Nc4 66.Bf4

123
b) 64...h2 was an easier win: 65.Kg2 Nxa5 66.Kxh2 Nc4 67.Bc1 a5 68.Kg2 a4 69.Kf2 d2 70.Bxd2 Nxd2–+

65.Kg3 h2 66.Kxh2 Nf3+ 67.Kg3 d2 68.Bxd2 Nxd2 69.f5 Nc4 70.Kf4 Nxa5

Finally taking that pawn!

71.Ke4 Nc6 72.Kd5 a5

And White resigned.

0–1

I was elated after the game, as I had managed to play the game in the spirit of my chess idol Tigran Petrosian and, in
a way, to imitate my opponent who was also a former student of my idol. This game illustrates why the ability to
understand and play endgames well are the core qualities that need to be developed by a student of the game of
chess.

124
 GAME 5
Gata Kamsky
Joël Lautier
New York 1991
This game was chosen for the tactical theme at the end. My opponent was one of the best French players at the time,
the young and talented GM, Joël Lautier.

1.e4 e6 2.d4 d5 3.Nc3

My mainstay system versus the French Defense. I was taught it when I first started playing chess at the Pioneer
Chess Palace in Leningrad in Russia. The system has served me well for many years, even though I have
experimented wildly with different sub-variations. Despite all the tactical and theoretical expertise required to play it
successfully, at the core it is a very strategic opening.

3...Nf6 4.e5 Nfd7 5.f4 c5 6.Nf3 Nc6 7.Be3 a6

One of the oldest main lines. The other old line in fashion in those days was the piece sacrifice after 7...Qb6 8.Na4,
but this is not a book on theory, so we will skip it.

8.Qd2 b5

125
9.dxc5

As mentioned earlier, this was an experimental move to see what happens if White simply exchanges his central
pawn. More challenging moves are of course 9.Bd3 and, somewhat surprisingly, 9.a3, both of which lead to heavy
theory. At the time this game was played, modern theory was not as developed as today, since there were no engines
to analyze with. A time of Wild West chess!

9...b4

Black wants to push the c3-knight before White develops his f1-bishop, attempting to prevent the Nc3–e2–d4
transfer. Theory at the time continued with the following line: 9...Bxc5. The most logical move for Black. It seems,
however, that my opponent wanted to do some experimenting of his own! 10.Ne2 0-0 11.Bxc5 Nxc5 12.Ned4 Ne4
13.Qe3 Nxd4 14.Nxd4 f6 15.exf6 Qxf6 16.g3 e5 17.fxe5 Qxe5 18.0-0-0 Bd7= with a roughly balanced game.

10.Nd1?!

I had this move in mind when I played dxc5. White is focusing on regrouping his minor pieces to a more aggressive
set-up aimed at the black kingside.
10.Na4! is the strongest move, giving White the better position. 10...Qa5 11.Nb6 Nxb6 12.cxb6 Bc5 13.Bxc5 Qxc5
14.Qf2 Qxf2+ 15.Kxf2 Rb8

126
Considered a good position for Black, however I disagree with this assessment. 16.Be2 Rxb6 17.c4 Black faces a
long-term defense in an inferior endgame because of White’s superior light-squared bishop. 17...bxc3 18.bxc3 Kd8
19.Rab1 Rxb1 20.Rxb1 Kc7 21.c4

21...dxc4
[21...d4?! 22.Bd3 h6 23.Be4 Rd8 24.Ke2 f5 25.Bxc6 Kxc6 26.Kd3 Bb7 27.c5! White’s advantage is threatening to
assume larger proportions, thanks to his initiative on the queenside. 27...Ba8 28.Rc1 Kc7 29.Nxd4

127
29...Bd5 (29...Bxg2? 30.Rg1 Be4+ 31.Kc3+–) 30.Kc3 g5 31.g3 Kd7 32.Kb4 Kc7 (32...Rb8+ 33.Ka5 Rb2 34.c6+
Kc7 35.Kxa6 Rxa2+ 36.Kb5 Rxh2 37.Kc5 Rd2 38.Ra1+–) 33.a4± and White is simply a pawn up with superior
piece activity.]
22.Rc1 Rd8 23.Ke3

23...h6 (23...Bb7 24.Ng5±) 24.Nd2 Na5 (24...g5 25.fxg5 hxg5 26.Nxc4±) 25.Bxc4 Kb6 26.Rb1+ Ka7 27.Bd3 Bb7
28.Rc1 Kb6 29.Ne4 With a small advantage for White, which a player like Magnus Carlsen would convert without a
doubt.

10...Nxc5?!

128
10...Bxc5 11.Bxc5 Nxc5 12.Nd4 Nxd4 13.Qxd4 Qe7! 14.Ne3 0-0 15.Be2 a5 16.0-0 Rd8! and all three results are
possible.

11.Bd3

11.Nf2!? seemed a more natural move, however I wanted to keep the knight on d1, with the possibility of going
Nd1-e3. 11...Bb7 12.Be2 Be7 13.0-0 0-0 14.Rac1 Ne4! and Black seems to have equalized.

11...Bb7 12.0-0 d4?!

An inaccuracy which does not make Black’s position worse, but gives White more opportunities, for example he
now has a new square for his d1-knight on c4.
12...Be7

This would be preferable, to see how White is going to solve the problem of his d1-knight’s further development.
13.Nd4 The most natural move. White needs to keep a blockading piece on the d4-outpost, keeping Black’s
129
potentially powerful b7-bishop closed in.
a) 13.Bf2? Ne4 14.Qe2 Nxf2 15.Nxf2 Qb6

Here Black has managed to exchange White’s key piece, the dark-squared bishop, and possibly equalized, although
White still has significant chances for a kingside attack.
b) 13.Nf2?! is possible, but leads to an easier game for Black: 13...d4 (13...0-0 is also perfectly fine. If White plays
the natural 14.Rad1, Black replies 14...d4! 15.Bxd4 Nxd4 16.Nxd4 Qd5 17.Nf3 Qxa2 18.Qxb4 Bxf3 19.gxf3 Rfb8
20.Ra1 Qxb2 21.Qxb2 Rxb2µ with the superior endgame.) 14.Bxd4 (14.Nxd4? Nxd4 15.Bxd4 Qd5

Suddenly Black is just winning. The d4-bishop will be taken with check, leaving White no time for the return
discovered check Bd3-b5+.) 14...Nxd4 15.Nxd4 Nxd3 16.Qxd3 Qd5 17.Nf3 Rc8

130
Black clearly has full compensation for the pawn as he has two superior bishops, while White’s extra c2-pawn is
really weak.
13...0-0 14.Rf3 (The rook lift looks critical, although 14.Nf2 is objectively safer: 14...Nxd4 15.Bxd4 Rc8 16.Rac1
Nxd3 17.cxd3 Qa5 18.b3=.) 14...Qc7
[14...Ne4 15.Qe1 Qc7 16.Nf2 Nxd4 17.Bxd4 Nc5 18.Ng4

I’m pretty sure this was the set-up I had in mind. White has managed to transfer his knight to the g4-square where it
guards against the ...f7–f6 advance and threatens a potential Ng4–f6 sacrifice, opening up the black king’s defenses.
While the engine evaluates this position as equal, it still feels more comfortable for White, since the black b7-bishop
is not doing anything. 18...Nxd3 19.cxd3 Bc5 20.Qe3 Rfc8 21.f5 Bxd4 22.Qxd4 Qc5 23.Qxc5 Rxc5 24.d4 Rc2
25.Rf2 Rxf2 26.Kxf2 exf5 27.Ne3 Rc8 28.Nxf5

131
With a double-edged endgame, for example 28...Rc2+ 29.Ke3 g6 30.Nd6 Bc6 31.Rf1 Rxg2 32.Rxf7 Rxb2 33.Rc7
White has the more active rook and superior knight, which gives him enough compensation for the sacrificed pawns.
33...Ba4 34.Nc8 Rxa2 35.Ne7+

35...Kh8 (35...Kf8 36.Nxd5 b3; 35...Kg7 36.Nxd5+) 36.Nxd5 b3 37.h4 Rh2 38.Nf6 Rxh4 39.Rb7 Bb5 40.e6 b2
41.Rxb5 axb5 42.e7 b1=Q 43.e8=Q+ Kg7 44.Qe7+ Kh6 45.Ng8+ Kh5 46.Nf6+ with a perpetual.]
15.Nf2 Rfc8

132
16.Rc1 a5 17.c3 Nxd3 18.Nxd3 bxc3 19.bxc3 Ba6 20.Rg3 Bf8!³ and it seems unlikely that White’s kingside attack
will be successful.

13.Bf2 Be7

14.b3!

The exclamation mark is for adapting to changed circumstances. It is not often that the white queen’s knight heads to
the c4-square in these types of French Defense positions.

14...f6?!

A clear concession. Black was worried about potential white access to the d6-square. He also wants to take control
of the center, however he seems overextended and his central pawns are weak. Therefore such a move, opening
things up, was premature.
133
14...0-0 was preferable. 15.Nb2

15...Qc7 (15...Na7 is an extremely interesting dynamic suggestion by the computer. Black threatens ...Bf3 and White
cannot take on d4 with his bishop because of the vulnerability of his king. He is therefore forced to capture on d4 in
unfavourable circumstances, opening up Black’s b7-bishop. 16.Nxd4 Nxd3 17.Qxd3 Nb5 18.Nc4 Qd5 19.Nf3 Qxd3
20.cxd3 Bxf3 21.gxf3 Nc3 22.Kg2 Rfd8 23.d4 Nd5

and Black should hold this endgame because of his better pawn structure and superior d5-knight which carries out a
few useful functions all by itself.) 16.Nc4 Rad8 17.Rae1 Na5 18.Bh4 Nxc4 19.Bxc4 Bxh4 20.Nxh4 Qb6

134
21.Kh1 (21.f5 d3 with counterplay) 21...Ne4 22.Qd3 Qc6 23.f5 White retains some aggressive possibilities on the
kingside.

15.exf6 gxf6 16.Nb2±

At this point White is much better, thanks to Black’s overextended pawns in the center and on the queenside.
Perhaps Black counted on his strong b7-bishop, coupled with the potential pressure on the g-file, but first he has to
solve the problem of his king in the center.

16...Qd7?!

16...Rg8 seems better and more logical. Black forces White to adopt a less comfortable deployment of his pieces.
17.Qe2 (17.Rae1 f5 18.Bg3 Rg6 19.Nc4 Kf8 20.Re2 a5 21.Rfe1 Kg7 22.Ng5 Bxg5 23.fxg5 Qxg5 24.Qxg5 Rxg5
25.Bd6 Nxd3 26.cxd3 Rg6 27.Rxe6 Rxe6 28.Rxe6 a4 29.Nb6 axb3! with an equal endgame) 17...f5 18.Bg3

135
18...Rg6 (18...Qc7? 19.Ng5‚) 19.Nc4 (19.Ng5 Nxd3 20.Nxd3 Qd5 21.Rfe1 Kd7 22.Rad1 h6 23.Nf3 Kc7 with a lot
of fight left in the position as Black is finally consolidating. His control over the a8-h1 diagonal will start to
influence the game more.) 19...Qc7 20.Ng5 0-0-0 21.Rae1 Kb8 22.Qh5 Rh8 23.Rf2 Bf8 24.Rfe2

White is still better, though the position is highly unbalanced in terms of pawn structure, and anything can happen
with kings placed on opposite flanks.

17.Nc4

17.Na4!± A computer move that did not occur to me. White’s a4-knight is powerfully placed on the side of the board
as it is ready to pounce on Black’s weak central pawns and at the same time prevents Black from castling long in a
view of possible Na4–b6 checks.

17...Rd8?! 18.Rae1

18.f5! is quite strong as well, however I did not want to release the tension so early. 18...e5 19.Qh6 Nxd3 20.Qh5+
Kf8 21.cxd3 Rg8 22.Qh6+ Ke8 23.Qxh7 Qd5 24.Bg3 Qf7 25.Qh6
136
White keeps an extra pawn, and has the solid plan of further improving his pieces on the central squares.

18...0-0 19.Bg3

Black finally castles kingside, and White has to hurry before Black completes his ...Kh8, ...Rg8–g7 regrouping
which would allow him to double on the g-file, creating counterplay. After some thought I realized that I must
transfer my queen to a more active position on the h5-square.

19...Kh8?!

19...Qe8 20.Qe2 f5

137
deserved attention, with the idea of transferring Black’s queen to h5 first, before White can accomplish the same
task, while keeping the rook on the d8-square to protect the far-advanced d4-pawn. 21.Bh4 Bxh4 22.Nxh4 Ne7
(22...Kh8 is bad due to 23.Na5! Nxa5 24.Qe5+ Kg8 25.Qxc5+– and Black is unlikely to be able to defend all of his
weak pawns.) 23.Qd2

White keeps the pressure on Black in different locations all over the board. Black has succeeded in sealing off
White’s d3-bishop from attacking on the kingside, but White has gained access to the central e5-square for his pieces
and the black queenside pawns are on dark squares and have become vulnerable. 23...Qb5 24.Ne5 Qb6 25.Qe2

138
White switches directions once again to the kingside. 25...Nd5 26.Qh5 The author of the book My System, GM
Nimzowitsch, would be happy with the way White has finally achieved his goal of placing his queen on h5, from
where it will exert strong pressure on the black kingside with a potential attack coming shortly. 26...Ne3 27.Qg5+
Kh8 28.Qh6 Kg8 29.Nc4

29...Qa7 The only move, as others lose. (29...Nxc4 30.Bxc4 Be4 31.Rf3 Bxf3 32.Rxe6+–) 30.Nxe3 dxe3 31.Bc4
Bd5 32.Bxd5 Rxd5 33.Rxe3 Rf7 34.Rg3+ Rg7 35.Rxg7+ Qxg7 36.Qxg7+ Kxg7 37.Nf3 Ne4 38.Ra1 Kf6 39.g3 Rc5

139
White has a lot of work ahead of him. It seems highly unlikely that he will be able to convert his extra c2-pawn into
a win.

20.Qe2 Rde8

21.Nh4?!

My turn to hesitate.
21.f5 immediately was better, as White transposes into the structure I wanted to achieve. 21...e5 (21...exf5 22.Nh4
leaves Black with a completely ruined kingside pawn structure for no compensation) 22.Ncxe5 The idea that I
missed during the game. White smashes his way through to the black king by brute force. 22...fxe5 23.Nxe5

140
23...Qd8 (23...Nxe5 24.Bxe5+ wins on the spot) 24.Ng6+!! Very flashy and strong. Black must suffer heavy material
losses to avoid immediate mate. (24.f6 Bxf6 25.Nf7+! Rxf7 26.Qxe8+ Kg7 27.Bc4 Qxe8 28.Rxe8 d3 29.cxd3 Bd4+
30.Bf2 Rxf2 31.Rxf2 is another win, although a slower one.) 24...Kg7 (24...hxg6 25.fxg6 Rxf1+ 26.Rxf1 Qd5
27.Bc4 Qg5 28.Rf7 Qxg6 29.Be5+ Nxe5 30.Qxe5++–) 25.Qg4 Nxd3 26.Nxe7+ Kh8 27.f6

27...Rf7 (27...Rg8 28.Nxg8 Rxg8 29.f7 Rxg4 30.Re8++–) 28.cxd3+–

21...Nxd3 22.cxd3 Bd8 23.Qh5 Kg8

141
24.f5!

Finally!

24...e5 25.Re4

White has lost his advantage of the bishop pair, and needs to reposition his pieces before Black consolidates and
creates problems for him. One of the ideas I had was to put my h4-knight back to work on its ideal square, e4. It will
take time, but when it happens, White will simply be winning because Black cannot protect all the holes in his camp
on the c5-, d6- and e6-squares which, taken together with his open king and terrible knight on c6, will doom him to
oblivion.

25...Rf7 26.Nf3?

An obvious mistake. I was so determined to bring my h4-knight to e4 that I refused to take other candidate moves or
lines into consideration. It could have cost me dearly.
26.Be1!!

142
I missed this idea completely, even though White activates his sleeping g3-bishop and can potentially transfer it to
the h6-square. 26...Ref8 27.Rg4+ Rg7 28.Rxg7+ Kxg7 29.Bd2 Kg8 30.Rf3 Rf7 31.Bh6 with a winning attack.

26...Ref8?! 27.Rg4+ Rg7 28.Rxg7+

28...Kxg7?

Black had to take with his queen.

28...Qxg7 29.Nd6 (29.Nfd2 Na7!

143
Finally getting that knight to do some real work, redirecting it to a great outpost on the c3-square. 30.Ne4 Bxe4
31.dxe4 Nb5 32.Be1 Nc3 33.Bd2 Kh8 34.Bh6 Qf7 35.Qf3 Rg8 Black has successfully solved all his problems. The
game becomes wide open and anything can happen.) 29...Ba8 30.Rc1 Na7 31.Be1!

White stands better anyway. If Black plays ...a5, then he won’t be able to play ...Nb5.

29.Nfd2 Be7

a) 29...Ne7? 30.Nxe5!+–
b) 29...Na7? is too late because of 30.Nxe5! fxe5 31.Bxe5+ and White is winning, e.g.

144
31...Rf6 32.Qg5+ Kf7 33.Qh4 Qe7 34.Bxf6 Qxf6 35.Qxh7+ Qg7 36.Qh5+ Kf8 37.Nf3 Bf6 38.Re1 Qf7 39.Qh6+
Qg7 40.Qf4+–.

30.Ne4

30...Kh8?

The final mistake of the game. Black misses the breakthrough tactics.
30...Qe8, trying to the chase White’s queen from its optimal position, was Black’s last chance, though even here
Black seems to be lost. 31.Qg4+ Kh8 32.Bh4 Ba8 33.Rf3 Rg8 34.Qh3

145
34...Qd8 (34...Qf8 35.Nb6+– The a4-knight joins the game with decisive effect by threatening to play either Nb6xa8
or Nb6–d5xf6.) 35.Ncd6 Rg7 36.Bxf6 (It’s never too late to make a mistake: 36.Rg3? Rxg3 37.Qxg3 Qf8 with
defensive resources.) 36...Bxf6 37.Nxf6 Qxf6 38.Ne8 Qg5 39.Nxg7

39...Nd8 (39...Qxg7 40.f6 Qf8 41.f7 Ne7 42.Qe6 Ng6 43.Qf6+ Qg7 44.f8=Q+ winning) 40.Nh5 Bxf3 41.Qxf3 And
the position is won.
White has better pieces and a material advantage with no counter-chances for Black.

31.Qh6+–

146
White gets to the Achilles heel of Black’s position, the f6-pawn.

31...Rf7 32.Bh4

Everything is ready for the final blow. Black sets a last trap.

32...Nd8

32...Qd8 33.Ncd6 Bxd6 34.Bxf6+ Rxf6 35.Nxf6 Qe7

36.Rf3 Qg7 (36...e4 37.Rh3+–) 37.Rg3 with mate to follow.

33.Ncd6!

White wins material.

33.Nxf6?? This careless move allows 33...Qc6 and suddenly it is Black who is winning.

147
33...Bxd6 34.Bxf6+

34...Kg8

34...Rxf6 35.Qxf6+ Kg8 36.Rf3+–

35.Rf3 Rg7 36.Bxg7 Bxe4 37.Rg3 Bxf5 38.Bxe5+

Black resigns.

1–0

148
 GAME 6
Gata Kamsky
Garry Kasparov
Tilburg 1991
This was the first game of my minimatch versus Kasparov. Having lost really badly to him earlier the same year at
the Linares super-tournament didn’t inspire confidence, but this time I was assisted by a Russian coach hired by my
father to help me with my terrible opening preparation. The coach’s name was name Polovodin, and he was to be
working with me until the end of the year.
Of course choosing to play 1.e4 against Kasparov’s pet Sicilian Scheveningen, which actually made Kasparov
World Champion in the 24th and last game of his match vs. Karpov, was quite a gamble, but we felt that this choice
would be unexpected, as I was playing the London System on a regular basis and would want to avoid my feared
opponent’s preparation.

1.e4 c5 2.Nf3 e6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 d6 6.Be2 a6 7.a4 Nc6 8.Be3 Be7 9.0-0 0-0 10.f4 Qc7 11.Kh1 Re8
12.Bf3 Rb8

This is a tabiya position of the Scheveningen Variation of the Sicilian Defense. So far, everything is well known.
Black prepares his ...b7–b5 push and guards against possible sacrificial breakthroughs by White in the center with
e4–e5. During my preparation with my coach, we had looked at this position and decided against the plan of a
kingside attack. The decision was also influenced by the fact that I didn’t want to lose my first game of the mini-
match with White, therefore trying to play solid chess was the choice.

13.Qd2 Bd7 14.Rad1

14.Nb3 would be White’s usual course of action. 14...b6 15.g4 Bc8 16.g5 Nd7∞

14...Na5

14...b5? 15.axb5 Nxd4

149
16.b6! An important intermezzo.16...Rxb6 17.Bxd4 Suddenly there’s no defense against e4–e5, catching Black’s f6-
knight off guard. 17...Rxb2? 18.e5+– White wins by tactics.

15.b3!

The exclamation mark is for the audacity of White’s idea.

Usually it’s Black who plays on the queenside, and any structurally-weakening moves by White on that flank would
favor Black. However timing is on White’s side as Black’s d7-bishop is not optimally placed, and this gives White
time to regroup.
15.Qf2?! Standard, but too passive. 15...Nc4 16.Bc1 b5 (16...e5!?=) 17.axb5 axb5 18.b3 Na5ƒ with a complex game
ahead.

15...Rec8

The correct rook goes to the c-file.

150
15...Rbc8?! This looks more logical, leaving Black’s other rook prepared for the fight in the center, however it is not
correct. 16.Nde2 b5?

17.e5! The point! Black cannot capture on e5 twice, because he loses his d7-bishop, as the bishop’s defender, the f6-
knight, is eliminated. (There is no point including 17.axb5?!, as that would allow 17...Bxb5, getting out of tactical
danger.) 17...dxe5 18.fxe5 Qxe5 19.Bd4 Qc7 20.Bxf6 Bxf6 21.Qxd7 Bxc3 22.Nxc3

22...Qxd7 (22...Qxc3 23.Bh5! Kh8 24.Bxf7 Rf8 25.Bxe6+–) 23.Rxd7 Rxc3 24.axb5 axb5 25.Ra7 Nc6 26.Bxc6 Rxc6
27.Rfxf7 with a winning four-rook endgame.

16.Nde2 Be8

The idea is to hold White’s pieces along the c-file, while Black prepares the ...b7–b5 push.
16...Nc6 was more accurate, following the rule that any piece which is not functioning well should be repositioned.

151
17.g4 (17.Na2 b5 18.axb5 axb5 19.g4 Ne8 20.Nac1 Nd8!? 21.c3ƒ and Black still has some problems to solve, since
his standard queenside counterplay has not resulted in any concrete achievements.) 17...b5 18.g5 Ne8 19.h4 bxa4
20.Nxa4 Nb4 21.c3 Nc6 22.Rb1 Na5 23.Qa2

But this is certainly what Black wants, to open the position up with White’s king starting to feel insecure.

17.Ba7!

a) 17.e5?µ dxe5 18.fxe5 Nd7 19.Qe1 Nc6! and Black’s compact pieces are getting dangerously close to springing
out in spectacular fashion.
b) 17.g4?! b5 18.axb5 axb5 19.g5 Nd7³

17...Ra8 18.Bd4

152
I was very uncomfortable with White’s position, as I felt that it was slipping from my grasp. I also adamantly
refused to go with the attacking g4-plan, knowing full well my opponent’s feeling for nuances in the positions
resulting from such a structure. Hence I decided to start looking for a draw. I would have been perfectly happy with
the repetition after 18...Rab8 followed by 19.Ba7.

18...Nc6!

No draw! My opponent feels my hesitation and feeds off it to try to win.

18...Nd7? 19.f5±

19.Bxf6!

Objectively not the best move, but I wanted to eliminate queens against an opponent famous for his middlegame
skills. I felt much safer in the harbor of the endgame.

153
19...Bxf6 20.Qxd6 Qb6?!

And now Black gambles! It is obvious that Black equalizes easily after the queen exchange, but at the same time,
Black’s chances of a win diminish drastically as well. For example:
20...Qxd6 21.Rxd6

21...g5?! Quite a standard pawn break in such positions with the intention of giving Black’s darksquared f6-bishop
more space to operate. The move also attempts to lull White into playing the simple defensive move g2–g3, however
White has a spectacular refutation.
a) 21...Bxc3 22.Nxc3 Nb4 (22...Ne5 23.Nd5!±) 23.Ne2 Nxc2 (23...Rxc2 24.Rb6²) 24.Rc1 Nb4 25.Rxc8 Rxc8
26.Rb6 (26.Kg1 b5 27.axb5 Bxb5 28.Nd4 Nd3 29.g3 h6 with equality) 26...Nd3 27.Kg1 Bc6=
b) 21...Be7 This attempt to keep both bishops on the board is risky. If White manages to regroup and keep his extra
pawn, he will have the superior chances. 22.Rd2 Nb4 23.e5 Rc7

24.Rc1! (24.Ne4!? Rxc2 25.Rxc2 Nxc2 26.Nd6 Bxd6 27.exd6 Rd8 28.Bxb7 Rxd6 29.Rc1 Nb4=) 24...Rac8
(24...Bc5 25.Nd1±) 25.Ne4 Rxc2 26.Rdxc2 Rxc2 27.Rxc2 Nxc2 28.Nd6 Bxd6 29.exd6
154
29...b5! (29...b6 30.Bb7! — attacking the a6-pawn but also threatening Bc8 — 30...Kf8 31.Bxa6 Bc6 32.Kg1 g6
33.Bd3 Nb4 34.Nd4 Bxa4 35.Bxg6 hxg6 36.bxa4 Nd3 37.Nc6 Ke8 38.Ne5!² White keeps the advantage with hopes
of a win.) 30.a5 The simplicity of this position is deceptive. Every move counts and any mistake would be very
costly. Both sides must be careful, as White’s far advanced d6-pawn can be both an asset and a weakness. The game
is hanging in the balance. 30...Kf8 31.Be4 Nb4 32.Nd4 Bd7 33.g4 h6 34.Kg2

34...g5 ( 34...f6?? 35.f5±) 35.Kg3 gxf4+ ( 35...Ke8 36.h4 gxf4+ 37.Kxf4 f6 38.Ke3 Kd8 39.Nf3 Bc6 40.g5 fxg5
41.hxg5 hxg5 42.Ne5 Bd5 43.Kd4 Bxe4 44.Kxe4 and neither side can make progress) 36.Kxf4 f6 37.Ke3=
22.e5! (22.fxg5 Be5! 23.Rd2 Nb4³; 22.g3 gxf4 23.gxf4 Bxc3 24.Nxc3 Nb4 25.Ne2 Rxc2 26.Rb6 a5 27.Rxb7 Rd8=)
22...Be7

155
23.Ne4!! The move that I missed during the game. (23.Rdd1? gxf4 24.Nxf4 Rab8∞) 23...Bxd6
[23...gxf4 24.Nf6+! Kf8 25.Nxe8 Rxe8 (25...Nxe5 26.Bxb7 Kxe8 27.Rd4 Rxc2 28.Bxa8 Rxe2 29.Rdxf4±) 26.Bxc6

26...Bxd6 (26...bxc6 27.Rd4 with a winning game for White due to the completely ruined black pawn structure)
27.Bxe8 Bxe5 28.Bxf7 Kxf7 29.Nxf4 Ke7 30.Nd3 with a healthy extra pawn for White and very good winning
chances.]
24.exd6

156
24...gxf4! Pawns are more important than pieces in the endgame! (24...Rd8 25.fxg5 Rac8 26.Nf6+ Kf8 27.Rd1 Ne5
28.Be4 Kg7 29.Nxe8+ Rxe8 30.d7+–) 25.Nxf4! (25.d7 Bxd7 26.Nf6+ Kg7 27.Nxd7 Rc7 28.Nc5 Ne5 29.Bxb7 Ra7
30.Nxe6+ fxe6 31.Be4 f3 32.gxf3 Rf7 with a very exciting battle ahead, where both sides have equal chances to win)
25...Rd8 (25...Kg7 26.Nh5+±) 26.Nf6+ Kg7 27.N4h5+ Kh8 28.Rd1 Bd7 29.Bxc6 Bxc6 30.d7 b5 31.Rd6! Bd5 32.a5
Ra7 33.Kg1 Rc7 34.Rxa6 Rxc2 35.Nxd5 exd5 36.Rd6²

White manages to create some rather unpleasant threats for Black. 36...d4 37.g4! A very strong move, making sure
that the white f6-knight is supported. 37...Kg8 (37...Rc7 38.Nf6) 38.Rxd4 Kf8 39.g5 Ke7 40.Nf6±

21.e5

157
Immediately taking the chance to close down Black’s monster f6-bishop.

21...Be7

21...Rd8 This was another possibility, but it seems that my opponent wanted to leave his rook on the c-file to keep
the potential pressure on my weak c2-pawn. 22.Qa3 Be7 23.Qc1 Nb4 24.Ne4

24...Rac8 (24...Bc6 25.h3²) 25.Rxd8 Rxd8 26.Rd1 Rxd1+ 27.Qxd1 Qe3 28.Qd2! Qxd2 29.Nxd2 b5 30.axb5 axb5
31.c3 Nd3 32.Be4 Nf2+ 33.Kg1 Bc5 34.Kf1 Be3 35.g3! Despite massive exchanges, White has managed to keep his
extra pawn and Black must seek equality.

22.Qd2 Bb4

22...Nb4?! would allow White to regroup with advantage: 23.Nd4! Rd8 24.Nce2±.

23.Qd3!?

158
At this point I was getting quite optimistic about not losing the game!
23.Qc1

I wasn’t too thrilled about playing such a passive move, however objectively it is much stronger as White simply
wants to play Ne4 and plant the knight on d6, keeping an extra pawn with the advantage. 23...Ne7
(23...Nxe5 24.Nd5! exd5 25.fxe5 Bc6 26.Bxd5 Bxd5 27.Rxd5 with a serious advantage for White) 24.Rd3!
An important intermezzo, to first take control of the weak e3-square.
24...Qc7 25.Rfd1

25...Rab8 (25...Ba5 26.Qa3 Nf5 27.Be4 Bc6 28.Bxf5 exf5 29.b4 Bb6 30.Qb3 Rd8 31.a5 Ba7 32.Nd5 Bxd5 33.Rxd5
Rxd5 34.Qxd5 White has finally managed to stabilize his advantage. 34...Qxc2 35.Qd8+) 26.Na2 Bc5 27.c4 Ba7
28.Qd2 b5 29.cxb5 axb5 30.Rc3 Qb6 31.a5 Qf2 32.Rxc8 Rxc8 33.Nac3 Nf5 34.Ne4 Qh4 35.g3 Qe7 36.Nd6 and
White has consolidated with an advantage.

23...Rd8 24.Qe4!
159
Temporarily pinning Black’s c6-knight to the defense of his b7-pawn. White’s c3-knight is finally ready for
regrouping.

24...Qc5!?

A psychological ploy by Black. I suspect that my opponent knew that for me to play a move like Nb1 would go
against my every instinct as a player. Objectively though 25.Nb1 gives White a huge advantage, as Black’s b4-
bishop is denied the vital d2-square and White has the potential to regroup the b1-knight further to either the c4- or
e4-square, but during the game I just couldn’t bring myself to play this move against the World Champion!

25.Na2?

25.Nb1!± Rxd1 26.Rxd1 Qf2

27.c3! (27.Nbc3!?±) 27...Be7 28.Nd2 Rd8 29.Qc2 f6 30.exf6 gxf6 31.Ne4 Qe3 32.b4 Bg6 33.Rxd8+ Nxd8 34.a5
Nf7 35.Qd2 with a position that is close to winning.

160
25...Bd2! 26.c3!

Thankfully I found in myself the strength to make this pawn move, taking control of the dark squares in the center.
Now I felt positive that the game would not be lost.

26...a5! 27.Nac1!?

I was just happy that I could bring that knight back into the game!
27.Qc2 was simpler. 27...Be3 (27...Qe3? would have been a serious mistake because of 28.Ng3!+–) 28.Be4!

An important intermediate move to protect White’s queen on c2 against possible ...Nb4 threats. 28...h6 (28...g6
29.Nac1 Rxd1 30.Rxd1 Rc8 31.Nd3 Qa7 32.Ne1! Nb4 33.Qb1 Na6 34.Nc2 b5 35.axb5 Bxb5 36.Ned4 Bxd4
37.Nxd4 and White’s position is finally stable with an extra pawn and the advantage.)

27...Bxc1!

A very practical choice by Black, otherwise Na2 by White would end the career of Black’s bishop on d2.
161
27...Qe3 28.Qxe3 Bxe3 29.Nd3±

28.Rxc1

28...Rac8

28...Rd2 was tempting, but White has everything under control. 29.Rfd1 Rad8 30.Rxd2 Rxd2 31.h3 Qb6 32.b4!
axb4 33.cxb4 Qxb4 34.Qxb4 Nxb4 35.Rc8 Kf8 36.Nc3 Ke7 37.Rc7+ Rd7 38.Rxb7 Rxb7 39.Bxb7 and Black has a
long fight ahead for a draw.

29.Ng3?

At this point I was running low on time and started to make rushed decisions. It is amazing how Black kept
exploiting my hesitation to make small improving moves.
The big problem for White in this structure is the queenside pawn formation. He desperately needs to move it
without weakening the dark squares, therefore it was imperative to go for a b3–b4 push!! The move in the game
spoiled a large amount of my earlier work.
a) 29.Rfd1 Qb6! 30.Qb1 Qe3²
b) 29.Qb1!? b5 (29...Qe3 30.Ng3±) 30.axb5 Qxb5 31.Rcd1 Rxd1 32.Rxd1 h6 33.h3 Ne7 34.Nd4 Qc5 35.Rd3 and
White still has to work really hard to make use of his extra pawn.
c) 29.Rb1! Simplest and strongest. 29...b5 30.b4 Qb6 31.axb5 Qxb5 32.c4 Qb6 33.c5

162
33...Qb8 (33...Qa7 34.b5±) 34.bxa5! (34.b5 Nb4∞) 34...Qa7 35.Qe3 Qxa5 36.Rfc1 and the formerly weak c2-pawn
is now officially a monster. White has serious chances of success.

29...Nb4!„

I have to give credit where it is due: Black has finally managed to make White commit an error and now seizes the
opportunity for activity. White has to be careful.
29...Nd4? 30.Qxb7!+–

30.Rcd1!

A good practical decision. Short of time, I decided to go for mass exchanges and prevent further black activity by
opposing Black’s control of the d-file. Furthermore I’m ready to play f4–f5 and create some threats of my own.
a) 30.Qxb7 Nd3 31.Rc2 Nxf4∞
b) 30.Rfd1 Rxd1+ 31.Rxd1 Nd5! 32.Ne2 Qb6∞
163
30...Qxc3?

A rushed move caused by excitement. Black forgets about White’s active possibilities.
30...Rxd1 31.Bxd1! Qxc3 (31...Nd5? 32.c4±) 32.Qxb7 Nd5=

Black’s activity is sufficient compensation for the missing pawn. The position is dynamically equal.

31.Qxb7?

And White makes a mistake in the move order.

31.Rxd8 Rxd8 32.Qxb7 transposes to the correct position in the game.

31...Qxb3?

It seems we both underestimated White’s resources in time trouble.

Correct was 31...Nd3 32.Ne4

164
32...Qd4!! The only move to keep the game in a delicate balance, although it’s probably impossible to find in time
trouble, even for the World Champion. 33.Qa6 Black still has to navigate carefully through turbulent complications,
as some lines below will show. 33...Rc2! (33...Bc6? 34.Nd6 Bxf3 35.Rxd3 The point of 33.Qa6.) 34.h3 Bc6
35.Qxa5 Bxe4 36.Bxe4 Nf2+ 37.Kh2 Nxd1 38.Bxc2 Ne3 39.Rf2

39...g6! Very impressive. Black is two pawns down, but his superior pieces, especially the centralized queen and
knight, are extremely dangerous. It is White, in fact, who has to be careful in this position. (39...Nxc2? 40.Rxc2
Qxf4+ 41.g3 Qd4 42.Qc3 with advantage to White) 40.Qe1 Rc8 41.Bb1=

32.Rxd8 Rxd8 33.Qc7! Rd4!

165
34.Ne4!

Best.
34.f5?! is a very interesting alternative, although it is objectively worse than the game move. 34...Qxa4
[34...exf5 35.Nxf5! (35.Qb8? Qxa4 36.Nxf5 Qb5!µ) 35...Qd3 36.Ng3 with counterplay. Black has to be careful.]
35.fxe6 (35.f6!? also deserves attention) 35...fxe6 36.Bh5 Bxh5! (36...Qd7? 37.Qc5+–) 37.Nxh5 Qd7 38.Qxa5 h6=

34...Qd3 35.Rg1?

A big mistake that spoils all the previous effort in the game.
35.Re1! looks natural, but it was not easy to see how White can achieve any advantage. 35...h6

36.f5!± An absolutely fantastic resource! Suddenly the white e4-knight is going, NOT to d6, but to f6!
a) 36.Qxa5? Bc6–+ Black just wins on the spot.
166
b) 36.Nd6?! Rxf4 (36...Qd2!?) 37.Nxe8 Rxf3 is a draw: 38.gxf3 Qxf3+ 39.Kg1 Qg4+=.
36...Bc6! Black’s only defense, however White still keeps an advantage. (36...Nd5 37.Qd8; 36...exf5?

37.Nf6+!! gxf6 38.exf6 immediately winning the game, e.g. 38...Bxa4 39.Qb8+ Kh7 40.Qf8+–) 37.Nf6+ gxf6
38.fxe6 fxe6

39.exf6 Qg6 (39...Rd7 40.Qc8+ Kh7 41.Bxc6+–) 40.Bxc6 Nxc6 41.Qxc6 Qxf6 42.h3 After this quiet move giving
the white king an escape door, Black’s position is extremely uncomfortable. 42...Rf4 43.Kh2±

167
35...Rxe4?!

Yet another mistake in this wild game. It seems Black wanted to avoid White’s possible Nd6 at all costs.
35...h6!? 36.Nd6 Qe3! (36...Rxf4? 37.Nxe8 Rxf3 38.Nf6+!+–) 37.h3! (37.Qc8? Nd3 38.Qxe8+ Kh7 39.Be4+ Rxe4
40.Nxe4 Qxe4 and Black is just winning: 41.Rf1 Qe2! 42.Rb1 Ne1!–+) 37...Nd3 38.Rf1 Nf2+ 39.Rxf2 Qxf2 40.Qc8
Rxf4 41.Qxe8+ Kh7 42.Qb5 (42.Nxf7 Rxf3 43.gxf3 Qxf3+ resulting in a fair draw) 42...g6=

36.Qc8 Kf8 37.Qc5+ Kg8 38.Qe7!

After some thought, I decided not to repeat the position.

I realized that a position with an extra exchange would be extremely hard to convert, but at least I would not be
risking a loss and could always make a draw from a position of strength. This would be of great psychological value
for the next game!

168
38...g6 39.Qxe8+ Kg7 40.Bxe4 Qxe4 41.Rf1?!

A final inaccuracy, after which the game ends in a deserved draw. The only chance to play for a win was 41.Qd8,
with the idea of transferring the queen to the defense of the white king. Let’s take a took...
41.Qd8!!

41...Qxf4 (41...Nd3 is objectively better, although White is better after 42.Qh4 thanks to 42...Nxf4 43.Qf2! with the
idea of Rf1.) 42.Qf6+! Strongest. Of course not Qxa5 because of ...Nd3.
[42.h3!?² If Qf6+ wasn’t winning, this move — which I missed — would have been an amazing resource. Black
cannot capture White’s e5-pawn. 42...h5! (42...Qxe5? 43.Rf1!+– Simple and best. Black’s b4-knight is too far off
and Black cannot protect his kingside without giving away the a5-pawn. After it falls, White’s advantage is
decisive.) 43.Rd1 Nd5 44.Qxa5 Ne3 45.Qd2 Qxa4 46.Rb1 Nd5 47.Rf1 Qa7² and Black should hold, thanks to the
strong d5-knight.]
42...Qxf6 43.exf6+ Kxf6 44.Rc1! Nd3!

45.Rc2!!+– An impossible move to see for a human player. However it exists, and gives White a big advantage,
169
since it is not easy for Black to make use of his kingside pawn majority and keep his very important a5-pawn alive.
(45.Rc3 Nb2=) 45...e5? 46.Rc6+ Kf5 47.Ra6 e4 48.Rxa5+ Kf4 49.Ra7 f5 50.Kg1 Ke3 51.a5+–

41...Nd3!

Now it’s a simple draw.

a) 41...Nc2? 42.Qb5 Ne3 43.Qe2+–
b) 41...Qe2? 42.Qb5 Nd3 43.Qb1 h5 44.h3±

42.Qb5

Setting a last trap.

42.Qd8 Qe2 43.Qf6+ Kh6 44.Rb1 Qc2=

42...Qe3!=

Black is accurate until the end.

42...Nxf4? 43.Qb2! Nd3 (43...Qxa4 44.Qf2 g5 45.h4! h6 46.g3+–) 44.Qd2±

43.h3 Nf2+ 44.Kh2 Qxf4+ 45.Kg1 Nxh3+

170
And a draw was agreed due to the coming perpetual.

½–½

An exciting first game of the minimatch which brought me the confidence to play against the reigning World
Champion. After going astray and making intuitive decisions, I still managed to play for an advantage. Of course, if
I had seen the idea with f5 and Nf6 earlier, things might have gone differently, but overall it was a good result for
me. I also drew the second game of the mini-match comfortably and perhaps that gave me the confidence needed for
our next encounter in the Dortmund super-tournament the following year.

171
172
 GAME 7
Gata Kamsky
John Fedorowicz
Los Angeles 1991
This game was played at the 1991 US Championship, where I became US Chess Champion for the first time. I had
quite mixed feelings about the title, but that’s a story for another time.

1.d4 Nf6 2.Nf3 g6 3.Bg5

Returning to my trusty Torre system.

3...Bg7 4.Nbd2 0-0 5.c3 d6

John Fedorowicz had a unique aggressive positional style that made him dangerous for any opponent. Prior to my
arrival in the US John was considered one of the best American players; he was the winner of my first tournament in
the US, the 1989 New York Open. Here he prefers the King’s Indian attacking set-up with ...d6 followed by ...e5.

6.e4 Nc6

One of the more aggressive ways to develop the b8-knight. Now the game transposes into a sort of Spanish game
where White has developed his queenside instead of his kingside. The position is very dynamic.

7.Be2

173
A quiet developing move. In later games, I would play Bb5 in true Spanish style, leaving the e-file open for the rook.

7...h6 8.Bh4 e5 9.0-0 Qe8

Black makes his ambitions clear. He gets out of the pin on his f6-knight and prepares to follow up with ...Nf6–h5–f4
and the ...f7–f5 pawn push, creating a problem for White’s h4-bishop. Since there is a threat to trap it with ...g6–g5
and ...f5–f4, White would have to take on f5, when ...gxf5 would open the g-file for a direct attack on the white king.

10.dxe5

I decided to exchange the central pawns mainly to discourage Black’s aggressive kingside pawn advance.

10...dxe5

11.b4

174
The situation has cleared and we can see the plans for both sides. White wants to play on the queenside, while Black
will try to get to the white king on the kingside.
Since the position has a semi-closed character, the minor pieces will play the crucial role, so both players will have
to be extremely careful with selecting the destination squares for their knights.
Black will try to hold back with developing his c8-bishop as it is already useful where it stands.
Also, instead of the game move, it was possible for White to exchange his h4-bishop for Black’s f6-knight
immediately, as the black knight is more useful at the moment, however I envisioned the position where White will
play f2–f3 and put his bishops on the f1- and f2-squares, ideal locations for them where they will protect the king
and help with queenside expansion.
11.Bxf6?! White shouldn’t play this move as it helps Black later with his kingside initiative. 11...Bxf6 12.Nc4 Bg7
13.b4 (13.Ne3 f5 simply helps Black to push his f-pawn further with tempo) 13...f5 14.Bd3

14...Kh8
a) 14...Be6 is another, probably worse, approach. I analyzed the following nice lines: 15.Qe2 a6 16.a4 Rd8 17.b5
Rxd3 18.Qxd3 fxe4 19.Qxe4 Rf4 20.Qe3 Bxc4 21.bxc6 Bxf1

175
22.cxb7! (22.Rxf1? Qxc6³) 22...Qb8 (22...Bc4? 23.Qc5!! Qb8 24.g3 Rxf3 25.Qxc4+ Kh7 26.Qxa6 e4 27.a5 Bxc3
28.Rd1 Rf8 29.Rd7+ A typical crazy computer line, but obviously all these sharp tactics are completely
unnecessary.) 23.Rb1 Bb5! 24.axb5 Qxb7 25.Qd3² White’s superior structure and more active pieces promise him a
long-term advantage.
b) 14...f4!?‚
15.Qe2 f4 16.Rfd1 (16.h3 g5 17.Nh2 h5‚) 16...g5 17.h3 g4 18.hxg4 Bxg4‚ and Black has obviously just got much
closer to achieving his goal of building an attack on the white king.

11...Nh5 12.Re1

12...Kh8?!

Possibly a waste of an important tempo.

a) 12...Nf4 seems the best alternative. Perhaps Black thought that I wanted to put my e2-bishop on the c4-square,
176
which is quite standard, but for the moment there is a better candidate for that ideal spot, and that is White’s d2-
knight! I’ll give the following examples to show you some sample plans for both sides. 13.Bf1

A curious position. White’s h4-bishop, while vulnerable, plays a big role in hitting e7 and d8, which are vital squares
for Black. The only way to get rid of that bishop is for Black to play ...g5, but that leaves the f5-square weak and
nullifies any chances of a serious black attack on the kingside. 13...b6 This move and the idea of developing Black’s
c8-bishop to the a8-h1 long diagonal is perhaps the best plan.
[13...a6 14.a4 (14.Nc4?! would lead to unnecessary complications: 14...g5 15.Bg3 Bg4 16.Ne3 Rd8 17.Qb3 Bxf3
18.gxf3 h5! 19.h3 Ne7!³ and the position is very double-edged) 14...Bg4 15.h3

15...Be6 (15...Bh5?! 16.Qc2 g5 17.Bg3 Qd7 18.Nh2 Rad8 19.Nb3² White has managed to set up good defenses on
the kingside before going on the counter-offensive on the queenside. Still, the game is far from over, and Black has
to play ...f5 at some point to create counterplay.) 16.a5² White’s queenside expansion becomes dangerous and Black
has to start thinking about defense against a possible Nd2–b3–c5; 13...f5? With his queenside under-developed, this
early push of Black’s f-pawn is usually dubious. 14.exf5 gxf5 15.b5 Nd8 16.Nxe5! Bxe5 17.Bxd8±]
14.Nc4 (14.a4?! Suddenly not that great, since it can be met with 14...a5 and the white queenside is under pressure,
177
as b4–b5 gets nowhere, and in fact creates a perfect post on c5 for the black knight.) 14...Bb7 15.Nfd2 f5!?

16.f3!?
[16.exf5 leads to complicated lines: 16...gxf5 (16...Rxf5? 17.Qc2± simply leaves Black with a weak isolated e5-
pawn and perfect outpost on the e4-square for the white minor pieces) 17.b5 Nd8 18.Nxe5 Bxe5 19.Bg3 Ng6
20.Bxe5 Nxe5 21.f4 Ndf7 22.fxe5 Ng5! with strong counterplay for Black.]
16...fxe4 17.Nxe4 Ne7 18.Qc2 Nf5 19.Bf2 Qf7 20.a4ƒ
b) 12...f5? is too early: 13.exf5 Bxf5 14.b5 Nd8 15.Bc4+ Kh7

16.Nxe5! Bxe5 17.Nf3±

13.Qc2

178
13...f6?!

At first sight this appears to be a very timid move. It also caused me to think that John was not happy with his
position, since his original plan of playing ...f5 could not work here, but a deeper look shows that Black takes the d8-
and e7-squares under control and creates access to the f7-square for his queen and possibly rook.
13...f5 14.exf5 gxf5 15.b5 Nb8 16.Nxe5 Bxe5 17.Bc4! Nd7 18.Nf3±

Black is hopelessly pinned along the e-file, and White will recapture his missing piece with interest, e.g. 18...Nhf6
19.Qd2 Kh7 20.Bxf6 Rxf6 21.Re2 Qe7 22.Qd5 Qg7 23.Nxe5 Rd6 24.Qf7 Nxe5 25.Rxe5 Qxf7 26.Bxf7 Bd7 27.Rae1
Kg7 28.Re7 Kf6 29.Bb3 Bxb5 30.h4 with a winning endgame.

14.Nc4 Nd8?!

Played in the spirit of the King’s Indian Defense! Black’s d8-knight is headed for the e6-square where it will be
much better placed than previously.
179
I thought that 14...Qf7 was better, with the strategic idea of developing coordination between all his pieces. 15.Nfd2
Nf4 16.Bf1 Be6 17.Ne3 Ne7 18.f3 h5

19.Nb3 (19.Bc4 Bh6 20.Bxe6 Qxe6 with an equal game, as 21.g3?? is terrible due to 21...Nh3+ 22.Kg2 Bxe3
23.Rxe3 g5; 19.Rad1 Bh6 20.Nb3 b6 21.c4 g5 22.Bf2 Rad8 23.Rxd8 Rxd8 24.c5 g4∞ Black has obvious
counterplay on the kingside.) 19...b6 20.c4

This move completes the transformation of the game into a fully-fledged King’s Indian Defense! The position is
very complicated, and Black’s kingside initiative should not be underestimated. 20...g5! (20...Bh6?! 21.Nd5 Ng8
22.Bg3±) 21.Bf2 f5 22.Nd2 Rad8 23.exf5 Nxf5 24.Nxf5 Bxf5 25.Ne4 g4‚ with a very dangerous attack for Black.

15.Nfd2 Nf4 16.Bf1 Nde6 17.f3 Bd7 18.a4 Rd8 19.Bf2²

180
The time has come to evaluate the position as both sides have completed the development of their pieces. Black is
worse, due to his inferior minor pieces. White’s queenside expansion continues unperturbed, while Black’s kingside
initiative is non-existent. Perhaps Black hoped to contain White’s aggression, but he is unlikely to be successful
because all the white pieces are participating in that area.

19...a6?!

19...b6 20.a5 Bc6 21.g3 Nh5 22.axb6 axb6 23.Ra7 Rf7 24.Ne3 Ng5 25.Bg2 Rfd7 26.Nb3 Bf8 27.Nd5 Qf7 28.c4
Ng7 29.h4 N5e6 30.b5

White succeeds in blocking the center and opening the queenside, giving him a big advantage. 30...Bxd5 31.cxd5±

20.Na5!

White finally starts to engage Black’s weak queenside pawns, probing for additional weaknesses that can be
exploited.

181
20.g3 This is another typical computer move that I simply did not consider as it means playing on the flank where
the opponent is stronger, but these human concepts simply don’t work with computers which find tactics, regardless
of strategic nuances. 20...Nh5 21.f4

21...exf4 (21...Rg8?! is too passive: 22.f5 Ng5 23.h4 Nh7 24.Be2 gxf5 25.exf5 Bf8 26.Kh2 Ng7 27.Bg4² and
strangely enough, it is Black who comes under pressure on the kingside, thanks in large part to the f5-pawn and
dominating the e4-square.) 22.g4 Ng5 23.gxh5 gxh5 24.Kh1 f5 25.exf5 Bc6+ 26.Kg1 Qd7

White’s seriously compromised king’s defenses give Black hopes of winning the game!

20...Bc8 21.b5

182
This move opens a line and creates the the additional possibility of b5–b6, with the idea of dislodging Black’s c7-
pawn from protecting the ‘KID’ d6-square.

21...axb5 22.axb5 Ng5 23.Kh1 f5

Finally Black has had enough of defending and decides to go over to counterattack.
23...Rf7

This was perhaps more accurate, with the idea of defending the critical d6-square with ...Bg7–f8. 24.b6 (24.Ndc4!?²
A very logical alternative, as this knight can now either join the fight on the queenside or to go to the e3-square
where it will keep an eye on the vital f5-square.) 24...c6 25.Ndc4 Bf8 26.Nb3 Rfd7 (26...Be6 27.Nc5 Bxc5 28.Bxc5
Rfd7 29.Be3 Nd3 30.Bxd3 Rxd3 31.Nb2 R3d6 32.Na4 Bc8 33.f4² Suddenly it’s the black king that starts to feel
uncomfortable, given that the g7-bishop defender is gone.) 27.Ra8²
White maintains a small advantage, but it’s not clear how to make further progress.

24.b6
183
24...fxe4?!

This makes things easier for White as Black gives up his plans of attack with ...f4 and ...g5.
24...c6 25.Bc5
[25.Nac4 fxe4 26.Nxe4 Nxe4 27.Qxe4 (27.fxe4 Qe7=) 27...Bf5! With precise play Black equalizes. 28.Qe3 Bd3
(28...Nd5? 29.Qc1±) 29.Bg3 Bxf1 30.Rxf1 Kh7 31.Qe4 Qe6 32.Rfd1 Rd5!∞]
25...Rf7 26.Nac4 Rfd7

Black is quite solid and the position is double-edged, for example 27.Ra8 Nfe6 28.Ba3 f4 29.Nb3 h5 30.Nc1 Nf7
31.Nd3 g5 32.Be2 Bf6 33.Nf2 Kh7 with counterplay. Black has managed to successfully regroup all of his pieces,
blocking White on the queenside while keeping the center closed. However, due to his greater space for
maneuvering, White still maintains a certain advantage.

25.Nxe4 Nxe4 26.Qxe4 cxb6

184
27.Nc4!

White goes after Black’s b6-pawn with his knight and not his f2-bishop.
27.Bxb6 Bf5 would give Black exactly what he wanted, a break in the center. 28.Qc4 Rc8 29.Qb3 g5!ƒ

27...Qc6?

After this mistake, Black’s position becomes critical.

27...Bf5 was best, a perfect mix of defense and offense! 28.Qxb7 Rb8 29.Qc7 Rf7 30.Qd6 Rd7 31.Qc6

31...Nd3 ( 31...b5? 32.Nxe5 Bxe5 33.Bd4, winning immediately. 33...Bxd4 34.Rxe8++–) 32.Bxd3 Rxd3 33.Nxe5
Qxc6! 34.Nxc6 Rc8 35.Ne7 (35.Ne5 Rdxc3 36.Bxb6 g5 37.Bd4 R3c7

185
This is perhaps the best try for Black. White has an extra pawn, but it is only 3 v 2 on the kingside, and Black has
two active bishops. The only question will be if White can convert his advantage, but it doesn’t seem likely. 38.g4
Bc2 39.Kg2 Rd8 40.Bb6 Rd2+ 41.Kg3 Rb7 42.Ba5 Rd5 43.Bc3 Kh7 44.Rac1 Ba4 and Black’s defense holds.)
35...Bxc3 36.Nxc8 Bxa1 37.Rxa1 Bxc8 38.Bxb6 Kg7

and this endgame can probably be held.

28.Qxc6 bxc6 29.Bxb6 Rd5

186
30.Bc7!

A key move in my opponent’s time trouble. Now Black loses his pawn, and the fact that his minor pieces are inferior
to their white counterparts is an important factor as well.

30...Nd3

30...Re8 31.Ra8 Bd7 32.Rxe8+ Bxe8 33.Bxe5 Bxe5 34.Nxe5 Rd8 35.g3 White is simply a pawn up.

31.Bxd3 Rxd3 32.Nxe5

32.Bxe5 seems more logical, to exchange a pair of bishops, however I felt that the opposite-colored bishops,
together with the pair of rooks, would give me best chances due to the exposed black king and possibility of an
attack, something that would be unpleasant for my opponent.

32...Bxe5 33.Bxe5+ Kg8 34.Bd4±

187
This bishop closes off the black rook on the d3-square and prepares for the infiltration of my rooks on Black’s weak
seventh rank.

34...Rd8 35.Ra7 Rd7

Fortunately for me, Black cannot play 35...c5, otherwise he would have had good drawing chances. 36.Bxc5 Rxc3
37.Bb6!

37...Rd7 ( 37...Rf8 38.Bd4 Rd3 39.Rg7+ Kh8 40.Ba1+– winning more material) 38.Re8+ Kf7 39.Rxc8!+–

36.Ra8 Rd8 37.h4 c5 38.Bf6 Rf8 39.Re7 Rd7 40.Kh2!

188
Patience! Black has no useful moves.

40...Rxe7

40...c4 41.Rxc8 Rxe7 42.Rxf8+ Kxf8 43.Bxe7+ Kxe7 44.Kg3 transposes into a lost endgame for Black.

41.Bxe7 Re8 42.Bxc5+–

White is two pawns up. Rook or bishop endgames are both won. The rest requires no comment.

42...Kf7 43.Ra7+ Ke6 44.Bd4 Re7 45.Ra8 Rc7 46.Kg3 Kd5 47.Rb8 Kc4 48.Rb6 Bf5 49.h5 gxh5 50.Rxh6 Rh7
51.Rxh7 Bxh7 52.Kh4

1–0

I have included this game as I felt it was very indicative of my early approach to the Torre System and the whole
189
Petrosian-like approach to chess.

Presenting my 1991 US Championship trophy to my first American sponsor, James Cayne, former CEO, Bear &
Sterns

190
191
192
 GAME 8
Alexander Beliavsky
Gata Kamsky
Belgrade 1991
This game is memorable for me for a couple of reasons — I got the chance to play the famous GM Beliavsky, and
the provocative way I played with my queenside rook throughout the game (in fact the theme of the game is
‘dancing rooks’). Hopefully readers will enjoy the game for aesthetic reasons as much as for its positional
maneuvering, but I must warn you that it is a very complex game and requires a lot of analysis to understand the true
nature of what is going on. Part of the complexity lies in the middlegame, where the pawn structure is unbalanced,
with most of both sides’ pieces looking for a favorable way to enter the game. The other, probably even more
complex, part is the endgame, where strange moves that seem to make no sense nevertheless follow a certain logic
and plan. I would even say that the endgame is by far one of the most complex that I have ever played. I hope that by
going through the different lines, and on occasion trying to guess the best plan or move, the reader will be able to
develop his or her own endgame skills.

1.d4 Nf6 2.c4 g6 3.Nc3 d5

The Grünfeld Defense had been my mainstay opening ever since the Karpov-Kasparov matches because I liked the
variety of positions — from closed to semi-closed to open — that could arise on the board.

4.Nf3 Bg7 5.Bg5

Beliavsky goes for this super-solid line which doesn’t challenge Black’s opening set-up but rather attempts to steer
the game into a maneuvering phase, where my distinguished opponent hoped to use his advantage in experience.

5...Ne4 6.Bh4 Nxc3 7.bxc3 dxc4 8.e3 b5!?

193
Quite audacious and unprepared! I have always hated studying for hours in preparation for a game as it makes chess
more like a science than a creative outlet. Alas, with the unstoppable march of computer progress, the element of
creativity can now only be found at home, where top grandmasters must spend many hours deeply studying the
positions resulting from the computer’s suggestions. The engine’s first or second lines are considered to be the
perfection that humans must strive to achieve. When I gaze with nostalgia at the 19th and early 20th centuries, when
chess maestros were considered to be truly intellectual geniuses, creating wonders on the chessboard, this seems
such a downfall for chess.
One of the all-time chess giants and geniuses, the American grandmaster and World Champion, Bobby Fischer,
recognized the potential danger of chess becoming a matter of scientific preparation and invented his own variant of
the game, Fischer Random Chess. Perhaps this represents the future, as the shuffling of pieces on the first rank,
resulting in numerous possible starting positions, reduces the effect of preparation to zero. This would, in effect, free
grandmasters from the prison of mathematical precision and enable them to truly create something again, while
being able to enjoy other pursuits in life.

9.a4 c6 10.Nd2

White commits himself too early by revealing his plan of transferring this knight to the c5-square.
10.Be2 was more logical, awaiting Black’s choice of development.

10...a6

Black decides to play simply and to hold on to his extra pawn on c4 and weather the storm. A sort of caveman
approach, but why not! White doesn’t have too many plans in the position; he will have to play the thematic e3–e4
pawn push at some point, hoping to gain space with the further advance of the e-pawn and possibly expand on the
kingside.

11.Be2

194
11...Ra7!?

The start of a bold plan to develop the a8-rook, since the direct 11...Nd7 is impossible because the pins after 12.Bf3
would lead to the collapse of Black’s queenside pawn formation. However the idea is not new, and had been played
before in several games. Black keeps his knight on b8 and bishop on c8 to ensure that he can always recapture on b5
with the c6-pawn. With only one possible weak square on c5, which can be controlled by ...Nb8–d7, Black prepares
for the freeing break in the center, after which he will simply be better, as White will have no compensation for the
sacrificed pawn.

12.0-0?!

An inaccuracy.
White should have started harassing Black’s b8-knight immediately with 12.Bg3, e.g.

12...e5!? An interesting computer suggestion. Black counter-sacrifices his e-pawn in exchange for freedom for his
195
queenside pieces.
a) 12...Nd7 would be a riskier choice, e.g. 13.Bf3 Qb6 14.Rb1 e5 15.Nxc4 Qc7 16.Nxe5 Nxe5

17.Bxe5 (17.dxe5 Be6 18.0-0 0-0 19.Qd6 bxa4 Black potentially equalizes.) 17...Bxe5 18.dxe5 Bf5! (18...0-0
19.Qd6 Rd8 20.Qf6 Bf5 21.Ra1 Re8 22.0-0 Re6 23.Qh4 Rxe5 24.Rfd1 Kg7=) 19.Rc1 0-0 20.Qd6 Rd8

with equality for Black. 21.Qxc7 Rxc7 22.Ra1 Rd3 23.Rc1 Rd8 24.0-0 (24.a5 Re7 25.Bxc6 Rxe5 26.Bb7 Rc5 27.0-
0 Rd7µ) 24...bxa4=
b) 12...Bb7!? This funny move is among the computer’s top suggestions. The advantage of placing the rook on the
a7-square is clear: it protects the b7-bishop and frees the a8-square for Black’s queen, similar to the typical
maneuver in the Reti where the white queen goes to a1. Play might continue 13.Bxe5 Bxe5 14.dxe5 Bf5 15.e4 Bc8
16.0-0 0-0 17.f4 Nd7 18.Nf3 Nc5 19.Qc2 Nb3!

196
Black correctly doesn’t go after White’s pawns and intends to exchange heavy pieces because the pawn formations
favor Black in the resulting endgame. 20.Rae1 (20.Rad1 Rd7=) 20...Re8 21.f5 (21.h3 Rd7 22.Kh1 Qe7 23.f5 a5
24.e6 fxe6 25.fxg6 e5! 26.gxh7+ Qxh7 27.Nh2 Kh8 28.Rd1 Nc5 We have reached a complex position with
possibilities for both sides.) 21...Rd7 22.e6 fxe6 23.fxg6

23...e5! A key move in Black’s defense. Preventing White’s e4–e5 is vital, otherwise Black would be stuck with his
terrible c8-bishop, while facing White’s increasingly dangerous attack on the kingside. 24.gxh7+ Rxh7³

12...0-0

The opening phase is over.

197
Black’s position has no weaknesses while White is hard-pressed to find not just compensation for the pawn, but also
a plan of how and where to arrange his pieces.

13.Bf3

With this move White reveals his decision to switch to a more defensive and prophylactic approach to minimize the
damage.

13...Rd7!?

Having played ...Ra7 already, it was difficult to stop there. I started thinking about putting it on a more suitable
square.
Something like 13...Qc7 was surely better, preparing the ...e7–e5 push, which would have pressed White hard for
answers, for example 14.Qb1 f5 15.Bg3 e5 16.Qb4 f4 17.axb5 cxb5 18.exf4 exf4 19.Bh4 h6 20.Bd5+ Kh8 21.f3 g5
22.Bf2 Nd7 23.Rfe1 Nf6 24.Be4 Rd8 Black is fully developed and White has no compensation for the pawn.

14.Qb1

The immediate 14.Ne4 was more to the point. 14...Qc7

198
15.Qe2 (15.Nc5 Rdd8 16.Qc1 f5 17.axb5 cxb5 18.Qa3 Rde8 19.Rfe1 e5 20.h3 Qb6! 21.Qb4 Bf6 22.Bg3 f4 23.Bh2
Nc6 24.Qa3 fxe3 25.fxe3 exd4 26.exd4 Nxd4µ The tactical justification behind Black’s central pawn break.)
15...Rdd8 16.Rfe1 f5 17.Nc5 Rde8 A thematic move in this position. After the successful transfer from a8, the rook
takes a direct part in the fight for the center after the ...e7–e5 push. 18.e4! Bf6 Another important move! Eliminating
the strong h4-bishop is necessary to weaken White’s hold over the dark squares.

19.Bxf6 exf6 20.Qb2 Nd7 21.Qa3 fxe4 22.Nxe4 Re6 23.Nc5 Rxe1+ 24.Rxe1 Nb6 25.axb5 axb5 26.Qa5 White
finally obtains full compensation for the sacrificed pawn.

14...Qc7

199
15.Ne4?!

a) 15.Qb4?! This move is questionable, as we’ll see later. White’s queen, while looking active on the b4- square, is
removed from play on the kingside. 15...e5 16.Ne4 f5 17.Nc5 Rdf7 18.h3 h6 A critical move for Black if he is to
make progress.

The threat of capturing the h4-bishop forces White to make concessions regarding his plans on the queenside, or to
risk his kingside being overrun. 19.Bg3 f4 20.exf4 (20.Bh2? Keeping the tension would be a mistake as White’s
queen on b4 is far removed from the action. 20...fxe3 21.fxe3 Qe7 22.axb5 axb5 23.Ra8 exd4 24.exd4 Qe3+ 25.Kh1
Rxf3!–+ Black’s attack breaks through.) 20...a5! An important intermediate move which leads to a complex
position. Other responses make life easier for White.
[20...exd4 21.f5 Be5

200
22.cxd4 (22.fxg6?? This loses immed iately. 22...Rxf3–+) 22...Bxg3 23.fxg3 Rxf5!? Black sets a small trap.
(23...Bxf5 24.g4 Bc8 25.Rae1°) 24.Qc3! (24.g4? Rxf3!–+) 24...R5f7 25.g4 Re8 26.Rfe1 Rfe7 27.Rxe7 Qxe7
28.axb5 axb5 29.Ra8 with clear counterplay because of Black’s open king and queenside knight and bishop both
stuck on their starting squares.]
21.Qa3
[21.Qb1 exd4 22.f5

22...Bxf5! 23.Bxc7 Bxb1 24.Bxb8 Bd3 25.Nxd3 cxd3 26.cxd4 (26.Bd6 Rd8µ) 26...Rxb8 27.axb5 cxb5 28.Bd5
(28.Rxa5 Rf4µ) 28...a4 29.Rfd1 Rd8 30.Bxf7+ Kxf7 31.Rxd3 Rxd4 32.Re3 Rd5³

201
White has to defend his position very carefully now. 33.Rb1 (33.Rc1 Be5µ) 33...Bf8 (33...Be5!? 34.Kf1 Bd6 is also
a great try.) 34.Rf3+ Kg7

This endgame is extremely difficult for White to save. 35.Ra1! (35.Kf1? a3 36.Ke2 Be7 37.Re3 Kf7 38.Reb3 b4
39.Rxb4 Bxb4 40.Rxb4 Ra5 41.Rb1 a2 42.Ra1 Ke6 43.Kd2 h5 44.Kc3 h4 45.Kb2 Kf5 with a technically winning
position, as the pawn endgame arising after 46.Rxa2 would obviously be lost for White: 46...Rxa2+ 47.Kxa2 Ke4
48.Kb2 Kd3–+) Now White has good drawing chances as the rook endgame arising after 35...a3 36.Rfxa3 Bxa3
37.Rxa3 b4 38.Ra7+ Kf6 39.Rb7 is just equal.]
21...exd4 22.f5 Qb6

202
23.axb5 (23.Ne6 b4 24.cxb4 axb4 25.Rab1 Bxe6 26.fxe6 Rxf3 27.gxf3 Na6 28.Qa2 Qc5 29.Rfc1 Re8 30.Rxc4 Qd5³
Black seems to be better in this wild position thanks to his very strong connected pawns.) 23...Bxf5
[23...cxb5 24.Ne6 (24.Bxb8 Qxb8 25.Bd5 Bxf5 26.Qxa5 dxc3–+) 24...Bxe6 25.fxe6 Rxf3 26.gxf3 Qxe6 27.Qxa5
Qf5 28.Rab1 Nc6 29.Qb6 Qxf3 30.Qxb5 dxc3 31.Qxc4+ Kh8

32.Rbd1! (32.Rb6 Nd4–+) 32...Ne5 33.Bxe5 (33.Qc7 Qf5ƒ) 33...Bxe5 34.Rd5 Bg7 35.Qd3 Qf4 36.Qg3 Qe4 with an
equal game, as long as Black avoids the exchange of queens.]
24.Bxb8 (24.bxc6 Nxc6 25.Bd5 dxc3³) 24...Qxb8 25.bxc6 dxc3 26.Bd5 Kh7 27.Bxf7 Rxf7 28.Qxa5 c2!µ The
bishops are stronger than rooks in this position.
b) 15.Qb2!?

203
15...e5
[15...Rdd8 16.Rfb1 (16.Qa3 An ideal square for White’s queen. It covers the d6and e7-squares and is ready to join in
along the a-file any time. 16...Rde8 17.Bg3 e5 18.Ne4 f5 19.Nd6 Re6 20.Nxc8 Rxc8 21.axb5 axb5 22.d5 Ree8
23.Rfb1 Qd7 24.dxc6 Nxc6 25.Rd1∞

White gets a certain activity that seems to give him adequate compensation, at least temporarily, for the missing
pawn.) 16...Rfe8! (16...Bf5?! The attempt to provoke White into closing off his f3-bishop by playing 17. e4 does not
solve the problem. 17.e4 Be6 18.Bg3 Qd7 19.Qa3 Rfe8 20.h3 Qc8 21.axb5 axb5 22.Qa5!² White keeps unpleasant
pressure on Black’s queenside.) 17.Nxc4 Bf5

204
Black is forced to enter a tactical skirmish, as otherwise White is just better thanks to his superior pawn structure,
piece activity and now equal material. 18.Bg3 (18.e4 bxc4 19.exf5 Qf4 20.Bg3 Qxf5 21.Qe2 Qd3 22.Rb4 Qxe2
Capturing the c3-pawn wasn’t good, due to Rab1 followed by Rxc4. 23.Bxe2 Nd7 24.Rxc4 e5!=) 18...e5 19.Nxe5
Bxe5 20.dxe5 Bxb1 21.e6 Qe7 22.exf7+ Qxf7 23.Rxb1 Qf6 and White’s monster pair of bishops more than
compensates for the missing exchange. Black will have to fight for a draw.]
16.Rab1!

Suddenly the threat of Nxc4 forces Black to make decisions. 16...exd4

[16...Bb7?! 17.Ne4 Re8! (17...f5 18.Nc5 Rd6 19.Be2!± recreates the threat to capture on c4, but this time with the
bishop! Black is much worse. 19...Bc8 20.Bxc4+±) 18.Nf6+ Bxf6 19.Bxf6 e4 20.Be2 Qd6 21.Be5 Qe6 22.Bg3ƒ
This square looks like an ideal position for White’s bishop, until Black’s knight gets to d5 of course.]
17.cxd4
[17.exd4? f5! 18.Nxc4 (18.h3 f4 19.Be4 Rdf7 20.Nxc4 g5!! 21.Bxg5 bxc4 22.Qxb8 Qa5! 23.Bh4 Qh5 24.Bxc6 Kh8³
finally capturing the bishop) 18...h6 19.axb5 axb5 20.Ne5 Bxe5 21.dxe5 g5³]
205
17...c5!

18.axb5 The most accurate.

[18.Bg3?! Qd8! 19.axb5 (19.Bxb8 cxd4 20.Qc1 — or 20.exd4 Rxd4µ — 20...dxe3 21.Ne4 Bb7µ) 19...cxd4 20.Nxc4
dxe3 21.Qa2 exf2+ 22.Bxf2 (22.Kh1 axb5 23.Rxb5 Na6=) 22...axb5 23.Rxb5 Na6 and Black survives. 24.Nb6?
(24.Bg4? Rc7!=) 24...Rd2³]
18...cxd4
[18...axb5!? 19.Bg3 Qa5! (19...Qd8 20.Bxb8 cxd4 21.Qxb5 c3 22.Nc4 d3 23.Qa4 c2 24.Rb5

and in this incredible position White keeps his advantage, since he can always give back the piece for Black’s two
passers in the center.) 20.Qxb5 Qxd2 21.Qxc5 Na6 22.Qxc4 Qa5 23.Ra1 Qb4 24.Qc6 Qe7 25.Rxa6 Bxa6 26.Qxa6

206
with an obvious advantage for White. However it’s not clear if it’s enough to win.]
19.b6 Qc5 20.Qc2! c3 21.Nb3 Qc4 22.Nxd4 Bxd4 23.exd4

23...Nc6! (23...Bb7 24.Rb3 Bxf3 25.gxf3 Qxd4 26.Bg3± White has a big advantage in his far more dangerous passed
pawn supported by the equally strong g3-bishop.) 24.Bf6 (24.Bxc6 Qxc6 25.Rb3 Bb7 26.f3 Rc8 27.Bf2 Rd5! 28.Rc1
Rb5=) 24...Bb7 25.Rfc1 Nxd4 26.Bxd4 Bxf3 27.Qxc3 Qxc3 28.Bxc3 Be4 29.Ra1 Rc8 30.f3 Bb7 31.Kf2 Rc6 32.Ba5
Rxc1 33.Rxc1 f6

207
With a slight advantage to White, however Black should equalize with accurate play.

15...Rdd8?!

Rushed play, based on intuition. Black should have started active play in the center immediately.
15...f5 16.Nc5 Rd6 was better, with the idea 17.Bg3 e5! and the position is unclear, for example 18.dxe5 Rdd8 19.e6
Qe7 20.Qb4 Bxe6 21.Bh4 Bf6 22.Bxf6 Qxf6 23.Nb7 Rd2 24.Rfd1 Rxd1+ 25.Rxd1 f4!³ and White has yet to fully
equalize.

16.Qb4

16.axb5 cxb5 17.Qb4 f5 18.Qc5

is an extra possibility that would not exist had Black played ... f5 on move 15, but it doesn’t pose any problem:
18...Rd7! (18...Qxc5 19.Nxc5 Rde8=) 19.Bg3 e5 20.Qxc7 Rxc7 21.Nd6 f4 22.exf4 exf4 23.Nxc8 Rcxc8 24.Bd5+
Kh8 25.Bh4 Bf6 26.Bxf6+ Rxf6 27.Rfe1 Rd6 28.Bb7 Rcd8 with advantage for Black, whereas White has to play
208
very carefully. Hence White’s reluctance to go in for such complications.

16...f5 17.Nc5

17...Rde8!

The rook dance continues! Black finally prepares ...e5.

18.Rfd1 Bf6?!

And I change my mind! Thinking about safety first is not always good.
The bold 18...e5 was required. 19.d5 e4 20.d6

20...Qa7! The move that I missed during the game. (20...Qf7 21.Be2 f4 22.axb5 cxb5 23.Nxa6! An important
breakthrough for White that evens the game. An interesting line could go 23...Bxa6 24.d7 Nxd7 25.Rxa6 Ne5
26.Qxb5 Rb8 27.Rb6 Rxb6 28.Qxb6 f3 29.Bf1 fxg2 30.Bxg2 Nf3+ 31.Bxf3 exf3 32.Qc6 Bxc3 33.Bg3 Kh8
209
34.Rd5 (avoiding ...Qf5) 34...Bf6 35.Qxc4 Qe6 36.h4 Qh3 37.Qf1 Qg4 38.Kh2 Bxh4 39.Qh3 Bxg3+ 40.Qxg3 Qe6
41.Qd6 Qf7 42.Kg3= Finally an equal position!!) 21.Be2 Nd7 22.Nxd7 Bxd7 23.Be7 Rf7 24.axb5 cxb5 25.g3 Bc8!µ

The bishop goes back to guard duty on its starting square, freeing the d7-square for the queen to escape the X-ray of
White’s a1-rook!

19.Bxf6 Rxf6

210
20.Ra2

Too mild.
20.e4! e5 21.exf5 gxf5 22.Re1

22...e4?!
a) 22...Rh6 23.dxe5 Rxe5 24.h3 Nd7 25.Nxd7 Bxd7 26.axb5 cxb5 27.Rxe5 Qxe5 28.Rd1 Be8 29.Rd5ƒ
The position opens up and Black’s king is wide open. White has better chances to recoup his losses while keeping
his attack going.
b) 22...Qd6!? 23.Rxe5 Rxe5 24.dxe5 Qxe5 25.h4 Rf8!³ keeps Black’s hopes for an advantage alive. (25...a5?
26.Qxa5 Qxc5 27.Qc7+–) 26.Rd1 a5!µ
23.Bxe4! We both probably missed this nice tactical computer motif. 23...fxe4 24.Nxe4 Nd7 25.Nxf6+ Nxf6
26.Rxe8+ Nxe8 27.Qc5 Nf6 28.axb5 axb5 29.Ra7 Qf4 30.h3 Ne4!? 31.Qxc6 Qxf2+ 32.Kh2 Qg3+ with a well-
211
deserved draw for both players.

20...Nd7

And the response is just as respectful. At that time I still hadn’t learned how to play boldly for a win vs. famous
grandmasters.
20...e5 21.Rda1 e4 22.Bd1 g5

White can only watch Black’s preparations for the assault on the kingside. 23.g3 Qf7 24.axb5 cxb5µ

21.Rad2?!

Having already spent so long over his moves, White didn’t have much time left at this point and started to play
general purpose moves.
21.Nxd7 was preferable, for example 21...Bxd7 22.g3 Qd6 23.Qa5 Bc8 24.Raa1 Qd8 25.Qb4 Qc7 26.h4 h6 27.Bg2
e5!?

212
Before White plays f4. 28.axb5 cxb5 29.dxe5 Rxe5 30.Rd4 Kg7 31.Rad1 g5 32.hxg5 hxg5 White is still faced with
the unpleasant task of defending his position. His counterplay is limited as Black controls most of the important
squares.

21...Qb6

21...Nxc5 would have been more accurate, though it is difficult to ruin such a position for Black with non-committal
moves. 22.Qxc5 Qd6

23.g3 (23.Qb6 g5 24.g3 f4 25.exf4 gxf4 26.g4 Bxg4!µ) 23...Rd8 24.Qb6 Rd7 25.Ra1 Bb7 with advantage for Black.

22.g3 Nxc5 23.Qxc5

White decides to exchange queens to prevent any possible black attack on the kingside. Only two results are possible
now.

23...Qxc5 24.dxc5 Kf7 25.axb5 axb5!?

213
25...cxb5

At first sight this move seems much stronger, however I was afraid to leave White with an open file and such a
monster pawn on c5, and anyway in view of fantastic defensive resources by White, it does not win. 26.Bd5+
(26.c6? e5 27.c7 e4–+) 26...Kg7! (26...e6?! is not good, since it is better for Black to leave his pawn on e7 to prevent
the weakening of the seventh rank. 27.Bc6 Re7 28.Rd8 Rc7 29.Bd7! with great drawing chances) 27.f4 e5 28.Kf2 e4
29.Ba8 Rf7 30.Rd6 Rc7 31.Bc6 (31.c6 a5µ) 31...Re6 32.Rxe6 Bxe6 33.Rd6 Kf7 34.Ke2 Ke7

We have reached a critical position for this endgame as White is quickly approaching zugzwang. 35.h3! (Walking
the king into the d-file with 35.Kd2? would be a blunder as it allows Black to pin White’s d6rook after 35...Bd7!–+.)
35...Bd7
a) 35...h5 As a general rule, putting your pawns on the same color of squares as your bishop is bad, and here is no
exception, although stopping any g4-counterplay also brings some extra ideas. 36.h4! It seems that Black cannot
breach White’s defenses! 36...Bd7!? This idea allows Black to get rid of his bad bishop, however the rook endgame
is going to be extremely tricky.

214
[36...Rc8 The idea of exchanging rooks does NOT win!! 37.Ke1! (37.Bb7 Rxc5 38.Rxa6 b4 39.cxb4 Rb5 40.Bc6
Rxb4 41.Ra7+ Kf6 42.Ra6 Rb2+ 43.Kd1 Rb3 44.Ke2 Ke7 45.Ra7+ Kd8!–+) 37...Rd8 38.Rxd8 (38.Bb7? Rxd6
39.cxd6+ Kxd6 40.Bxa6 Kc6–+ Black captures White’s bishop.) 38...Kxd8 39.Kd2 Bf7 40.Bb7

40...Kc7 (40...a5 41.Bc6 Be8 42.Bd5=) 41.Bxa6 Be8 42.Kc2 Kc6 43.Kb2 Bd7 44.Ka3 Kxc5 45.Bb7 Be6 46.Kb2 b4
47.Kc2! Kb6 48.Ba8=

and Black cannot make progress!!! Draw!]

37.Bxd7 Rxd7 38.Rxg6

215
38...a5 ( 38...Rd3 39.Rg7+! and White even starts to play for a win!) 39.Rg7+ Kd8 40.Rxd7+ Kxd7 41.g4!= and it
seems that the resulting pawn race ends in a draw as well!
b) 35...Rc8 36.Bb7 Rxc5 37.Rxa6 b4 38.cxb4 Rb5 39.Ra7

and thanks to the trick 39...Kf6 40.g4! White has great drawing chances.
36.Bxd7 Rxd7 37.Rxa6 Rd3 38.Rb6 Rxc3 39.Rxb5 Kd7 40.Rb6 Rd3 41.c6+ Kc7 42.Rb4 c3

216
43.Rc4 Rd2+ 44.Ke1 Rh2 45.Rxc3 Rxh3 46.Kf2 Rh2+ 47.Kf1 Rd2 48.Rc5 Rd6³ Black wins the pawn, however his
chances of winning this 4 vs. 3 rook endgame are very small.

26.Ra2

Of course. White’s pawn structure is improved and the extra pawn’s influence is lessened.

26...e5 27.Ra7+ Re7 28.Ra8 Bd7 29.Rd8 Be8µ

Following the Dutch Defense motif, Black’s bishop is better placed on e8 than on c8, although right now it is still
more like a big defensive pawn in the chain. Black has managed to consolidate his position while keeping the extra
pawn. Now it was time to look for the way to improve my pieces.

30.e4

30.Ra1 Trying to keep control of the a-file gives up the c5-pawn. 30...e4 31.Be2 Rfe6 32.h4 Re5 33.Kf1 Rxc5
217
34.Ke1 Rd5 35.Rb8

35...h5 (35...h6 keeping the option of playing ...g6–g5 is a strong alternative) 36.Raa8 Re6 37.Ra7+ Bd7 38.Rbb7
Red6 39.Ra2 (39.Rb8 Rd2–+ and Black makes progress in getting his rooks into White’s half of the board.) 39...Kf6
(39...Ke7 40.Rc7 Rf6 41.Rca7 Kd6 42.Rb7) 40.Rb8 Be6 41.Rb6 Bf7

42.Rb8 (42.Ra8 Rd2–+) 42...Rd8 43.Rb6 R5d6 44.Raa6 Be8 45.Rb7 Ke6 46.Ra2 R6d7 47.Rb6 Kd5 48.Rba6 Kc5
49.R6a5 Bf7 50.Ra6 Be6 51.Rb2

218
51...Rg8! Black opens up a second front on the kingside, deciding the game. 52.Rba2 g5 53.hxg5 Rxg5 54.Ra8 h4–+

30...f4

31.g4!?

White attempts to block the board. The alternative was to play 31.Bg4 h5 32.Bc8 fxg3 33.fxg3 Rf3 34.Rc1 Ra7
35.Rd2 Ke7 and the white pawns are defenseless.

31...g5

Of course. White cannot be allowed to play g4–g5.

32.h4

219
32...Rh6?

32...Ra7! Taking control of the a-file was critical. 33.hxg5 Rg6 (33...Re6!? with the idea ...Ra3/ ...Kg6, was very
strong as well, e.g. 34.Rc8 Ree7 35.Kg2 Ra3 36.Rc1 Kg6 37.Rd8 Kxg5µ) 34.Rc8 Rxg5 35.Kf1

35...Rg7 36.Rdd8 Bd7 37.Ra8 Rxa8 38.Rxa8 Rg8 39.Ra7 Rd8 40.Be2 Kf6 41.Ke1 Be8 42.f3 h5 43.gxh5 Bxh5
44.Rc7 Be8 45.Ra7 Bd7 46.Kf2 Ke6 47.Bf1 Rh8 48.Kg1!µ and White maintains some sort of fortress.

33.hxg5 Rg6 34.Kf1

34.Ra1! would most likely lead to a draw, as the activity of the white rooks fully compensates for the loss of the
pawn. 34...Rxg5 35.Kf1 Rg6 36.Raa8 Bd7 37.Rf8+ Kg7 38.Ke2 Be6 39.Rfd8 Rc7 40.Re8 Kf6 41.Red8 Kg5 42.Rd6
Bxg4 43.Bxg4 Kxg4 44.Rd1 Kg5 45.Rf8 Kh6 46.Re8=

34...Rxg5?

220
Black should have taken advantage of the a-file with 34...Ra7!

35.Ke2

35.Ra1 Bd7 36.Raa8 Rg7³ leads to a similar almost equal position as in the lines above.

35...Rg6?

35...Ra7!

36.Rh1 Kg7 37.Ra1!

White finally achieves the ideal setup for his pieces.

37...h6 38.Raa8 Bf7 39.Ra6 Rc7 40.Rd6 Rf6 41.Rd8 Kg6

At this point I realized that I had completely misplayed my advantage and given White too much. Since my rooks
221
are fully tied up defending pawns, it was time to active the strongest endgame piece, the KING!

42.Kd2?

The beginning of a strange plan by White.

I suspect that my opponent wanted to press his advantage and infiltrate with his king all the way to b6, where it
would capture my c6-pawn and lead to victory.
Possibly my feeble performance in trying to realize my advantage had given him grounds for hope, however the real
reason why this move puts White in danger is that it leaves his f3-bishop unprotected. In many lines, Black is able to
use this to his decisive advantage.
42.Bg2 would have made more sense, to liberate the f3-bishop from guard duty and relegate that role to the f2-pawn,
fully blocking the kingside from any potential black diversion. The king would find the f2-square a good shelter.
42...Kg5 43.f3 h5 44.gxh5 Bxh5 45.Kf2µ with advantage for Black, but very likely not enough to win the game.

42...Kg5 43.Kc2? Be6 44.Re8 Rg6

45.Ra1?

This move is probably the critical mistake. White releases Black’s c7-rook from defense and allows him to improve
his position.
45.Kd2! Bxg4 46.Ke2 Be6 47.Ra1! Kf6 48.Rh8! Despite being two pawns down, White manages to control Black’s
movements effectively and restrict his rooks.

45...Kf6?!

Fortunately this move doesn’t throw away the win. I didn’t realize that this position is almost reciprocal zugzwang, a
favorite with endgame study composers!!
45...Rcg7!!

222
The position is a sort of zugzwang. White’s a1-rook is tied down to keeping control of the a-file and eyeing a
possible check on g1 should Black decide to capture on g4. His king has to keep control of the d2-square, and his e8-
rook is tied to e-file so that it can recapture on e5 if Black’s e6-bishop moves. 46.Rf8 (46.Kb2 Rd7 47.Kc2 Rd3
48.Be2 Rh3–+) 46...Rg8 47.Rxg8 Rxg8

and White cannot hold this zugzwang without conceding critical squares. 48.Be2 (48.Ra6 Bxg4 49.Bxg4 Kxg4
50.Rxc6 Kf3–+; 48.Rh1 Ra8 49.Rh5+ Kg6 50.Rxe5 Kf7!! 51.Kb2 b4 The final breakthrough, all made possible
because of the v ulnerability of White’s f3-bishop.) 48...f3 49.Bxf3 Rf8–+

46.Rh1?

Simplifies Black’s task.

a) 46.Rh8!

223
was probably White’s best chance. The idea is to exchange his g4-pawn for Black’s h6-pawn and transpose into a
four-rook endgame which will be extremely difficult for Black to win. 46...Kg5! Black forces White’s a-rook to
leave the a-file.
[46...Rc8?! 47.Rh7 Bxg4 48.Bxg4 Rxg4 49.Rxh6+ Kg5 50.Rh7! is less clear. (50.Re6? Rg2 51.Rxe5+ Kg4–+)]
47.Rh1 Ra7 48.Kb2 b4!–+ and Black breaks through on the queenside.
b) 46.Rd8?! would have left the game in balance: 46...Bxg4 47.Rd6+ Be6 48.Ra8 Rg1!³

49.Kd2?? Rd7 (49...Ke7 50.Bh5=; 49...Rf1?! allows 50.Bg4²) 50.Rxd7 Bxd7 51.Rf8+ Kg7 52.Rd8 Bh3 53.Rd6 Rf1
54.Ke2 b4–+
c) 46.Rf8+!? would have forced Black to find only moves to win the game: 46...Rf7

224
47.Rh8! Exchanging rooks would be a mistake.
c1) 47.Rxf7+? Kxf7! 48.Ra7+ Ke8!! 49.Kd2 Bxg4 50.Bxg4 Rxg4 51.Rc7 Rg6 52.Ke2 Re6 53.Rh7 Kf8 54.Rc7 Kg8
55.Kf3 h5–+
c2) 47.Rd8!? Kg5 48.Rh1 (48.Rd6 Bd7 49.Rh1 Rxd6 50.Rh5+ Kg6 51.cxd6 Rf6 52.Rxe5 Rxd6µ) 48...Ra7! 49.Kb2
b4! Exploiting the poor location of White’s f3-bishop. 50.cxb4 c3+ 51.Kb1 Ra2µ
47...Rfg7 48.Rf8+ Rf7 49.Rh8 Kg5!

50.Re8 (50.Rh1? allows Black to claim the a-file. 50...Ra7!) 50...Rfg7!!µ White is in zugzwang again.

46...Ra7!–+

225
Now it’s over. Black is winning, thanks mainly to White’s f3-bishop.

47.Kb2 b4?!

Rushed.
47...Kg7!! would have eliminated White’s last hopes. 48.g5 (48.Bd1 f3! 49.Bxf3 Rf7 50.Rh3 h5!! A finesse that
wins the game. I obviously didn’t see all these lines in the game, but they are too pretty to skimp over.) 48...hxg5
49.Bh5 (49.Reh8 Re7–+) 49...Rh6–+

48.g5+!

My opponent finds resourceful moves that I didn’t think could work.

48...hxg5 49.Rxe6+?

White finally succumbs to the pressure.

49.Bh5!

226
would have posed difficult questions for Black, who must now play very precisely to secure victory. 49...bxc3+
a) 49...Rh6?? is obviously impossible. 50.Rxe6+ Kxe6 51.Bg4++– and it is White who wins.
b) 49...Rg8 also wins for Black, however, the path to victory is much more difficult. 50.cxb4 c3+ 51.Kxc3 Ra3+
52.Kd2 Ra2+ 53.Kd3 Rxf2 54.Rxg8 Bxg8

55.Bd1!? Kg6 56.b5 cxb5 57.c6 Be6 58.Rh8 b4!! 59.c7 g4 60.Re8 (60.c8=Q Bxc8 61.Rxc8 g3–+) 60...Bd7! and the
threat of mate on b5 forces White to give up the chase. 61.Rxe5 g3 62.Rd5 Bc8 63.Rd8

227
63...Ba6+ (63...Bb7 64.c8=Q Bxc8 65.Rxc8 g2 66.Kc4 Kg7!!–+ 67.Rc7+ Kf6 68.Rc6+ Ke5 69.Rg6 Kxe4 70.Bh5
f3–+) 64.Kd4 g2 65.Ke5 f3 66.Bxf3 Rxf3 67.Rg8+ Kf7 68.Rxg2 b3 69.Kd6 Bc8 70.Rd2!µ and questions remain if
Black can win this!!
c) 49...Bf7 At first, this seems the easiest and best solution, or at least I thought so during the game. However White
has the following idea: 50.Rc8! (50.Bxg6 bxc3+ 51.Kxc3 Ra3+ 52.Kb2 Rb3+ 53.Kc2 Bxg6 54.Rh6 Kg7 55.Rxg6+
Kxg6–+) 50...b3

51.Bg4! White’s control of the light squares with his bishop is more important than Black’s g6-rook.
[51.Rxc6+ Kg7 52.Bxg6 Bxg6 (52...Ra2+? 53.Kb1 Bxg6 54.Rc7+ Kg8 55.f3 g4 56.Rd1 and it is White who wins!!
56...Ra8 57.Rd6+–) 53.Ra1!? (53.Re1 g4 54.Rd6 g3 55.fxg3 f3–+) 53...Rxa1 54.Kxa1

228
54...g4! (54...Bxe4 55.Re6 with advantage to White) 55.Rd6 Bxe4 56.c6 Bxc6 57.Rxc6 g3 58.fxg3 f3!–+ wins for
Black]
51...Kg7

52.f3 (52.Rch8 Ra2+ 53.Kb1 Bg8 54.f3 Rc2 55.Bf5 Rf6 56.R8h5 Rxf5 57.exf5 Rg2!–+ The dancing rook has been
working like a magician in this game and now it puts the finishing touch to Black’s victory.) 52...Ra2+ 53.Kb1 Rf6
54.Rc7 Rc2!–+
50.Kxc3 Rg8 51.Rxg8 Bxg8 52.Bg4 Kg7

229
and Black should win. Some sample lines: 53.Rd1 Ra3+ 54.Kc2 Ra2+ 55.Kc3 Rxf2 56.Rd7+ Kf8 57.Rd8+ Kf7
58.Rd7+ Ke8 59.Rg7 Kf8 60.Rxg5 Rg2 61.Rf5+ Bf7 62.Bh5

62...Rg3+! 63.Kd2 Rd3+ 64.Kc2 Rd7 65.Bg4 Re7 66.Rf6 Ra7 67.Rxc6 Ra2+ 68.Kc3 Ra3+ 69.Kc2 Rg3 70.Be6 f3
71.Rc8+ Ke7 72.Bxf7 Kxf7 73.Rc7+ Ke8 74.Rc8+ Ke7 75.Rc7+ Kf6 76.Rc8 Rg7 77.c6 Rf7 78.Rd8 f2 79.Rd1 Ke6
80.Rf1 Rf3 81.c7 Kd7 82.Kd2 Kxc7 83.Ke2 Rf4–+

49...Kxe6 50.Bg4+ Kf6 51.cxb4

230
51...Kg7!

Black’s second rook joins in the game and with it White’s fate is sealed.

52.Rh5 Rh6 53.Rxg5+ Rg6 54.Rxg6+ Kxg6 55.f3 Kf6 56.Kc3 Ke7 57.Kxc4 Ra3!

Zugzwang. White is forced to watch Black bring his king over to the queenside.

58.Bh5 Kd7 59.Bg4+ Kc7 60.Bh5 Kb7 61.Bg4 Ra1 White resigned. 0–1

61...Re3 wins too:

231
62.b5 (62.Bh5 Ra3 63.Bg4 Ra1 64.Be6 Rd1 65.Kc3 Rf1 66.Bg4 Ka6 67.Kc4 Rc1+ 68.Kb3 Kb5 69.Be6 Rb1+
70.Kc3 Rxb4–+) 62...cxb5+ 63.Kxb5 Rc3 64.Bh5 Kc7 65.Bg4 Rc2 66.Bh5 Rh2 67.Bg4 Rb2+ 68.Kc4 Kc6 69.Bh5
Rc2+ 70.Kb4 Rxc5–+
The game started slowly but then developed into a highly nuanced endgame. I was particularly impressed with how
White managed to outplay me in the endgame at first, activating both rooks, but shortage of time and the constant
need for vigilance took their toll and he started to make mistakes.

232
 GAME 9
Gata Kamsky
Alexei Shirov
Dortmund 1992
This game is included because it was played during an unofficial rivalry between Alexei and me in tournaments back
in the ’90s. Separated in age by only couple of years, we were both considered serious future contenders for the
world crown. Furthermore, the participation of the World Champion, Mr. Kasparov, served as additional motivation
to ‘show’ yourself. I was particularly pleased with the way this game was decided. Quite unlike my usual style, I
decided to use my opponent’s style against him. Many years later, I realized of course how deeply Lasker’s
techniques were embedded in my chess, but in my younger days a lot of decisions were made on the spur of the
moment.

1.d4 d5 2.c4 c6 3.Nf3 Nf6 4.Nc3 e6

Shirov has always been considered one of the world’s leading experts on the Botvinnik line of the Slav, which suited
his style perfectly. I preferred to sidestep a theoretical discussion, owing to my lazy approach to the openings.

5.cxd5 exd5 6.Bg5 Be7

These days people are no longer afraid to play 6...Bf5 immediately, followed by 7.Qb3 Qb6. The endgame turns out
to be quite playable for Black, despite his chronic doubled f-pawns.

7.Qc2 g6 8.e4

233
Now we enter the line which was quite popular in those days. I felt that the endgame with extra space and more
active pieces was exactly what I needed vs. my fiery opponent, but he had different plans.

8...0-0?!

Really risky. Black is gambling with this move.

a) 8...Nxe4? is possible but less desirable. 9.Bxe7! The point. 9...Kxe7 (9...Qxe7?? 10.Nxd5! And suddenly it’s
game over.) 10.Nxe4 dxe4 11.Qxe4+ Be6 12.Bc4

12...Re8 (12...Qa5+? 13.Kf1! Qf5 14.Qe3 Nd7 15.Re1 Rae8 16.d5! cxd5 17.Bb5± White has a huge attack.) 13.Bxe6
Kf8 14.0-0 Rxe6 15.Qf4 Kg7 Black is OK. (Not to be confused with András Adorján’s famous book of that name.)
b) 8...dxe4 is the main line and a better move. 9.Bxf6 Bxf6 10.Qxe4+

234
10...Kf8 (or 10...Qe7) 11.Bc4 Kg7 12.0-0 Nd7 13.Bb3 Nb6 14.Qf4 h5 15.Rfe1 Bf5

Now a position arises with equal chances. White has some squares for his knights, but Black has no weaknesses and
potentially the superior endgame due to his better structure.

9.e5

235
A logical reaction. Black had probably counted on his lead in development and the fast ...f7–f6 central pawn break,
or possibly the ...c6–c5 break, leading to a favorable game, but White has additional resources.

9...Ne4

In view of what happened later it was more prudent to set up a defense around Black’s king, e.g. 9...Ne8 10.Bh6
Ng7 11.h4 Bg4 12.Bd3 Nd7, although White’s chances are definitely superior due to his extra space in the center
and strong potential attack on the kingside.

10.Bh6 Re8

11.Bd3

This turned out to be a new move at the time. I didn’t know it then, but there was a game played by former World
Champion M. Tal, who was an occasional trainer of my opponent. The move is natural of course, and seems to force
236
Black to exchange his e4-knight, reinforcing White’s center after b2xc3. However...
11.h3 Nxc3 12.bxc3 c5 13.Bb5 Nc6 14.0-0 Bf5 15.Qd2 Qa5 16.Bxc6 bxc6 17.dxc5 Qxc5 18.Nd4 Bd7= Tal,M-
Renet,O Cannes 1989.

11...Nxc3?!

a) 11...Na6!? is an interesting recommendation by the computer. Black seems to win an additional tempo for
development. If he somehow manages to open the center, his queenside majority might make life difficult for White.
12.a3

12...f5
[12...Bg4!? is a more active possibility. Black has the superior development and White has to be careful. The
following line shows how easy it is to err: 13.Bxe4! (13.Nxe4? dxe4 14.Bxe4 Bxf3 15.Bxf3 Qxd4³ and White’s
whole center collapses.) 13...dxe4 14.Qxe4 f5 15.Qd3! Bxf3 (15...Nc7 16.Rd1 Qd7 17.h3 Bh5 18.Rd2 Rad8 19.0-0ƒ
with the better pawn structure for White) 16.gxf3 Nc7 17.Qc4+ Nd5 18.Rc1 with an extra pawn, but breaking
through Black’s solid defense remains a daunting task. 18...Rc8 19.h4 Qd7 20.Qb3 Kh8

237
21.Nxd5 Qxd5 (21...cxd5 22.Kf1 Rxc1+ 23.Bxc1 h5 24.Kg2 Kg8 25.Bf4 Rc8 26.Qe3 Kf7 27.Bg5 Bf8 28.Bf6 Rc2
29.Re1 Ke6! 30.Qg5 Qf7 and it is far from easy for White to convert his advantage) 22.Qxd5 cxd5 23.Ke2 Kg8
24.h5 Kf7 25.hxg6+ hxg6 26.f4 Ke6 27.Bg5±]
13.h4 Nc7 14.Rd1 Ne6

15.Be3!
[15.Kf1 b6 16.Kg1 a5 17.h5 (17.Ne2 c5 18.Qc1 Ba6 and Black is completely fine) 17...g5∞; 15.h5 g5 16.g4 fxg4 is
nothing for White.]
15...Kh8!? Vacates the g8-square for the rook, thus preparing for White’s coming h4–h5. (15...Ng7 16.g3 b6 17.0-0
a5 18.Ne2 Qd7 19.Ne1 Nh5 20.Ng2 Ba6 21.Bxa6 Rxa6 22.Ngf4 White finally manages to encircle black’s e4-
knight, and threatens to capture it with f2–f3. 15...b6 16.h5 g5 17.g4! A visually pleasing move that undermines
Black’s e4-knight.) 16.Kf1 (16.h5 g5 17.g4 fxg4 18.Nh2 Nf4 19.Nxe4 Nxd3+ 20.Qxd3 dxe4 21.Qxe4 Be6 22.Rg1
Rf8 23.Nxg4 Bd5 with counterplay) 16...Bd7 17.g3 a5 18.Kg2 Rc8 19.Qb3 b5 20.Ne2

238
20...Qb6 (20...a4 21.Qa2 Rg8 22.Nf4 Nxf4+ 23.Bxf4 Be6 24.Be3 Qb6²) 21.Nf4 Nxf4+ 22.Bxf4 c5 23.dxc5 Nxc5
(23...Bxc5 24.Bxe4 fxe4 25.Ng5‚) 24.Qa2 Ne4 25.Rhf1 Be6 with complex play.
b) 11...f5!? This is another interesting possibility. Black tries to hold on to his central e4-knight for as long as
possible, because it hinders White’s advance on the kingside. In addition, Black needs to prepare his own central
pawn push ...c6–c5.

The play now revolves around White’s plans to capture Black’s e4-knight with f2–f3, while trying to prevent Black
from preparing the ...c5 break and liberating the e4-knight. The position is complex, but White should have the
superior chances due to his greater space for maneuvering. 12.h4
[12.0-0 Nd7! (12...Be6 13.Ne2 g5 14.Qc1 f4 15.h4!+–, breaking the defensive pawn chain) 13.g3 Nf8 with decent
play for Black, e.g. 14.h4 Ne6 15.Rfd1 Bd7 16.Ne2 c5 17.dxc5 Bxc5 18.Bxe4 fxe4 19.Ng5 Rc8 20.Qd2 Bc6 21.b4
Nxg5 22.Bxg5 Be7 23.Bh6 d4 24.Nxd4 Ba4 25.Rdc1 Qd5 26.Nf5 Rxc1+ 27.Rxc1 Qxd2 28.Nxe7+ Rxe7 29.Bxd2
Bc6!=]
12...Be6 13.Rd1! Very strong. White leaves his king in the center, as castling long would put it in a possible danger
zone. The king will find a better hiding place on g2. 13...Na6 14.Qc1 Bf7 15.h5 Nc7 16.hxg6 Bxg6 17.Kf1 Ne6
239
And now White has two different plans to improve his position: 18.Kg1!? This is often useful as a prophylactic
move, removing the king from all possible tactics after a future ...c6–c5 and leaving the f-file open for potential
action.
[18.g3!? Rc8 19.Kg2 c5 (19...b5 20.Ne2 c5 21.Nf4 Nxf4+ 22.Qxf4 c4 23.Bb1 b4 24.Nh4 Qd7 25.f3 White captures
black’s e4-knight.) 20.dxc5! Bxc5 (20...Rxc5 21.Be3 Rc7±) 21.Bxe4! fxe4 22.Nh4‚ White manages to open the
board on all fronts with very strong threats against the black king.]
18...Rc8
[18...Kh8 19.Ne2 Rg8 20.g3 Rc8 21.Nf4 Qd7 22.Nh4 Nxf4 23.Qxf4 Qe6 24.Nf3 Bf7 25.Bxe4 fxe4 26.Ng5 Bxg5
27.Bxg5 Rg6 (27...h5 28.Rh4 Kh7 29.Kg2 Rg7 30.Rdh1 Kg8 31.g4!+–) 28.Qh4±]
19.g3 c5 20.Be3!

White needs to retain control of the d4-square as it is integral to his advantage and further plans for play on the
kingside. 20...c4?!
[20...a6 21.Kg2 b5 22.dxc5 Bxc5 23.Bxe4 fxe4 24.Nh4 Bxe3 (24...d4 25.Nxg6 hxg6 26.Bxd4! Bxd4 27.Qh6 Kf7
240
28.Qh7+ Ng7 29.Rh6!! Rc6 30.Ne2) 25.Qxe3 d4 26.Qh6 Qg5 27.Qxg5 Nxg5 28.Rxd4 Rxe5 29.Nxg6 hxg6 30.Rd6
Re6 31.Rxe6 Nxe6 32.Nxe4 Rc2 33.Rd1 Rxb2 34.Rd6 Kf7 35.Rxa6± Somehow White has finished only a pawn up,
but Black still has to suffer for a long time before he can reach a draw.]
21.Be2 Qd7 22.Nh4

22...Kf7 (22...b5 23.Nxg6 hxg6 24.Nxe4 fxe4 25.Bg4 Kf7 26.f3 exf3 27.Rf1 Rf8 28.Rxf3+ Ke8 29.Qf1! Rc6 30.Rh7
Rxf3 31.Qxf3 Kd8 32.Qf7+– and White has finally infiltrated on the kingside.) 23.Nxe4 fxe4 24.Nxg6 hxg6 25.f3!‚
The kingside is finally open and White’s pair of bishops will be monsters in the coming battle.

12.bxc3

12...c5

12...Bg4 would probably be a waste of time since White would not fear doubled f-pawns, in fact the potential f4–f5
break would only make things better for him! 13.h4!? Bxf3 14.gxf3 Bxh4 15.f4 with a decisive attack.
241
13.h4!

Correct, otherwise Black develops and equalizes.

Now White will have to enter complications, although they are favorable for him.

13...cxd4?

Perhaps the first major and only mistake of the game.

a) 13...Bg4? 14.h5 Bxh5 15.Rxh5 gxh5 16.Bxh7+ Kh8 17.Qf5 Qd7 18.Qxf7 Bd8 19.Qg6+–
b) 13...Nc6 14.h5 c4 15.Bxg6 fxg6 16.hxg6‚ with the idea 16...hxg6 17.Qg6.
c) 13...c4! It was absolutely critical to remove’s White’s d3-bishop from its attacking position. 14.Be2 Bf5 (14...Nc6
15.h5 Bf5 16.Qc1 b5 17.Ng5 f6 18.g4 fxg5 19.gxf5 gxf5 20.Rg1 Kh8 21.Rxg5!‚) 15.Qd2

15...b5!
[15...Nd7?! Too timid. Black has to create counterplay or risk being overrun on the kingside. 16.h5 Nf8 17.Nh2!!

242
An absolutely amazing resource which proves the adage that sometimes in order to go forward, one has to step back
first! Black’s f5-bishop is the key to holding his position, so removing this bishop is White’s top priority, as we have
seen in earlier lines. 17...Qd7 18.Nf1! The knight retreats even further, in order to be transferred to the key e3-
square, where it will attack Black’s d5-pawn and finally help White with his main idea, the f2–f4–f5 pawn push,
opening up Black’s kingside defense. 18...b5 19.hxg6

19...Bxg6 (19...fxg6 20.a3! A very important move, the reason for which will be shown later. 20...a5 21.Ne3 b4
22.g4 Be6 23.f4 bxc3 24.Qxc3 Reb8 25.Bd1 Ra6 26.0-0+– White’s queen, while not best known for its role as
blockader of a pawn, nevertheless stands well on the c3-square, and it will take a while before it can be dislodged. In
addition, White’s pawn mass is threatening to obliterate Black.) 20.Ne3± White is ready for his attack and Black is
in big trouble.]
16.a3 (16.h5? b4 17.hxg6 Bxg6 18.Kf1 bxc3 19.Qxc3 Nc6 and it is Black who is fighting for an advantage.) 16...a5
17.h5 Nc6 18.hxg6 Bxg6 19.Qf4 b4

243
20.Ng5!
[20.Nh4?! Every tempo is important. This move is too slow. 20...Bxh4! 21.Rxh4 b3 Black is better now. 22.Bh5!?
(22.Bg5? doesn’t work: 22...b2 23.Rd1 Qb6 24.Bf6 Nxd4! 25.cxd4 b1=Q 26.Qh6 Qxd1+ 27.Kxd1 — 27.Bxd1 Qxf6
— 27...Qb1+ 28.Kd2 c3+ 29.Ke3 Qc1+ 30.Kf3 Qxh6–+) 22...Qe7 23.Bg5 b2 24.Rd1 Qxa3 25.Bf6 b1=Q 26.Rxb1
Qxc3+ 27.Kf1 Qd3+ 28.Kg1 Qxb1+ 29.Kh2

29...Re6!! and Black is just in time. 30.Qh6 ( 30.Bxg6 fxg6 31.Rxh7 Kxh7 32.Qh4+ Kg8 33.Qh8+ Kf7 34.Qg7+
Ke8) 30...Rxf6 31.exf6 Qb8+ 32.g3 Qf8–+]
20...Qb6 (20...b3 21.Bh5! This move helps to eliminate the key defender of the black position, the g6-bishop.)
21.Bh5 Nd8

244
The point behind ...Qb6 — Black’s queen is transferring to the kingside!! 22.Nxh7! White’s attack provides
sufficient counterplay to balance Black’s passer on the queenside.
[22.axb4 Bxg5 23.Qxg5 Ne6 24.Qf6 axb4 25.Rxa8 Rxa8 26.Bxg6 fxg6 27.0-0 b3 28.f4 Ra7 29.f5 Rf7

30.Qxf7+ Kxf7 31.fxe6+! Ke8 (31...Ke7? 32.Bf8+ Ke8 33.Bd6 wins instantly) 32.Rf8+ Ke7 33.Rf7+ Ke8 with a
typical crazy computer draw!]
22...bxc3 (22...Bxh7 23.axb4 Bg6 24.Qg3=) 23.Bxg6 (23.Qg3? Kxh7 24.Bf3 Rh8 25.Bf8+ Kg8 26.Rxh8+ Kxh8
27.Qh3+ Kg8 28.Bxe7 Qb2 29.Bf6 Qxa1+ 30.Bd1 Bh7 31.Qg3+ Kf8 32.Qg7+ Ke8 33.Qh8+ Kd7 34.Qf8 Nb7!
35.Qxf7+ Kc6 36.Qxh7 Qxa3 37.Be7 Qc1 38.Qg6+ Kb5 39.Bg5 Qb2 and Black is better) 23...Qxg6

245
24.Bg5! The game now erupts in wild fireworks where attack and defense and counter-attack will all balance each
other out in the end! 24...Bxg5 (24...Ne6? 25.Nf6+ Bxf6 26.Bxf6+–) 25.Nxg5 Rb8 26.Rc1 Rb1 27.Rh3! c2 28.Rg3
Rxc1+ 29.Qxc1 Kf8!

30.Nxf7 Qe4+ 31.Re3 Qxg2 32.Nxd8 Qg1+ 33.Kd2 Qxf2+ 34.Kc3 Rxd8 35.Rh3 Kf7 36.Qxc2 Qxc2+ 37.Kxc2
Rg8= and after the fireworks we have an exciting equal rook endgame!

14.h5!!‚

246
White ignores the destruction of his center and focuses entirely on the black king. All my pieces are in play and
positioned perfectly for the attack, and I judged that Black’s under-developed queenside would not be in time to
protect his king. In such positions, every tempo is critical.
a) 14.cxd4? Nc6 and now it is only Black who can be better.
b) 14.Nxd4 Nc6 15.Nxc6 bxc6 16.h5 Bg5 and again Black is better.

14...g5?!

a) 14...Qb6 15.Rb1± (15.hxg6 hxg6 16.Rb1 Qe6 17.cxd4±)

b) 14...Nc6 15.hxg6 hxg6 16.Bxg6+–
c) 14...Qa5 15.hxg6 hxg6 16.Nxd4±
d) 14...Qc7

This move is the computer’s first line of defense, however the resulting endgame is still better for White. 15.hxg6
247
hxg6
[15...Qxc3+ 16.Qxc3 dxc3 17.gxf7+ Kxf7 18.Be3 Rh8 19.Bd4 Be6 20.Bxc3 Nc6 21.Rh6 Rag8 22.Rb1! A very
important sneaky move to break down Black’s defense; the protection of the black c6-knight must be undermined
first.

22...b6 23.Kf1 Rg4 24.Rc1! And now it becomes clear why it was important to move Black’s b7-pawn, as the c6-
knight is no longer protected and White is threatening to penetrate on the seventh rank with Rc1–c7. This forces
Black to remove the guard of the critical g5-square, making White’s game practically winning. 24...Bc5 25.Bd2!
Ke7 26.Bg5++– White transfers his bishop to the perfect f6-outpost, netting a pawn and potentially helping with his
kingside pawn advance.]
16.Bd2!! Kg7 17.Ke2

17...Bg4 (17...Bf6 18.Kf1 Bxe5 19.Bh6+ Kg8 20.Bxg6 Qc4+ 21.Kg1 d3 22.Bxd3 Qxc3 23.Bh7+ Kh8 24.Re1+–)
18.Bh6+ Kg8 19.Bxg6 Qxe5+ 20.Kf1 with a devastating attack.

15.Bxh7+ Kh8
248
16.Bg6!!

Very accurate. There’s no time to recapture the d4-pawn yet.

a) 16.cxd4 Nc6, idea ...g4 with counterplay
b) 16.Nxd4? Nc6!³ A perfect example of why thinking about material sometimes backfires. Black is better now.
(16...Qb6 17.Bg6 fxg6 18.hxg6 Kg8∞)

16...Be6™

16...fxg6 17.hxg6 Kg8 18.Bxg5! Bxg5 19.g7+– and White has got to his prize!

17.Nxd4 Qc8

17...Bf8 18.Bxf8 Rxf8 19.Qd2! This move is a thematic idea in this position, as it prepares another key idea where
White retreats his bishop to the c2-square and follows with Qd2–d3, preparing the long range battery against the
black king.

18.Qd2!!

249
An absolutely key idea which White had to foresee before entering this whole line starting with h4–h5. White needs
to eliminate any defenders on the kingside, the primary target being Black’s g5-pawn.

18...fxg6

18...Rg8 19.Bxg5 Bxg5 20.Qxg5+– Qxc3+ 21.Ke2 Qc4+ 22.Kf3 Qc3+ 23.Qe3 Qxe3+ 24.Kxe3 Nc6 25.Nxe6 fxe6
26.f4 with a winning endgame.

19.hxg6 Kg8 20.Rc1?!

During the game I thought this was a brilliant move, demonstrating that White doesn’t have to hurry after all the
tactical fireworks. However, it’s not the top choice. Luckily for me, this inaccuracy doesn’t prevent White from
winning the game.
a) 20.Bxg5 Bxg5 21.Qxg5 Qxc3+ 22.Ke2 Qxd4 23.Qh6 Qb2+!
b) 20.Kf1! This move was simpler and stronger. White defends against a possible ...Qxc3 with check and prepares
for his a1-rook to go directly to e1. 20...Nc6 21.Bxg5 Nxe5 22.Bxe7 Rxe7 23.Qh6, winning.

20...Nc6 21.Bxg5 Qc7

21...Nxe5 22.Bxe7 Rxe7 23.Qh6+–

22.Bxe7

250
22...Rxe7

22...Qxe5+ 23.Kf1 Rxe7 24.Re1 Qf6 25.Rxe6 Rxe6 26.Nxe6 Qxe6 And now both moves win for White: 27.Rh8+
(27.Qh6 Qf6 28.Qh7+ Kf8 29.Qh8++–) 27...Kxh8 28.Qh6+ Kg8 29.Qh7+ Kf8 30.g7++–.

23.Kf1!

The only move to win, but sufficient. Black is helpless against White’s threats on the h-file. The game is won.

23...Rae8 24.Re1!+–

A final demonstration of how strong White’s position is. He can allow such quiet strengthening moves.
24.Qh6 Qxe5 25.Nxe6+–

24...Rg7
251
24...Nxe5 25.Qh6+–

25.Qh6 Kf8 26.Rh4

26.Qf4+ was stronger: 26...Kg8 27.Nxe6 Rxe6 28.Rh8+ Kxh8 29.Qf8+ Rg8 30.Qh6+ Qh7 31.Qxh7#.

26...Nxd4 27.Qh8+ Rg8

27...Bg8 28.Rf4++–

28.Qf6+ Bf7 29.Qxf7+ Qxf7 30.gxf7 Kxf7 31.Rxd4+–

and Black resigned as White’s two extra pawns ensure a technical win.

1–0

It’s strange how after such a sacrificial attack, White won by transposing into a technical endgame, but it’s just
another example of the law of physics which says that energy is never lost, it just changes into another state.

252
 GAME 10
Gata Kamsky
Garry Kasparov
Dortmund 1992
While the tournament overall was not a success for me, this game was, and remains, one of my most memorable. A
victory over a reigning World Champion who lost very few games in his career is always a highlight!

1.d4 Nf6 2.c4!

The exclamation mark is for the surprise value. I didn’t play main lines much and my decision to go into a main line
King’s Indian was influenced in large part by the fashionable line that occurred in the game. White scored well with
it and I was curious to see how the World Champion would react.

2...g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 0-0 6.Be2 e5 7.0-0 Nc6 8.d5 Ne7 9.Nd2

This main line was very popular at the time. Modern main lines of course focus on 9.Ne1, after which theory goes
on into the 30s. I’m not going to go into detail, because it was all well-known and heavily played in those days.

9...a5 10.a3 Nd7 11.Rb1 f5 12.b4 Kh8

This move prepares the transfer of the awkward e7-knight to a better location. Black leaves the d7-knight in place to
guard against a potential c4–c5 break.

13.f3 Ng8 14.Qc2 Ngf6 15.Nb5 axb4 16.axb4 Nh5 17.g3 Ndf6 18.c5 Bd7 19.Rb3 Bh6

19...Nxg3?!²

20.Rc3

253
We have reached a topical position. I was very nervous but didn’t see how White could lose this. Of course my
opponent’s famed and feared preparation was a serious cause of concern for me. I decided that I would try my best
and, if I was going to lose, I would lose in glory.

20...Bf4!?

A new move at the time.

Previously 20...fxe4 had been played on several occasions, but it seems that this exchange favors White as the f3-
square is opened up for the d2-knight.
My opponent played the text move quite quickly and his typical grimacing made it clear that it had all been analyzed
and that it was a pity to waste such analysis on a player like me. Mr. Kasparov’s facial contortions were famous and
quite annoying, but since no rulebook exists for things like this, players were pretty much forced to tolerate such
behavior.
20...fxe4 21.fxe4 Bh3 22.Re1
[22.Rd1 Ng4 23.Nf3 Ra1 24.Bxh6 Rxd1+ 25.Qxd1 Nxh6 26.cxd6 cxd6 27.Nc7

254
27...Nf7 (27...Qe7 28.Ne6 Bxe6 29.dxe6 Qxe6 30.Rc7 White has a slight pull, due to his strong e2-bishop in an open
position and Black’s awkwardly placed knights.) 28.Bf1 (28.Ne6?! Qb6+ 29.Kh1 Nf6!! 30.Nfg5 Nxg5 31.Nxg5
Bd7³ Black is better since his queen is really well placed on b6, while White’s king is quite weak.) 28...Bd7
(28...Bg4?! This move looks natural, but is not the best, since this bishop will be cut off from the queenside. 29.Ne6
Qb6+ 30.Kg2 Re8 31.Qa4 Nf6 32.Qa5!± White eliminates Black’s best piece, his queen, leading to an advantage for
White.) 29.Ne6 Qb6+ 30.Kg2 Ra8 31.Nd2!?

An interesting computer idea. The following line gives an example of how murky this position is. 31...Qxb4
(31...Ra2 32.Bc4 Ra8 33.Rf3!ƒ The sudden transfer of the rook to the kingside makes White’s attack dangerous.
33...Kg8 34.Rf2²) 32.Rb3 Qa4 33.Be2 Bxe6 (33...Nf6?? 34.Qf1+–) 34.dxe6 Ng5 35.Nc4 Qc6 36.Nxd6 Nxe6
37.Bxh5 Nd4 38.Nf7+ Kg7 39.Nxe5 Qxe4+ 40.Bf3 Ra2+ 41.Kh3 Qxe5 42.Rxb7+

255
42...Kf6 (42...Kh6 43.Qc1+ g5 44.Bg4± A curious position now arises with both kings on the h-file. Despite Black’s
ominous pieces in the center, White’s king is safer!) 43.Rb6+= leading to a perpetual since Black’s king cannot risk
avoiding the checks.]

22...Qd7!? (22...Ng4 seems a logical follow-up to Black’s previous move. 23.Nf3 Ra1 24.Bxh6 Rxe1+ 25.Nxe1
Nxh6 26.Nf3 Nf7! and I think that Black’s position is actually preferable, thanks to the numerous dark-squared
weaknesses around the white king.) 23.Nf3
[23.cxd6 cxd6 (23...c6!? 24.dxc6 bxc6 25.Nc7 Qxd6 26.Nf3 Bxc1 27.Nxa8 Bh6 28.Nb6 Qxb4 29.Rb3

256
29...Nxe4!! 30.Rxb4 Be3+ 31.Kh1 Nf2+ 32.Kg1 Nd3+= leads to an exciting draw by perpetual) 24.Nf3 (24.Rc7??
Be3+ 25.Kh1 Rac8!–+) 24...Nf4!? leads to massive complications]
23...Bxc1 24.Rxc1 Nf4 25.Ng5?!² (25.Bc4!? N4h5 26.Ng5±; 25.cxd6 cxd6 26.Ng5 Nxe2+ 27.Qxe2 Bg4 28.Qd3
Ne8 29.Nc7 Rc8 30.Nce6± with advantage to White, thanks to his very strong e6-knight) 25...Nxe2+ 26.Qxe2 Bg4
27.Qc4 Ne8∞ 0–1 (40) Epishin,V (2615)-Polgar,J (2550) Vienna 1991.

21.cxd6!?

I wanted to play something that my opponent might have overlooked in his preparation. It was obvious that 21.gxf4
or any move protecting g3 would have been heavily analyzed.
a) 21.gxf4!? Nxf4

22.Bc4
[22.Nb3 N6xd5! 23.exd5 Qg5+ 24.Kf2 Qg2+ 25.Ke1 Qxe2+ 26.Qxe2 Nxe2 27.Nxc7 (27.Kxe2 Bxb5+ 28.Ke1 Bxf1
29.Kxf1 Ra2µ) 27...Nxc3 28.Nxa8 Rxa8 29.cxd6 Nxd5 30.Bb2 Nxb4 31.Bxe5+ Kg8 32.Nc5 Bb5³]
257
22...N6h5 The most natural choice, which I thought was the course my opponent’s preparation would have taken.
This line leads to a draw, but White has to be careful.
[22...Bxb5?! A computer suggestion, but illogical. Despite all the advances in the past decade, computers still have
difficulty evaluating King’s Indian positions. 23.Bxb5 N6xd5 24.exd5 Qg5+ 25.Kf2 Qh4+ 26.Ke3±

White’s king has escaped with extra material and the much better position. Black is ruined.
26...Qxh2 (26...Nxd5+ 27.Ke2 c6 28.Bc4 Qxh2+ 29.Rf2 Nf4+ 30.Ke1 Qh4 31.Nb3 winning) 27.Bc4 dxc5 28.bxc5
Rad8 29.Nb3 Qh3 30.Kf2+–]
23.Nb3 fxe4

24.fxe4 (24.Bxf4 Rxf4 25.fxe4 Qg5+ 26.Kh1 Rxf1+ 27.Bxf1 Rf8 28.Bc4 Qh4‚) 24...Qh4 (24...Nh3+ 25.Kh1 Qh4
26.Nxc7 Rxf1+ 27.Bxf1 Rf8 28.Qe2 Rf2 29.Qd3 Rxh2+ 30.Kxh2 Qf2+ 31.Bg2 Qg1#) 25.Kh1 Nh3 26.Qg2 Rxf1+
27.Bxf1 Rf8 28.Rf3 Nf2+ 29.Kg1 Nh3+

258
30.Rxh3 Bxh3 31.Qxh3 Qf2+ 32.Kh1 Qe1 (32...Nf4 33.Be3 Qe1 34.Bxf4 Rxf4 35.Kg1 Qf2+ 36.Kh1 Qe1=)
33.Kg1™ 33...Qf2+ with a draw by repetition.
b) 21.Rf2!? is another intriguing possibility, but let’s leave the analysis of this position to the theorists.

21...Nxg3?

Suddenly Black errs.

a) 21...cxd6!

This simple recapture is, of course, the main line, leading to a more or less equal position. Obviously my opponent
wanted more from the game and the psychological aspect is important. As I would play more and more games in
open tournaments where my opponents would be much lower rated and fear me, I would be faced with very solid
and rather passive play for a draw. It seems that I adopted the same strategy in this game against the World
Champion, baiting him to take extra risks to try to win. 22.Nc7 (22.Nxd6? This pawn is untouchable. 22...Ba4
23.Qb1 Bxd2 24.Bxd2 Qxd6µ with an extra piece for Black) 22...Rc8 23.Ne6 Bxe6!

259
[23...Qb6+?! This active-looking move is not great, as the queen on b6 helps White develop his Nd2 with tempo.
24.Kg2

24...Rxc3 (24...Qxb4? 25.Nc4 Bxe6 26.Ba3 Qb5 27.dxe6 Bh6 28.Bxd6+–) 25.Qxc3 Bxe6 26.Nc4 Qc7 27.gxf4
White’s two bishops, especially the dark-squared one, are becoming monsters as the center is blown wide open.
27...Bxd5 28.fxe5 fxe4 29.exf6±]
24.dxe6

24...Be3+! A very important intermezzo. Black has achieved one of the ideal locations for the KID bishop. 25.Kg2
Rxc3 26.Qxc3 f4 The position is very dynamic, but Black has no bad pieces and White’s king is open. These factors
compensate for the deficiencies in Black’s structure. 27.Nb3 Bxc1 28.Rxc1 fxg3 29.hxg3

260
29...Nxg3! with an immediate draw.
[29...Qe7 is also possible: 30.Qc7 Qxe6 31.Bc4 d5 32.Nc5 Qc8 33.Qxc8 Rxc8 34.exd5 b6 35.Nd3 e4 36.Bb3 Rxc1
37.Nxc1 Ng7! 38.fxe4 (38.d6 exf3+ 39.Kxf3 Nf5=) 38...Nxe4 39.Kf3 Nd6 40.Kf4 Kg8 41.Ke5 Nf7+

42.Kf6 Nh5+ 43.Ke6 Nxg3 44.d6 Nf5 45.d7 Kg7 46.Ne2 Nd8+ 47.Ke5 Nc6+ 48.Kd5 Nd8 49.Ke5 with a draw,
since White cannot let the black king get closer to the d7-pawn.]
30.Kxg3 Nh5+ 31.Kf2 Qh4+ 32.Ke3 Qg5+ 33.Kf2 Qg3+ 34.Ke3 Qg5+= A draw by repetition is unavoidable.
b) 21...Bxb5?! 22.dxc7 Qd7 23.Bxb5 Qxb5 24.c8=Q Be3+ 25.Kh1! Raxc8 26.Rxc8 Rxc8 27.Qxc8+ Kg7 28.Qc7+
Kh6 29.Nc4 Bxc1 30.Rxc1+–
c) 21...Bxg3?! 22.dxc7

261
22...Qe7 (22...Bxh2+ 23.Kxh2 Ng4+ 24.fxg4 Qh4+ 25.Rh3+–) 23.d6! Qg7 24.hxg3 Nxg3 25.Nb3 fxe4 26.fxe4
Nfxe4 27.Rxf8+ Rxf8 28.c8=Q±

22.hxg3

22...Nh5!?

Black throws caution to the wind and goes all-in.

22...Bxg3 23.dxc7! Qe7 24.d6 Qg7 25.Nc4

262
25...f4 (25...Bxb5 26.Nb6 Bxe2 27.Qxe2 and White’s pawns are unstoppable) 26.Nxe5 Qh6 27.Bc4 Qg5 28.Nxd7
Nxd7 29.Bb2 with a decisive advantage for White.

23.gxf4 Nxf4?

And a final mistake.

23...Qh4 would still have posed some problems for White. 24.Rf2!

24...Ng3
a) 24...fxe4 25.Nxe4+–
b) 24...g5!? A resourceful move. Black attempts to open the g-file to assist in the assault against the white king.
25.fxe5 Simplest. (25.Nxc7 gxf4 26.Bf1 or 25.Bf1 wins as well) 25...Ng3 26.Rh2 Nxe2+ 27.Rxe2 Bxb5 28.Nf1
White has successfully defended against Black’s threats and is left with a huge material advantage.
c) 24...Nxf4 25.Bf1 Qg3+ (25...Bxb5 26.Rh2 Qg3+ 27.Kh1+–) 26.Rg2 Nh3+ 27.Kh1 Nf2+ 28.Rxf2 Qxf2 29.Qd1!
Bxb5 30.Bxb5 cxd6 31.Qe2 Qh4+ 32.Kg1+– winning.
263
25.Rh2 Nxe2+ 26.Rxe2 Bxb5 27.Nf1 Bxe2 28.Qxe2 cxd6 29.fxe5 dxe5 30.Rc7 with a winning position due to
Black’s helpless rooks which cannot protect all his structural weaknesses. 30...fxe4 31.Qxe4 Qxe4 32.fxe4 Ra1
33.d6 Kg8 34.d7+–

24.Bc4!!

The only move in the game that was difficult to foresee. It seems unbelievable that with most of White’s pieces on
the queenside, and the d2-knight blocking any help for the king, he can weather Black’s attempts on his king’s life,
but it is so. I remember my opponent’s incredulous facial expression turn to darkness when he realized that he had
pushed too far and sacrificed too much material.
24.Rf2 also leaves White with an advantage, but the road is murkier. 24...Nh3+
[24...Qg5+ 25.Kh1! (25.Kf1? loses outright: 25...Qh4 26.Ke1 Ng2+ 27.Kd1 Qxf2‚ ) 25...Qg3 26.Rh2 Nxe2

27.Qd1!! Another computer move that is incredibly difficult to find, after which Black is just lost. (27.Rxe2 Bxb5
28.Nf1 Qh3+ 29.Nh2 cxd6 and although White is better, there is plenty of play left.) 27...Nxc3 28.Nxc3 Qg5
264
(28...fxe4 29.Ncxe4 Qf4 30.dxc7+–; 28...cxd6 29.Nf1! The point. The black queen is captured.) 29.dxc7+–]
25.Kf1 (25.Kg2? Qg5+! 26.Kxh3 Qg1! 27.Rg2 f4+–+) 25...Nxf2 26.Kxf2 Qh4+

27.Kg2! The only move to play for an advantage.

a) 27.Kf1 Qh1+ (27...fxe4 28.Nxe4 Bh3+ 29.Kg1 Qe1+ 30.Kh2+–; 27...f4 28.Bc4!) 28.Kf2 Qh4+=
b) 27.Ke3?! f4+ 28.Kd3 Bxb5+ 29.Nc4 Qf2–+
27...f4 28.Bf1! (28.Nf1? Rf5!! Black manages to bring his last reserves into the attack, completely reversing the
course of the game and winning! 29.exf5 gxf5–+ 30.Nh2 Rg8+ 31.Kf1 Qxh2 32.Ke1 Rg1+ 33.Kd2 Bxb5 34.Rd3
Rg2 35.Kc3 cxd6 36.Rd2 Bxe2 37.Qxf5 Qh6³ with an advantage for Black) 28...Rf5 (28...Qg3+ 29.Kh1 Qh4+
30.Kg1 Qg3+ 31.Bg2+– White successfully defends against Black’s attack.) 29.exf5 gxf5

30.Bc4!! This bishop move seems to be the key winning move in many similar positions. All White’s pieces move to
the queenside leaving the king alone. The king, like the captain of a sinking ship, leaves the kingside last and moves
to safety on the queenside to win the game. 30...Rg8+ (30...Qe1 31.Nf1) 31.Kf1 Qh1+ 32.Ke2 Rg2+ 33.Kd3 e4+
34.fxe4 fxe4+ 35.Nxe4 Rxc2 36.Rxc2+–
265
White has more than sufficient material compensation for the queen.

24...Nh3+

a) 24...Qg5+ 25.Kf2 Qg2+ (25...Qh4+ 26.Ke3 Qg5 27.Nb3! fxe4 28.Qxe4+–) 26.Ke1 Qg3+ 27.Kd1+– The white
king escapes to safety.
b) 24...Qh4 25.Nb3 Nh3+ 26.Kg2±

25.Kh1!

25.Kg2 Qg5+ 26.Kxh3 f4+ 27.Kh2 Qg3+ 28.Kh1 Qh3+ 29.Kg1 Qg3+=

25...Qh4

26.Nb3!!
266
In keeping with the line, White moves pieces away (!!) from the kingside to protect the king. In those days I was a
calculating monster and had already seen the win against best Black play. The rest of the game is a demonstration of
accuracy and liquidation of any potential black activity.

26...fxe4 27.Qh2!

White finally brings the queen over to support the king. The pawns in the center do not matter.
27.fxe4?? Nf2+ 28.Kg2 Bh3+ 29.Kg1 Qg4+–+

27...Rf5

27...Ra2 28.Nd2 e3 29.Rxe3+– Black’s attack has been rebuffed, leaving White with a decisive material advantage.

28.f4!

28.fxe4 Rh5 (28...Qxe4+ 29.Qg2 Rxf1+ 30.Bxf1+–) 29.Rcf3! Qxe4 30.Nd2 Qg4 31.Rf8+ Rxf8 32.Rxf8+ Kg7
33.Be2 still wins the game.

28...Rh5™ 29.Qg3!

Giving Black a choice of whether to exchange queens or forfeit his attack on the king.

29...Qxg3

Black resigns himself to his fate.

a) 29...Nxf4+ 30.Qxh4 Rxh4+ 31.Kg1 Ra2 32.Nd2 Rg4+ 33.Kf2 Rg2+ 34.Ke1+–
b) 29...Qf6 30.fxe5 Qg7 31.dxc7+–
c) 29...Qd8 30.dxc7 Nxf4+ 31.Kg1 Qf6 32.Bxf4 exf4 33.Rxf4 Bf5 34.Be2 Rg5 35.Bg4+–

30.Rxg3 exf4 31.Bb2+ Kg8 32.dxc7!!

Finishing the game with a flourish. White’s unstoppable pawns in the center are more important than the rook.

267
32...Bxb5

32...fxg3 33.d6+ Be6 34.Bxe6#

33.Bxb5 fxg3 34.Kg2

The last finesse. White’s king finds cover behind the black pawn, and now nothing can stop White’s pawns.

34...Ng5

34...e3 35.Bd7+–

35.d6 Rh2+ 36.Kxg3 Rxb2 37.Bc4+ Kg7 38.d7

Black resigned. 1–0

268
Overall not much happened in the game, bar the fact that my opponent was the current World Champion. It is only
fair to say that our future games resulted in a huge winning score for Mr. Kasparov. This was the one and only game
where he underestimated me as his opponent and paid the price for it.

269
 GAME 11
Gata Kamsky
Vladimir Kramnik
Lucerne 1993
Another milestone for me, both personally and as a team player. Lucerne hosted the World Team Championship in
1993 with the participation of the top ten teams in the world. Russia was obviously considered the favorite, but the
US team, as always in good fighting team spirit, managed to steal the gold. That was, and remains to this day, my
only team gold medal. My game against Kramnik helped our team draw the match vs. Russia and at the same time, I
broke the psychological hold he had had over me since my defeat to him earlier in the year at Linares. Kramnik was
Mr. Kasparov’s personal protégé and student, and was heavily pushed as such. It is quite ironic that despite losing
matches to both me and Shirov, Mr. Kramnik was personally handpicked by his former mentor to play a world
championship match against him. Thus Mr. Kasparov’s prophecy came true: Mr. Kramnik became World Champion
by winning that match and would later avoid a rematch, causing Mr. Kasparov to retire in frustration over the matter.
Of course, Mr. Kramnik would also participate in one of the most scandalous world championships in the history of
the royal game, the infamous ‘toiletgate’ match vs. Mr. Topalov, but that is not a story for me to write about. Back in
1993, we were basically considered as rising contenders for the world championship title, but still with a lot to
prove.

1.d4 d5 2.c4 c6 3.Nc3 Nf6 4.Nf3 e6

An invitation to one of the sharpest lines of all, the Botvinnik Variation of the Semi-Slav Defense. White has the
option of refusing and steering the game into more positional lines starting with 5.e3, which usually lead to exciting
chess in the Meran or Anti-Meran variations. I felt that I was ready for the Botvinnik lines, but my opponent had
other ideas.

5.Bg5 h6

My opponent chooses to delay employing the Botvinnik Variation until our PCA Candidates match that was to take
place the following year in New York.

6.Bxf6

Many years later, players would discover yet another exciting sub-line in this sharp Slav variation running 6.Bh4 g5
7.Bg3 b5 where, despite lots of games and really heavy theory, no conclusive verdict has been reached to this day.
270
However, faced with a surprise by my opponent, I decided to go for a solid approach. White does not strive for
much, and pins his hopes on his better development and slightly more space in the center, which gives him the
freedom to choose the arena of future action. Black, on the other hand, has a slightly cramped position, like a coiled
spring, ready to unleash two murderous bishops on White’s center. The battle will be played out around how the
center is to be opened and when.

6...Qxf6 7.e3 Nd7 8.Bd3

One of the main lines.

Other moves include 8.Qc2 or 8.Rc1, but the purpose of this book is not to go into a theoretical discussion, but rather
into the complexities of the middlegame and endgame.

8...g6

Standard development of the bishop, even today. Black prepares pawn breaks in the center which will open this
dark-squared beauty to full effect.

9.0-0 Bg7

10.e4

Another main line for White is 10.Rc1, when after 10...0-0 he has a choice. He can either play 11.b4 and follow the
standard anti-Grünfeld idea of expanding on the queenside, which makes sure that Black’s g7-bishop is out of work,
or he can play the text move, 11.e4.
After 10.Rc1 0-0 11.e4 (11.b4 dxc4 12.Bxc4 e5 13.Bb3 is approximately equal, though White has a temporary
advantage in development) 11...dxc4 12.Bxc4 (12.e5 Qe7 13.Bxc4 c5! 14.Re1 cxd4 15.Qxd4 Nb6 and Black is OK,
for example 16.Ne4 Bd7 17.Nf6+ Kh8! A very cool move, just ignoring White’s f6-knight. 18.Bb3 Bc6 19.Qf4 Rfd8
20.Nd4 Bd5 21.Nxd5 Nxd5 22.Qe4 Rac8=) 12...e5 13.d5 Nb6

271
we have reached the main position of this line. There have been a lot of attempts by White to demonstrate a small
plus, but none has really succeeded. 14.dxc6 (14.Bb3 is an older move that leads to a complicated middlegame with
lots of games and analysis which we will not go into here.) 14...bxc6 15.Be2

15...h5! More accurate than the immediate ... Rd8 which I played against van der Sterren in the third game of our
FIDE Candidates match in 1994, although I eventually equalized in that game and it ended in a draw. The idea is to
meet 16.Qc2 with 16...Bh6 and White’s heavy artillery is dislodged from attacking Black’s target on c6. (15...Rd8
16.Qc2 Bg4 17.Nb1 Bxf3 18.Bxf3 Qe6!? 19.b3 Rd6 20.Rfd1 Rad8 21.Rxd6 Rxd6 22.Nd2 h5 23.Be2 Bh6 24.Rd1
Qd7 25.Nf3 Rxd1+ 26.Bxd1 Qd6 27.Be2 Bf8 28.g3 Kg7 29.Kg2 Be7 30.Qc3 Nd7 31.Bc4² with a rather annoying
small plus for White. The opposite-colored bishops may even create extra difficulty for Black’s defense as White’s
c4-bishop is better placed.) 16.Qc2 Bh6 17.Rcd1 Qe7 18.Nb1!

272
There is still plenty of play left, since White’s knights have very strong squares available for occupation.
18...Bg4 19.Nbd2 Rfd8 with counterplay.

10...dxc4 11.e5!?

The first critical position.

My idea during preparation was to prevent Black’s freeing ...e6–e5 push. In the ’90s computers were quite awful, so
when it came to preparation everyone was left to their own devices, in other words you had to use your own head. I
had looked at the general outlines of the position and figured that the central pawn structure was good for White,
who also had more space for maneuvering and a possible plan of attack on the kingside with h4–h5 (Black could
hardly risk playing ...h6–h5 because of the importance of the g5-square for White’s knights). Black’s only source of
counterplay lay in a timely ...c6–c5 break, which could be countered by White’s d4–d5 push, for example. So the
game revolves around these motifs.

11...Qe7 12.Bxc4
273
Black now has the option of playing an early 12...c5 before castling short.

12...0-0

Played after considerable thought. My opponent’s intuition served him well.

12...c5?!

This move is indeed too risky to be played with an uncastled king.

13.Re1! This quiet centralizing rook move seems to be the strongest.
[13.d5!? An interesting pawn sac, which is probably what my opponent feared during the game. 13...Nxe5 14.Nxe5
Bxe5 15.Re1 Now Black faces a difficult choice, which is considerably harder to make during an actual game.

15...Bxc3!? This move looks quite logical, with the idea of further simplifying the position and liquidating into an
even heavy piece endgame, where any possible defects in Black’s pawn formation can be exploited by White only in
limited fashion.

274
a) 15...Bd4? 16.Nb5±
b) 15...Bxh2+ This computer move looks pretty greedy and something that only a beginner would consider, however
it is playable, thanks to the numerous exchanges that have left fewer pieces on board, thus giving Black good
chances to defend. 16.Kxh2 Qh4+ 17.Kg1 Qxc4 18.Ne4!

The only way for White to exploit his advantage in development and the dark square weaknesses in Black’s pawn
chain.
18...0-0 19.d6 Qd4 (19...f5 20.Rc1 Qd5 21.Nxc5 Rd8 22.Qxd5 exd5 23.Nd3! Bd7 24.Rc7 Bb5!² And now either
25.Nf4 or 25.Nc5 gives White a certain advantage in continuing the fight for a win.) 20.d7 Bxd7 21.Qxd4 cxd4
22.Nf6+ Kg7 23.Nxd7±. While Black has three pawns for the minor piece, White’s chances are preferable due to his
coming strong knight blockade on the d3-square.
16.bxc3 0-0 17.Qb3

17...e5 (17...Kg7 18.dxe6 Bxe6 19.Bxe6 Rfe8! 20.Re5 — or 20.Bc4 Qxe1+ 21.Rxe1 Rxe1+ 22.Bf1 Rd8 23.Qxb7
Rdd1 24.Qb5 Rxf1+ 25.Qxf1 Rxf1+ 26.Kxf1 Kf6= — 20...fxe6 21.Rae1 b6 22.Rxe6 Qd7 23.g3 Rad8 with
approximate equality as White doesn’t have enough material to attack the slightly open black monarch.) 18.Re3 Qd6
275
19.Rae1 Re8 20.f4

20...e4! 21.Rxe4 Rxe4 22.Rxe4 Bf5 23.Qxb7 Bxe4 24.Qxa8+ Kg7 25.Qxa7 (25.g3 Qb6 and Black gets sufficient
active counterplay to neutralize White’s extra pawn.) 25...Bxd5

26.Bxd5 (26.Bf1 Qxf4 27.Qxc5 Bxa2=) 26...Qxd5 27.h3 Qd1+ 28.Kh2 Qd2 29.Qxc5 Qxf4+=]
13...cxd4 14.Qxd4 0-0 15.Ne4! Nb6 16.Rad1!

276
The difference compared to the position after 10.Rc1 is obvious. Here White’s d1-rook prevents Black from
developing his c8-bishop via the d7-square. 16...Nxc4 17.Qxc4 b6 18.Qc3! Bb7 (18...Ba6 19.Nf6+±) 19.Nf6+ Bxf6
Black is forced to take on f6 in view of the coming Rd7 double attack. 20.exf6 Qc5 21.Qd2! It is important to keep
the queens on, as White gets the well-known superior combination of Q+ N vs. Q+B which, combined with the far-
advanced f6-pawn, grants him a big advantage. 21...Qh5 22.Ne5±

Black’s b7-bishop is going to be neutralized by White’s f3-pawn. Black will be struggling for whatever counterplay
he can find, while White can enjoy building up his spatial advantage for either an attack or maneuvering on both
flanks.
22...Bd5 23.b3 b5 24.f3 Qg5 25.Qxg5 hxg5 26.Kf2 Rfc8 27.Rc1 a5 28.Kg3 a4 29.bxa4 bxa4 30.a3+– The white king
strolls up to grab Black’s g5-pawn and the position becomes a technical win for White.

13.Qe2!

A move with a threefold purpose, the kind that makes a pro player very happy!

277
(i) White prevents Black’s immediate ...c5 break; (ii) he leaves himself options for the optimal deployment of his
rooks, and (iii) hints at the possibility of exchanging Black’s c8-bishop via Bc4–a6, should the possibility arise,
since Black’s queenside pawns are a bit vulnerable, given the temporary lockout of Black’s g7-bishop.

13...b6 14.Rfe1

An immediate 14.Ba6 would be dubious, not because of wasting time moving an already developed piece, but
because of the following line:
14.Ba6?! Bxa6 15.Qxa6 c5 16.Qa3 Rae8³ Black is simply better, as White’s pawn center collapses.

14...a6!

It is interesting to note that despite playing right before the advent of the computer revolution, we still played moves
that modern computers agree to be optimal. Since White was ready to play 15.Ba6, Black stops the threat. During
the game I felt that Black had played too many preparatory pawn moves in order to develop his c8-bishop, and that
such a slow plan should not work, however his strategy is sound. Many years later, in 2011, in game two of my
return Candidates match vs. Topalov, I myself would employ a similar restrictive pawn chain as Black, forcing my
opponent into passive play which didn’t suit his style, to win a critical game.

15.Bd3

In similar vein, I sidestep Black’s possible expansion on the queenside with ...b5/...c5 and prepare my d3-bishop to
meet its black counterpart on the e4-square, hoping to prevent the ...c5 break completely.
15.Rad1 Bb7 This move is most likely to lead to the game continuation. However if White persists in keeping his
c4-bishop on the current diagonal to push for a d5 counter if Black breaks with ...c5 there’s not much he can do.

278
16.a3 c5 17.d5 b5 18.Bb3 c4 (18...Rae8!?³ Developing the rook seems even stronger as Black prepares his entire
army for the opening of the center and coming battle, where his strong pair of bishops will give him an advantage as
the white knights have no useful outposts anywhere.) 19.Bc2 exd5 20.Nxd5 Qc5³

Black starts to take over the game.

21.b4!
[21.Nb4 This doesn’t work in view of some deep tactics, though at first it seems that Black cannot prevent his king
being opened up with White’s e5–e6 break. 21...Nxe5!!µ 22.Nxe5 Rae8 23.Nd7 Qg5

279
24.f4!? An attempt to simplify into a superior endgame by tactical means. (24.Qf1 Rxe1 25.Rxe1 Rd8 26.Rd1
Bxb2µ Black’s bishop pair dominates White’s two knights and bishop.) 24...Qxf4 25.Qxe8 (25.Qf1 Rxe1 26.Qxe1
Bd4+ 27.Kh1 Qg4 28.Qg3 Qxd7–+) 25...Rxe8 26.Rxe8+ Kh7 27.Nf8+ Bxf8 28.Rxf8 Kg7 29.Rfd8

29...Qg5!! 30.R8d2 (30.R1d2 a5 31.Na2 Bxg2–+) 30...Bf3 31.Nxa6 Bxd1 32.Rxd1 Qf6 33.Nb4 Qxb2–+ Black has
restored the material balance and dominates the position thanks to White’s terrible piece coordination. In addition,
White is likely to lose one of his remaining minor pieces, leading to a completely winning game for Black.]
21...cxb3 22.Bxb3 Qxa3 23.e6 Qxb3 24.Ne7+ Kh7 25.exd7 Rad8

280
The game is roughly equal as White’s far-advanced passed pawn compensates for the missing material. The most
logical scenario would be the following: 26.Rd6 Qb4 27.Red1 a5 28.Ne5 Qc5 29.N7c6 Bxc6 30.Nxc6 Rxd7
31.Rxd7 Qxc6 32.h4 a4 33.R7d6 Qe8 34.Qxe8 Rxe8=

People who play the Grünfeld as Black will recognize the familiar drawing endgame pattern!

15...Bb7 16.Be4²

Prophylaxis is the name of the game! I had the feeling that my opponent was extremely uncomfortable, as he had a
passive position and was looking for ways to equalize. This clearly didn’t help the rest of his team as board one is
very important, at least in terms of team psychology. Black cannot free himself easily with the programmed ...c6–c5.

281
16...Ra7?!

A modern approach to developing the rook, reminiscent of the method I employed in the above game vs. GM
Beliavsky!! However not all rooks are made to dance!
16...Rab8 17.a3! The idea is to provoke Black into playing 17...a5 which would severely weaken the b5-square,
setting it up as an excellent outpost for the white knights! (If 17.Rac1 the following computer line illustrates the
possibilities: 17...b5 18.Red1 c5 19.Bxb7 Rxb7 20.Ne4 cxd4 21.Nd6 Rbb8 22.Rc6 Rfd8 23.Rxd4 Nb6 24.h4!² While
White’s advantage is not large, it is not easy for Black to play this position, as one tiny inaccuracy can make his life
very difficult.)

17...Rfd8 Best. Black has to play the waiting game, while trying to find moves that can potentially improve his
position. (17...c5?! 18.Bxb7 Rxb7 19.Qxa6 White is just a pawn up. 17...a5?! 18.Rac1 Rfd8 19.Rc2! h5! 20.Qe3 c5
21.Bxb7 Rxb7 22.Nb5± and now we see how White’s Nb5 is ideally placed, thanks to that little weakening ...a6–a5.
17...b5?! is not what Black wants. 18.b4 imprisons the b7-bishop.) 18.b4 (18.Rad1 b5 19.b4 a5 20.Rb1 axb4 21.axb4
Nb6 22.Bd3 Ra8 23.Ne4 Nd5 24.Qd2 Ra3 25.Nc5 Rda8 26.h4 Bc8² and it is not easy for White to demonstrate how
282
to use his superior minor piece placement.) 18...a5 19.bxa5 bxa5 20.Reb1 Ba8 21.Rxb8 Rxb8 22.Qa6 Qd8 23.Rb1
Rc8! Black has almost equalized as his g7-bishop is joining in the game via the f8-square, reminding White of his
own weak a3-pawn. The game, however, is far from over and a lot of things can happen. White has superior minor
piece coordination and Black’s a8-bishop is still out of the game.

24.h4 A standard pawn break that intends to open a diagonal on the kingside for White’s e4-bishop. 24...Bf8 25.Qd3
Qe8 26.a4 Rd8 27.h5 g5

This is an example of how the game could continue, leaving White with some hope of an advantage.

17.Rac1 b5

283
Black puts the finishing touches to the queenside pawn structure before finally playing the long-delayed
programmed ...c6–c5 break.

18.Qe3!

Immediately reminding Black of potential pins on the a7-rook.

18...c5

18...Nb6 19.d5!? (19.Nd2!?² This move, with the idea of playing Nd2–b3–c5, completely blocking Black’s b7-
bishop, is a solid alternative to 19.d5 and guarantees White a long-lasting strategic advantage.) 19...cxd5 20.Qxb6
dxe4 21.Qxa7 exf3 22.g3±

White’s rooks are superior to Black’s bishops. Also, Black is unable to use the weak light squares around White’s
king as his pieces are pinned down on the queenside.
22...Ra8 23.Qb6 Rc8 24.a3!! Finding a way to open files for the rooks which spells trouble for Black. 24...h5 25.Na2
284
Rxc1 26.Rxc1 Qd7 27.Qd6±

19.d5?

I had been preparing this counter break for so long that I didn’t stop to look for alternatives — which there were!!
And better ones!!
a) 19.b3! How little attention such small pawn moves attract, and how much of a difference they can make!!
19...Rfa8 (19...Rc8 20.Bxb7 Rxb7 21.Ne4+–; 19...Bxe4 20.Nxe4 cxd4 21.Qxd4± Black’s queenside is in ruins.)
20.Bxb7 cxd4 21.Nxd4 Rxb7 22.Nc6 Qf8 23.Ne4 Nb6 24.Nd6! (24.f4 Nd5²) 24...Nd5 25.Qe4 Rd7 26.a4 bxa4
27.bxa4 h5 28.g3±

Advantage to White, simply because Black’s entire heavy arsenal has no moves!! White’s pair of knights on the
c6and d6-squares control the queenside, leaving White with the luxury of improving his position slowly. (28.Nc4)
b) 19.Bxb7 cxd4 (19...Rxb7?! 20.d5± White gets what he wants.) 20.Nxd4 Rxb7

285
21.Nc6 (21.f4?! Nb6 would leave White with only a slight spatial advantage, which could slowly be neutralized to
even the game. Potentially Black’s g7-bishop could even create problems for White.) 21...Qc5 22.Ne4 (22.Qe4 Qb6!
23.Ne2 Nc5 24.Qf3 Nd7=) 22...Qxe3 23.Rxe3

23...g5! Once again, a little pawn move can affect the whole position. Black undermines White’s e5-pawn, resulting
in a game where both sides have chances. 24.g3 g4!∞

19...Nxe5

Of course! Black captures a very important pawn.

I went for the line that I thought would cause Black maximum trouble.

20.Nxe5 Bxe5 21.dxe6

286
Now Black faces a difficult decision...

21...Bxe4?!

...and makes an inaccurate move. White is left with a superior Q+N vs. Q+B combination and when he sets up his
forces on the light squares, Black’s dark-squared bishop will be outplayed.
a) 21...Kg7! The only move that guarantees Black equality, as all other moves leave White with some pressure on
the black position.
22.Rcd1! The only move that poses questions for Black. (22.Nd5 Bxd5 23.Bxd5 Bd4 24.Qg3 Qd8!µ; 22.exf7 Bd4
23.Qd2 Qxf7 Clearly Black’s position is superior. 22.Bxb7 Bd4 23.Qd2 Rxb7 and it is only Black who can press for
an advantage.)
22...Bxe4

23.Qxe4
[23.Nxe4 Bd4 24.Rxd4 (24.Qg3 Qxe6 25.Nxc5 Qc4 26.Re4 Qxc5 27.Rexd4 Re7=) 24...cxd4 25.Qxd4+ f6 26.Nc5
287
Rd8 27.Qe3 Raa8 28.b4 Rd5³ White will have to struggle for sufficient compensation for the missing exchange.]
23...Bxc3 24.bxc3 Qf6!? (24...Qxe6 25.Qxe6 fxe6 26.Rxe6 Rf6 27.Rxf6 Kxf6 28.Rd6+ Kf7 and while this position
looks rather passive for Black, a draw is not far off. 29.h4 h5 30.Rc6 c4 31.a4 bxa4 32.Rxc4 a3 33.Ra4 a5 34.f3 Ke6
35.Kf2 Kd5 36.Ke3 a2 37.Kd3 Kc5 38.Rxa2 a4 with an easy draw for Black) 25.Qc6!

Posing a final question. 25...c4 (25...fxe6 26.Qxc5²) 26.Re3 fxe6 27.f3 e5 28.Rd6 Qg5 29.Qc5 Re7 30.h4 Qxh4
31.Re4 Rf4 32.Rxe5 Rff7 33.Rxa6 Rxe5 34.Qxe5+ Kh7 35.Ra8 Qf6 with an equal endgame.
b) 21...Bxc3?! 22.Qxc3 c4 23.b3!

White leaves his e6-pawn en prise to create discoordination among Black’s pieces. Capturing the pawn leads to a
small but lasting advantage for White, which is, again, critical in a team competition, as it guarantees lasting
psychological pressure on all the players on the opposing team! 23...cxb3 24.axb3 Bxe4 25.Rxe4 Kh7! Black’s
position is becoming more unpleasant with every exchange, which is unusual, since exchanges generally favor the
defense. But in this particular position, with White controlling major open files, any potential weakness in the pawn
formation would be mercilessly attacked.
c) 21...Bd4!? 22.Qxh6 Qf6 23.Re2 fxe6 24.Qxg6+ Qxg6 25.Bxg6 Kg7 26.Be4 Bc8 27.b3 e5²
288
No doubt my opponent considered this position for a long time. The question is whether Black’s two strong bishops
provide sufficient compensation for White’s extra pawn — and another factor has to be taken into consideration. In
those days, I was regarded as one of the best defenders in the world amongst top players, so giving me an extra pawn
in the endgame was a considerable risk. Yet, in this particular position, Black should have nothing to fear, as his d4-
bishop is just too strong, while White’s extra kingside pawns cannot really move, due to the presence of all the rooks
on the board. Opening up the position, especially the king’s position, would be suicide for White.

22.Nxe4

22...Bxb2!

Black was probably counting on 22...c4, but realized his mistake and found the correct move.
22...c4?! 23.Qxh6 Black probably missed that he cannot capture on b2. 23...Bg7 and now White has several options
which all lead to extra material with good winning chances.
289
[23...Bxb2? 24.Ng5 f5 25.Rcd1!

25...c3 (25...Qg7? 26.Qxg7+ and all recaptures are bad for Black. 26...Kxg7 27.Rd7++– or 26...Bxg7 27.e7+–)
26.Qxg6+ Qg7 27.Qxg7+ Rxg7 (27...Kxg7 28.Rd7++–) 28.e7 Rxe7 29.Rxe7 c2 30.Rde1 c1=Q 31.Rxc1 Bxc1
32.Ne6± with an extra pawn for White in the endgame and very good winning chances.]
24.exf7+!?
[24.Qe3 Bxb2 25.Rc2 Bg7 26.exf7+ Qxf7 (26...Rxf7 27.Qb6!ƒ) 27.Nd6 (27.Ng5 Bh6!=) 27...Qd7 28.Rd2 c3
29.Rdd1 Raa8!

and it seems that White will have to be careful not to push too hard. 30.Qe4 Rad8 31.f3 (31.Qd5+? Kh7µ) 31...Qc7
32.Qxg6 c2 33.Qe6+ Kh7 34.Qe4+ Kg8 35.Qe6+=]
24...Qxf7 25.Qd2 Rd7 26.Qc2² White has the advantage thanks to his extra pawn and unassailable knight on e4
which cements the defense, though Black obviously keeps some counterplay due to his pawn majority on the
queenside supported by the g7-bishop and other pieces.

23.Rxc5 Bg7
290
24.Rc6!

This move remains White’s only attempt to keep pressure in the position by trying to use his knight to establish a
light-squared strategy. If White manages to keep a hold on the e6-square, he might eventually strangle Black’s
defenses. I must add, though, that Black has adequate defensive resources here.

24...fxe6 25.Nc5

During the game I felt that this move was a serious error resulting from the onset of time trouble, and could have
undone the fruits of all the previous work. However the computer disagrees and states that even after what is
possibly White’s best move, 25.Qb3, Black can successfully defend his position.
25.Qb3!? Kh7 26.Rxe6 Qf7

Now White has the choice of moving his queen to either e3 or g3. 27.Qg3! Probably the stronger option, as it keeps
Black tied to the defense of his g6-pawn. (27.Qe3?! maintains the pressure on Black. 27...Rd7! Black sacrifices the
291
a-pawn in order to activate his pieces. 28.Rxa6 Re8 29.h3 Ree7! 30.Qb3 Ra7!

The point. Black exchanges the remaining pawns on the queenside to transpose into a 2 vs. 3 pawn formation on the
kingside with very good drawing chances.) 27...Rc7
a) 27...Bd4?! Misplaced activity that helps White by giving him time to protect his weak back rank. 28.Kh1 Rc7
(28...a5? 29.f3 b4 30.Nd6 Qg7 31.h4!‚) 29.Rxa6 (29.f4? Bg7 30.Nd6 Qxf4 31.Qxg6+ Kh8= White’s attack has been
rebuffed, leaving Black with good chances of an advantage.) 29...Re7 30.h3 Qf5 31.f3± leaves White with an extra
pawn.
b) 27...Qf5! might be best, but after 28.h3 b4 29.Nd6 Qg5 30.Qd3ƒ White retains a great initiative.
28.Rxg6!

28...Qf4 (28...Rfc8 29.Rxa6 Rc1 30.Qe3 Qf4 31.Rf1 Rxf1+ 32.Kxf1 Qxh2 33.Ng3±) 29.Qxf4 (29.Qd3!? allows
White to keep the queens on, which might be a good idea, since Black’s combination of R+B and pawn majority on
the queenside makes a white win in any endgame questionable. 29...Kxg6 30.g3 Qf5 31.Qd6+ Kh7 32.Qxc7 b4±)
29...Rxf4 30.Re6 Rc2 31.g3 Rf7 32.Rxa6 Bd4 33.Kg2 Bxf2 34.Rd1 Bd4+ 35.Kh3 Bg7± keeps White’s advantage.
292
25...Qf7?

Returning the favor. We were both short of time at this point, but obviously White keeps his plus when the knight
comes to e6. Nevertheless around here Vladimir offered a draw and I felt emboldened to continue the game. To
explain: in the early days, my opponent was known to offer draws in inferior positions. Sometimes such a tactic
might work, but once people found out, it backfired, because people realized that their position was superior. With
the passage of time, people sometimes manage to eliminate their mistakes, but many years later, when I had just
returned to competitive chess and faced Kramnik again at the Turin Olympiad in 2006, the same thing happened. I
was not sure about my position, but my opponent suddenly made a move and offered a draw. I realized that there
was something in the position. It was an audacious move, leading to a wild game, especially the endgame where I
missed the win. I’ll present the game later in the collection. Keep an eye out for it, now that you know the
background story.
25...Rf5!

Black simply equalizes as there are not enough attacking pieces to support White’s onslaught. 26.Nxe6 Re5 27.Qc1
Rxe1+ 28.Qxe1 Rd7 29.Rxa6 (29.g3 Rd6 30.Rxd6 Qxd6 31.Qe4 Kf7= and Black can look to the future with
293
optimism.) 29...Rd5!!

The key move that we both missed. The double threat of ...Re5 and ...Qd7, highlighting White’s back rank
weakness, evens the score. 30.f4 Qb7! (30...Qd7 31.Qe4) 31.Ra3 b4 32.Re3 Bc3 33.Qb1 Kf7 34.Rxc3 Qa7+ 35.Rc5
Rxc5 36.Nxc5 Qxc5+ 37.Kh1

37...Qf2!= The final touch.

26.Nxe6 Re8

294
27.Qb3?

Yet another mistake in such a seemingly simple position. Every tempo is critical, as Black still has a pawn majority
on the queenside and, given time to pin White’s e6-knight, can simply push it.
27.h3!! a5 28.Qc5 Rae7 29.Re3! A key tactical idea; the knight is untouchable.

29...Kh7 (29...Rxe6 30.Rcxe6 Rxe6 31.Qc8++–) 30.Nxg7 Qxg7 31.Rg3 Re5 32.Qb6±

27...Kh8?

27...a5!

295
The computer finds this amazing resource. The critical point is that once White creates a luft for his king and places
his rook on e3, Black is tied up and his queenside pawns become breakfast. So he has either to start pushing his
queenside pawns more quickly or prevent White from improving his position by denying the white rook the key e3-
square. 28.Qxb5 (28.g3 Rae7 29.Re4 a4 30.Qd5 Qf5=) 28...Bd4! 29.Re2 Rb7 30.Qd3 Rb6!

A series of forced (!) computer moves, leading to a drawn position.

31.Rxb6 Bxb6 32.Qc4 (32.Qd5 a4 33.h3 Bxf2+ 34.Rxf2 Qxe6=) 32...Bxf2+=

28.g3 Qf5

296
Black has lost too much time. He is unable to push his pawns and White starts to create too many tactical threats.

29.Re2± Rae7

29...Rg8 30.Nf4 Bf6 31.Ree6 The weakness of the black pawns on the sixth rank is critical.

30.Rxa6+–

A pawn up with a knight on e6, I finally felt that I could calm down and look at how my teammates were doing. One
board had lost and the other boards were likely draws, so a win on this board would let us draw the match vs. the
previous gold medallists and give us a fighting chance of gold ourselves.

30...Bf6

Now White just has to find a way to win another black pawn and it becomes a technical win. It was obvious that my
opponent felt that his position was resignable as he started to play faster.
297
31.Re3 Kh7 32.Rf3

White eliminates Black’s defender of the sixth rank, the f6-bishop.

32...Qe5 33.Nf4

Black loses a second pawn as his g6-pawn is too weak.

33...Qb2 34.Qd3 Re4

35.Kg2!

Prophylaxis until the very last moment!! To eliminate your opponent’s hopes while retaining all the pluses of your
position is a severe psychological blow.

35...Qd4

Black voluntarily gives up a second pawn and after

36.Qxb5 R8e7

he resigned, as he has absolutely no hope of saving his position.

1–0

After the 2-2 draw vs. Russia, our team sailed through the remaining matches without loss, winning our only gold
medal in the World Team Championship to this day. Many years later, after my return to professional chess and to
the US Olympiad team, I would remember our captain, IM John Donaldson, and recommend him to the US Chess
Federation for captaincy of the Olympiad team. Under his guidance, our team would take two bronze medals at the
Olympiad and finally, in 2016, without my participation, the gold medals in the historic Baku Olympiad, where there
would not be a single ex-Soviet or Russian player in the team.

298
 GAME 12
Gata Kamsky
Johann Hjartarson
Biel 1993
The following game was played in the Biel Interzonal tournament in 1993, part of the qualification system for the
World Championship. This was the last Interzonal before the system was abandoned due to cataclysmic changes in
the chess world which led to the split into two world championship cycles. See
https://en.wikipedia.org/wiki/Candidates_Tournament for more information.

1.d4

This was my second Interzonal. In the first, in Manila in the Philippines in 1990, I scored miserably, but it provided
me with invaluable experience and gave me confidence for the Biel event. Biel itself is a beautiful city with a huge
chess tradition; chess truly flourishes there.

1...Nf6

My opponent was an experienced and famous Icelandic former World Championship Candidate. I felt that I needed
to steer the play into more positional waters, as my opponent’s imaginative playing style was full of attacks and
sacrifices.

2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.f3

The Sämisch system. One of the pluses of this set-up is that White solidifies his center and prevents Black from
expanding for an attack on the kingside. Black is forced to make strategic choices and either enter Benoni-type pawn
structures after ...c7−c5 or block his bishop with ...e7−e5 and hope that White doesn’t develop his own kingside
attack after castling long.

5...0-0 6.Be3 Nc6

I would later employ the King’s Indian myself and choose to play 6...c5 positions. At this point in my life I was only
starting to realize just how important opening preparation and knowledge were. Given my distaste for spending
hours on opening preparation, I utilized Lasker’s approach and instead chose lines that would try to negate both my
opponent’s opening knowledge and playing style, and thus place us on a roughly equal footing.
299
7.Qd2

The point of this move was to scare Black into thinking that White is preparing to castle long and follow up with an
attack on the kingside. The psychological game is on!

7...a6

Preparing for the typical ...b7–b5 break in case White goes with the above-mentioned plan of castling long.

8.Bd3!

My opponent wasn’t happy with this move. He realized that I was going for short castling where Black’s ...a7–a6
can be considered a potential weakening of the dark squares on the queenside. Black has no choice but to play in the
center now.

8...e5 9.d5 Nd4 10.Nge2 Nd7!?

The idea of this knight move is manifold. Black protects the d4-knight, intends to further protect it with a ...c7–c5
pawn push and, if White takes on c6, possibly transfer a knight to e6 via f8, cementing the monster on d4. White has
to consider carefully how to deal with this black knight on d4. The alternative was to play 10...Nh5, trying to go with
the ...f7−f5 push, but since White hasn’t declared the position of his king, the h5-knight might potentially be a target
for White’s expansion. We were both in uncharted waters, and navigating the correct path for Black was not an easy
task, given his lack of space. White enjoys more space and better piece development with a clear plan for the
middlegame. Black has to be careful as the cost of a mistake is high and he could land with a strategically lost game.
a) 10...c5?! would be a mistake.

300
11.dxc6 Nxc6 (11...Nxe2? loses a pawn with no compensation: 12.cxb7 Bxb7 13.Bxe2) 12.0-0 Be6 13.b3 Rc8
14.Rfd1 Qa5 15.Rac1 Rfd8 16.a3!

White’s huge space advantage and Black’s lack of squares for his minor pieces give White a very comfortable
advantage. For example:
16...Nd7 (16...Qxa3?? loses at once: 17.Ra1 Qb4 18.Ra4 Qxb3 19.Nc1+–) 17.b4 Qc7 18.Nd5 Qb8 19.Bb1

301
19...Bxd5 (19...Re8 20.Ba2 Bxd5 21.Qxd5±) 20.exd5! The strongest solution. White’s pair of bishops, combined
with the potential c4–c5 break and the black queen’s limited mobility to come to the aid of its monarch, give White a
serious advantage. (20.cxd5 Ne7 21.b5! Rxc1 22.Rxc1 Rc8 23.bxa6 bxa6 24.Bd3 Rxc1+ 25.Nxc1 Qc8 26.Nb3 f5
27.Qa5+– White has a close to winning position.) 20...Ne7 21.Bg5 Bf6 22.Ng3

22...Bxg5 (22...b5?! 23.c5 dxc5 24.d6 Nd5 25.Ne4 Bxg5 26.Qxg5 Nf4 27.bxc5±) 23.Qxg5 f6 24.Qe3 b6 25.f4 Rf8
26.f5 g5 27.Ne4 Kh8 (27...Kg7? 28.Nxg5+–) 28.g4± White dominates completely. He has the option of pursuing
pawn breaks on both flanks, which, together with extra space for maneuvering, gives him a great chance of victory.
b) 10...Nh5!

302
Personally I prefer this move, as it is much more active. 11.0-0
b1) 11.0-0-0?! c5! Now it works. White’s king will feel quite uncomfortable with the opening of the queenside. The
game is very complex with chances for both sides, but Black would be the one pressing. 12.dxc6 bxc6 13.Bg5 Qa5!?

14.Be7 Re8 15.Bxd6 Rd8 16.Be7 Rd7 17.Ba3 (17.Bg5 Rb8–+ with the idea Rdb7) 17...Ne6 Black has great
compensation. A nice line could go 18.g3 Rb8 19.Qe3 Nd4 20.Bc2 Bf6 21.Rd2 Ng7 22.h4 Nge6 23.Bd1 Nc5 24.h5

303
24...Rb3! 25.Bxc5 Qxc5 26.Kb1 Rb4 27.hxg6 hxg6 28.Na4 Qxc4 29.b3 Qb5 with the initiative.
b2) 11.Bg5!? Perhaps this was the move that Black didn’t like. White can now capture on d4, doubling Black’s
pawns and creating a potential pawn mass on the kingside. 11...f6 12.Bh4 Nf4!? 13.Nxf4 Bh6 14.Nce2 Nxe2
15.Qxe2 Bxf4

It seems that the position has simplified into an equal middlegame. Black has managed to relocate his locked g7-
bishop on to a good diagonal, taking the c1-square from White’s rooks. However this is the King’s Indian! The
strategic plans remain the same and Black still has to look for resources to counter White’s basic plan of queenside
expansion. The game might proceed as follows: 16.0-0 c6 17.dxc6 bxc6 18.b4 Qb6+ 19.c5 Qxb4 20.Bc4+

304
20...d5! (20...Kg7? 21.Bf2! dxc5 22.Rfd1 and suddenly White is better, due to his superior piece activity.) 21.exd5
Qxc5+ 22.Bf2 Qd6 23.g3!?
[23.dxc6+?! Be6 with counterplay; 23.Rfb1 cxd5 24.Rb6 Qd8 25.Rd1 Kg7 26.Bxd5 Ra7 27.Bb3 Rd7 28.Rxd7+
Qxd7 29.Qc2

Black has compensation, but the position is unclear: 29...Qd2! (29...Rd8 30.g3 Bh6 31.Qc5‚ ) 30.Qc7+ Kh6 31.Qc5
Rd8 (31...Kg7? 32.Rb7+! Bxb7 33.Qe7++–) 32.Rxf6 Bb7 33.Rf8 (33.h4 Rc8‚ ; 33.g3? Qe2!µ) 33...Rd5 34.Qc4 Be3
35.Qh4+ Kg7 36.Qe7+ Kh6 37.Qh4+=.]
23...Bh6

305
24.Rfb1!! (24.Rab1 c5=) 24...cxd5 (24...c5 25.a4! Bd7 26.a5 Rfb8 27.Rxb8+ Rxb8 28.Bxa6 Bd2 29.Qd3 Bb4
30.Bc4 Ra8 31.a6± shows why it was important to go to b1 with the f1-rook.) 25.Rb6 Qd8 26.Rd1 Kh8 27.Bxd5
Ra7 28.Rxf6 Qxf6 29.Bxa7²

11...c5

306
With some advantage for White. Granted, none of these lines were seen during the game, but they demonstrate the
richness of chess in the simplest of positions. One move can make a huge difference!
12.dxc6?! I don’t like this. White simply gives up his center, surrendering space there for no compensation, and
allows Black to reinforce his central d4-knight, followed by expansion on the kingside. (12.Rab1 is a standard plan,
to open up the queenside with a b2–b4 break. The game is complex with chances for both sides, but this is certainly
better than what Black got in the game. The h5-knight is much more active.) 12...bxc6 13.Rad1 Be6 14.b3 c5
15.Nd5

15...f5?! (15...Bxd5 16.cxd5 a5 17.Kh1 Qd7 18.Nc3 f5 19.Na4 Qc7 20.exf5 gxf5 21.Rfe1 Kh8 22.Rc1 Qb7 23.Bxd4
cxd4 24.Rc6 Rad8 Black’s chances will be superior, as his mobile central pawn mass, especially if combined with
threats on the dark squares to the white king, is potentially devastating.) 16.Bg5 Nf6 17.Nxd4 cxd4 18.exf5 gxf5
19.g4!

307
This of course is a modern approach to chess. Computers have changed chess so radically by developing a very
concrete approach to the game. None of the modern top players can probably imagine what it was like to play chess
before engines became what they are today. Moves like 19.g4 would have been considered a great inspiration and
flight of fancy! Nowadays it’s a standard weapon to break up Black’s pawn chain and open the diagonal for White’s
d3-bishop. White’s king is in fact safer than its black counterpart. 19...f4 20.Qc2 h6 21.Nxf6+ Bxf6 22.Bxh6 Rf7
23.Qe2 Bg7 24.Qe4+– demonstrates the strength of the d3-bishop.

11.0-0 c6!?

11...c5 would be just as good, as Black doesn’t really threaten to take on d5.

12.dxc6!?

After serious thought I decided to change the structure. Nowadays I would probably go for a more logical follow-up
with Rab1, with the plan of proceeding with b2–b4, trying to open lines on the queenside. If Black plays ...c6–c5,
then he has essentially lost a valuable tempo.
308
12.Rab1 c5 with counterplay.

12...bxc6 13.b4

The idea behind White’s previous move. Black is going to play ...c6–c5, since with his knight on d7, he is not
planning on playing the more strategic ...d6–d5 break. White will counter a ...c6–c5 push with b4–b5, creating a
passed pawn, not only a protected passed pawn, but also one which blocks the queenside for any play there. The
game will either continue on the kingside, or enter lockdown and be a positional draw.

13...Bb7

13...Nb6? Just to show that, in general, it is a bad idea to allow White to play c4–c5 in such positions. 14.c5 dxc5
15.bxc5 Nd7 16.Na4

16...Ne6 (16...Rb8 17.Bc4 Qe7 18.Nxd4 exd4 19.Bg5! Bf6 20.Bxf6 Nxf6 21.Qxd4 and Black doesn’t have
compensation for his lost central pawn.) 17.Bc4!? Clearly Black has too many holes in his structure, so White has
clear targets to attack. 17...Qe7 18.Rfd1

309
18...Ra7 (18...Re8 19.Qd6 Qxd6 20.cxd6) 19.Qd6 Nf6 20.Qxe7 Rxe7 21.Nb6 Rc7 22.Rab1 Re8 23.Nxc8 Rexc8
24.Bxe6 fxe6 25.Nc3±

14.Rab1 a5!?

At this point Black conceived an extremely interesting plan which he demonstrated later in the game. I felt that
opening the a-file, when my rook had just left it, was not an option for me.

15.b5 c5 16.f4!?

Showing that I was willing to contest space on both sides of the board. Objectively this move is not best, but it does
have the advantage that White will not have to fear being mated as he might have if Black had managed his standard
...f7–f5–f4 push.
The computer suggestion 16.Na4?! is not great. The computer simply evaluates every aspect of the game, including
pawn structure, and underestimates Black’s potential play on the kingside. Eventually the engine recognizes that
White cannot make any serious progress with his play on the queenside as Black has successfully blocked it. 16...f5
310
17.Bf2 f4 18.Nec3 g5 19.Nd5 h5 and Black has a much clearer game plan.

16...exf4?!

Perhaps an inaccuracy. Maybe Black feared White’s f4–f5 pawn push, plus the exchange on f4 also opened his g7-
bishop, so the move is quite logical, but it also gives White’s minor pieces squares and the rook the open f-file to
operate on. As mentioned before, sometimes the best move in the position is not a concrete move, but a move that
leads to easier play!
a) 16...f5! During the game I was more worried about this simple push which seems to be correct. Black’s
fianchettoed bishops and strong central pawn formation, together with the d4-knight, provide a good basis for
counterplay.

17.fxe5 (17.Kh1!? is a move to consider if you want to avoid the wild complications ahead, however in that case a
really smart-ass question becomes why White allowed the opening to play out in this way in the first place!!)
17...dxe5
[17...Nxe5?!, with the idea of sacrificing a pawn for the pair of bishops, is not sound. 18.Nxd4 cxd4 19.Bxd4 Nf3+
20.gxf3 Bxd4+ 21.Kh1

311
21...Qf6 (21...Qh4 22.exf5 gxf5 23.Be4!! A cold shower on the black position. White simply wins here.) 22.Nd5
Bxd5 23.cxd5 fxe4 24.Bxe4²]
18.Bg5 Nf6!!
[18...Bf6 19.Bh6 Bg7 20.exf5 gxf5 21.Bxg7 Kxg7 22.Rbd1! Leaving the b1-square available as a retreat for the d3-
bishop. The position is very unbalanced. The current threat is to take on f5. The computer prefers White’s position
as it seems that he is just a little faster with his play against Black’s pawn mass in the center.

22...f4 (22...Nb6 23.Nxd4 cxd4 24.Ne2 e4 25.Nf4 Kh8 26.Ne6 Qe7 27.Nxf8 Rxf8 28.Bc2 d3 29.Qc3+ Qg7
30.Qxg7+ Kxg7 31.Bxd3 exd3 32.c5!± The dust has settled and White has managed to transpose into a favorable
endgame.) 23.Be4 Bxe4 24.Nxe4 Qb6 25.Rde1 Rae8 26.Rf2 Kh8 27.N2c3 Rg8 28.Nd5 Qg6 29.Kh1 a4 30.g3!

312
With all his pieces on their best squares, White now goes for the classic counter-push, which gives him an
advantage. Many years later, at the first Sinquefield Cup in 2013, in my first-round game against the reigning World
Champion, Magnus Carlsen, I went for pawn aggression on the kingside as Black, and after one inaccuracy my
opponent responded with a similar counter-push which demonstrated the reality of what was going on.]
19.exf5 e4! 20.Bc2 gxf5!∞

With an absolute mess on the board. Even with computer assistance this position is not easy to evaluate. My gut
feeling is that on the deepest level, the computer will find one of its famed miracle draws.
b) 16...Re8

313
17.f5 (17.fxe5 Nxe2+ 18.Bxe2 Bxe5 The white and black weaknesses balance each other out, resulting in a roughly
equal position.) 17...gxf5 18.Bg5 (18.exf5? e4 19.Bc2 Ne5µ) 18...f6 (18...Bf6 19.Bxf6 Qxf6 20.Nxd4 cxd4 21.Nd5
Bxd5 22.Rxf5±) 19.Bh4 f4 20.Nd5 Bh6³ and White’s pawn sacrifice is not quite sound.

17.Nxf4²

Now White has a pleasant game. I felt quite confident about my chances here as we were now playing for two
results, a white win or a draw.

17...Nb6?!

17...Ne5 is a more active and logical move. Knights are best placed in the center and the advance of White’s b-paw n
is blocked by Black’s b7-bishop. 18.Nfd5 Ne6 19.Be2 Nd4 20.Rbd1 Bc8! Relocating the bishop to a better square.
21.h3 Be6 22.Kh1 Rc8 23.Bg5 f6 24.Be3 Kh8! 25.a4 Re8² The position is obviously in White’s favor due to his
longterm strategic assets, but Black can put up a fight, for example

314
26.Bxd4 cxd4 27.Qxd4 f5 28.Qd2 (28.exf5 Bxf5 29.Qa7 Nxc4 30.Bxc4 Rxc4 31.Rf3 Rh4! An amazing resource, but
it is really difficult to do anything against Black’s strong pair of bishops.) 28...Nxc4 29.Bxc4 Rxc4 30.b6 Rd4
31.Qf2 Rxd1 32.Rxd1 Rf8 33.Rb1 fxe4 34.Qe3 Be5 35.b7 Qb8=

and the threat of ...Qxb7 keeps Black alive.

18.Ncd5?!

If I had known what my opponent was up to, I would probably have played a2–a4 to prevent his plan. My next few
moves were pure prophylaxis, cementing the center and space advantage. However this is inaccurate. The f4-knight
is blocking White’s movement, so it was better to play 18.Nfd5.

315
18...a4!?

The start of an original plan.

18...Nd7!? was an interesting idea, simply highlighting how awkward White’s f4-knight is.

19.Kh1 a3!

Now, suddenly, Black has more options. The b2-square, just made available by the march of the a-pawn, is
potentially very attractive for the black pieces, especially the rook, later in the game.

20.Nxb6 Qxb6 21.Nd5!

21.g4!? Another idea, to prevent Black from playing ...f7–f5. The knight goes to d5 anyway. However Black gets a
crucial tempo to exchange queens, when it will be almost impossible for White to break down Black’s defense.

21...Qa5! 22.Qxa5 (22.Nd5 Qxd2 23.Bxd2 Be5 24.Bh6 Bg7 25.Bf4 Rfd8³; 22.Qf2 f5 23.gxf5 Rxf5∞) 22...Rxa5
316
23.Nd5 Be5 24.g5 Kg7 25.Kg2 Ra4 26.h4 Ne6

27.h5 The only place where White can try to break through is the kingside.
[27.Rh1?! Bxd5 28.exd5 Nf4+ 29.Bxf4 Bxf4 30.b6 Rb8 and suddenly Black might be better. (30...Rb4?! not yet.
31.Rxb4 cxb4 32.Rb1 Bd2 33.c5 dxc5 34.d6 Rb8 35.d7 and only White can push.) 31.Rb5 Be5 32.Rhb1 Bb2 33.Kf3
Rb4³]
27...gxh5 28.Rh1 Kg6 29.Rh4 Nxg5 30.Rbh1 f6 31.Rxh5 Rh8 32.Rh6+ Kg7∞ and Black has some kind of fortress.

21...Bxd5 22.exd5

22...f5

Probably the key position in the game. I have decided to ignore Black’s d4-knight completely and try to open up my
pair of bishops. At this point I was reminded of a game of mine vs. GM Kaidanov, where I had two central knights
in the endgame vs. his two bishops and the bishops slowly tore my position apart, so I decided to adopt a similar
317
strategy and play around Black’s d4-knight.
22...Qa5!? With the idea of transposing into an endgame where Black can successfully establish a defense. 23.Qf2
Keeping the queens on the board is preferable, as White can create potential threats against Black’s king. (23.Qxa5
Rxa5 24.g4 Be5 25.g5 Ra4 26.Bd2 Kg7 27.Kg2 Re8 28.h4 Re7 29.h5 Kg8 30.Rf2 Rb7 31.Kh3 Re7 32.Re1 Raa7
33.Rff1 Ra4 34.Bc3 Raa7 35.Kg2 Kg7∞) 23...Qc3 24.Rbd1

24...f5
a) 24...Qb2 25.Rd2 Qc3 26.Bf4²
b) 24...Rae8 25.Bf4 Be5 (25...Rd8 26.h4 Qb2 27.Rd2 Qc3 28.h5 f5 29.Qg3 and White manages to get through to
Black’s weak pawns.) 26.Bh6!± A typical trick in such positions. White makes use of the unprotected black d6-
pawn.
25.Qd2! Suddenly reversing course. Black’s f-pawn is now on f5, which changes White’s strategy. This pawn can
now be successfully attacked by a g2–g4 push, which in turn will open up White’s d3-bishop and make White’s
prospects in the endgame much higher than in a position with Black’s pawn on f7. (25.h4 Rae8 26.h5 Qb2 27.Rd2
Qc3 28.Bf4 Be5 29.Bh6 Bg7 30.Bg5 f4„) 25...Qb4

318
26.Rb1 Qxd2 27.Bxd2 h5! An example of how this endgame could develop: 28.g3 Kh7 29.Kg2 Bh6 30.Bc3 Bg7
31.Rbe1

31...Rf7 (31...Be5!? 32.h3 Kg7„) 32.h3 Be5 33.g4 hxg4 34.hxg4 Kg7 35.gxf5 gxf5 36.Bxd4 Bxd4 37.Rxf5 Rxf5
38.Bxf5 Ra4 39.Bd3 Rb4

319
Despite White’s best efforts, which have netted him an extra pawn, these massive exchanges have led to a draw
because of bishops of opposite color.

23.Rbe1

23...Rfe8

23...Qa5? is too late as White can now prevent Black’s g7-bishop from going to e5, blockading the whole position.
24.Qxa5 Rxa5 25.Bf4 Rd8 26.Re7 Ra4

320
27.Rc7 (27.Rfe1 Nxb5; 27.g3 Rb4 28.Ra7 Rb2 29.Rxa3) 27...Rb4 28.Ra7 Rb2 29.Rxa3+–

24.Bg5

Going for the g2–g4 break immediately with 24.h3! was better, but in an important tournament like the Interzonal it
is not easy to make such radical decisions during the game. I was trying to play safe and prevent any black
counterplay. 24...Qa5 25.Qxa5 Rxa5

26.Bxd4! Rxe1 27.Rxe1 Bxd4 28.Re8+ Kf7 29.b6!+–

24...Be5

321
25.Qf2

White wants h2–h3 without allowing ...Bg3.

25.h3 Bg3 26.Re3 Be5 27.h4 (27.g4?! fxg4 28.hxg4 Rf8 Suddenly Black takes over the f-file.) 27...h5!∞

25...Qa5 26.Bd2 Qd8

27.g3

Another careful move. The immediate 27.g4 was also possible, but we were running short of time and, as
emphasized earlier, White wants to make the g2–g4 push after sufficient preparation.
27.g4 fxg4 28.Qf7+ Kh8 29.Rxe5 dxe5 30.Bxg6

322
30...Re7 (30...hxg6? is bad. After 31.Qxg6 White’s attack is unstoppable in view of the coming Bg5–f6 mate.)
31.Qf6+ Rg7 32.Be4 Nf3! The only move, transposing into an endgame where he has chances to hold. 33.Qxd8+
Rxd8 34.Be3! h5!

The computer shows that Black is worse, but has sufficient defensive resources. 35.Kg2 (35.Bxc5 Nd2 36.Re1 Nxc4
37.b6 Rb7 38.d6 Rbb8 with advantage for Black.) 35...Nh4+ 36.Kg3 Nf3 37.Bxf3 gxf3+ 38.Kxf3 Rg4 39.Rc1 e4+³

27...Ra7 28.Bh6 Qa5 29.Re3

323
Keeping with the spirit of the game called prophylaxis! Black is forced to stand by and watch White’s preparations.

29...Qd8 30.Qd2 Ra4

Intending ... Rb4–b2.

31.Rb1

Preventing Black’s ...Rb4. Chess is simple!

31...Qb6?

A very strange move which could have been very costly for Black. One excuse is that we were both pressed for time
here and wanted to play safe chess!

32.Bf4?

324
32.g4! Of course. With his last move, Black moved his queen from a perfect defensive position. 32...f4 (32...fxg4?
33.Qd1 Rb4 34.Qxg4 Rxb1+ 35.Bxb1 and the threat of Bxg6 decides the game.) 33.Bxf4 Rf8 34.Bxe5 dxe5 35.Rh3
Now White is simply a pawn up with the superior pawn structure and a winning game.

32...Qb8 33.Kg2?

A mistake that could have been costly. White allows the trade of dark-squared bishops which Black should have
accepted, since the resulting position is a fortress.

33...Kf7?

33...Bxf4 34.Rxe8+ Qxe8 35.Qxf4 (35.gxf4 Ra7=) 35...Qe7 36.b6 (36.Kf2 Kg7 37.Re1 Qd8=) 36...Ra8 37.h4 Rb8=

White cannot make progress.

34.Bg5 Qb6

325
34...Bf6? Black cannot offer the exchange of bishops now, due to 35.Rxe8 Qxe8 36.Bxf6 Kxf6 37.Qh6

when his king is far from safe.

35.Ree1 Raa8 36.h3

Finally my pawn was free to move and I wanted to threaten the g3–g4 break right before Black’s time control, to
make my opponent more nervous and take more drastic defensive measures. The psychological battle continues!

36...Kg8 37.Rf1 Qb7 38.Rf2 Ra4 39.Bh6

326
39...Raa8?!

Black is carelessly shuffling his pieces, as it seems that White is making no progress. However he should have
harassed White’s h6-bishop out of that spot: 39...Bg7! 40.Bf4 Be5 with counterplay.

40.Rbf1

40.g4 Kh8 41.Rbf1 Rg8 42.Kh1 Rae8 43.gxf5 gxf5 44.Bxf5 Qe7‚

This line shows how the dark squares are important. Black’s e5-bishop is now a monster.

40...Bg7 41.Bf4

41.g4 Very tempting, but I didn’t want to exchange the dark-squared bishop. 41...Bxh6 42.Qxh6 Qg7! Now for
instance: 43.Qxg7+ Kxg7 44.gxf5 Re3 45.Bb1 gxf5 46.Bxf5 Rf8! 47.Be6 Rxf2+ 48.Rxf2 Rc3 49.b6 Rxc4 50.b7
Rb4 51.Rf7+ Kg6 52.Kf2 h5 53.Ke3 h4
327
and we have reached a curious position where neither side can make progress. Draw.

41...Be5 42.Bg5 Ra4

43.g4!?

Finally. White has nothing else, so it’s now or never. Black has to solve concrete problems.

43...Rb4?!

Black plays in the spirit of the position, however he creates a weakness on f5 and allows his king to be opened up.
43...fxg4!? Black can actually take the pawn, since his position is very defensible. 44.hxg4 Qd7 45.Kh3 (45.Qd1
Rb4 46.Bc1 Ra8 with counterplay) 45...Rb4 46.Qd1 Rb2 47.Bh6 Bg7=

44.gxf5 gxf5 45.Kh1!

328
Of course White wants to go after Black’s king.

45...Kh8?

45...Rb2 was necessary. 46.Qe3 Qd7!=

A hard-to-see move. Black eyes White’s h3-pawn with his ...f5–f4 push. 47.Bf4 (47.Rxb2 axb2 48.Rg1 Bh2
49.Bh6+ Bxg1 50.Qg5+ Kh8 51.Qf6+ Kg8 52.Qg5+=) 47...Bxf4 48.Qxf4 Kh8 49.Rg1 Rxf2 50.Qxf2 Qf7 51.Qf4
Qf6 and Black manages to hold the position.

46.Rg1?

46.Bxf5! There was nothing wrong with taking the pawn. 46...Rb2 (46...Rg8 47.Qe3 Qg7 48.Bh6 Qe7 49.Bg4+–)
47.Qe3 Rg8

329
48.h4! Protecting everything and keeping the extra pawn. (48.Bg4? Qg7=) 48...Qg7 49.Bd3 Qd7 50.b6+–

46...Qf7 47.Qe3

47...Rg8?!

47...Rb2! It was absolutely necessary for Black to exchange this rook in order to diminish White’s threats on the
kingside. 48.Rfg2 Rxg2 49.Rxg2 Bg7 50.Qg3 Be5 51.Bf4?! (51.Qf2 Qh5 52.h4 Rg8 and the position becomes a
game with all three possible results.) 51...Bxf4 52.Qxf4 Qh5 Black is counterattacking.

48.Rfg2 Rg6 49.Bf4!?±

330
White has finally caught Black a bit off guard and the latter has problems protecting his weak pawns on d6 and f5
while keeping his king safe. Black’s b4-rook is temporarily out of play and this is a major factor.

49...Qg7?

The critical mistake in the game.

49...Bxf4 50.Qxf4 Rxg2 51.Rxg2 Qf8²

was absolutely necessary. 52.Qc1 Ra4 (52...f4 53.Qxa3 f3 54.Rg3 f2 55.Bf1 Rb1 56.Kg2

331
White has an extra pawn, and his a-pawn is a potential monster, while Black has limited activity due to the open
position of his king. If Black doesn’t find a concrete idea, White only has to watch for tactical threats to convert his
material advantage.
56...Ne2™ 57.Qf3! Nf4+ 58.Kxf2 Qf6 59.Kg1 Qd4+ 60.Kh2 h5 Black has practical chances.) 53.Qg5

53...Ra7! With the idea of meeting Rf2 with ...Rf7, protecting the f5-pawn. Black is worse, but it is not clear how
White can break through his defenses. (53...Rb4? 54.Rf2 Rb2 55.Kg2! Placing Black in zugzwang, when he is
forced to give up either his d6- or f5-pawn or create weaknesses around his king. 55...h6 56.Qg6± Qg7 is impossible,
due to 57.Qxg7+ Kxg7 58.Rxb2 axb2 59.b6 and the pawn queens.)

50.Bxe5 dxe5 51.d6!!+–

332
Suddenly it is White’s d-pawn that decides the game. Black’s pieces are too disorganized to mount a sufficient
defense.

51...Rb2

a) 51...Ra4 52.Rxg6 hxg6 53.d7 Ra8 54.Qg5 Qxd7 55.Qxg6

b) 51...e4 52.d7 Ne6 53.Qxc5 Qxd7 54.Qxb4 Qxd3 55.Kh2!+– Cool defense. Black is lost.

52.d7?!

Not the best, but a human move which still wins the game.
52.Bf1! The correct winning move, however, in order to make it work, one had to foresee the following difficult
computer tactics. 52...e4 53.d7! (53.Qxa3? would be a mistake: 53...Nf3!∞.) 53...Qxd7 54.Rxg6 hxg6 55.Rxg6 Ne6

56.Rg1 ( 56.Qh6+? Qh7 57.Rxe6 Qxh6 58.Rxh6+ Kg7 59.Ra6 Rxa2 leads only to a small advantage for White.)
56...Kh7 57.Qxa3 Qd4 58.Qa7+ Ng7 59.Qe7 Rxa2 60.b6 Rb2 61.b7 Rb3 62.Qc7 Qb2 63.b8=Q Rxb8 64.Rxg7+
333
Qxg7 65.Qxb8+–

52...Ne6?

Equivalent to resignation, however the more resilient 52...Qxd7 would also likely have lost the game: 53.Qxe5+
(53.Rxg6? hxg6 54.Rxg6 Qb7+ 55.Kg1 Nf3+ 56.Kf1 Nh2+=) 53...Qg7 54.Qe8+ Qg8 55.Qxg8+ Kxg8

56.Rxb2! (56.Rxg6+ hxg6 57.Rxg6+ Kf7 58.Rg2! Rb4 59.Rf2 Kf6 60.Kg2 Ke5 61.h4 should also be a win for
White.) 56...axb2 57.a4!

The pawn starts to run and decides the game. (57.Rxg6+? would be a terrible mistake; after 57...hxg6 it is White
who is in trouble, since 58.a4 is answered by 58...Nb3µ.) 57...Rxg1+! The most challenging for White.
[57...Nb3 58.Rb1! ( 58.Bxf5!? is winning as well, but the road to a win is a lot trickier. 58...Rxg1+ 59.Kxg1 Na5
60.Kf2 Nxc4 61.Ke2 h6 62.Bc2 Kf7 63.Kd3 Nd6 64.Kc3 Ke6 65.Kxb2 Kd7 66.Kb3 Kc7+– and White has to watch
for tricks, since his h-pawn’s queening square is not the same color as the bishop.) 58...Rd6 (58...Rg3 59.Bf1 Na5
60.Rxb2 Ra3 61.b6 Ra1 62.Kg2 Nb7 63.Kf2+–) 59.Bc2 Na5 60.Rxb2 Rd2 61.b6

334
61...Nb7 (61...Kf8 62.Rb5+–) 62.a5 Nxa5 63.Ra2 Nb7 64.Ra8+ Kg7 65.Bxf5 Nd6 66.Rd8 Rd4 67.Bc8 Nxc8
68.Rxd4 Nxb6 69.Rg4++–]
58.Kxg1 Nf3+ 59.Kf2 Ne5 60.Bxf5 Nxc4 61.Ke2 Kg7 62.Bc2 Na5 63.Kd2 Kf6 64.Kc3 h6 65.Kxb2 Kg5 66.Kc3
Kf4

White is winning, but still has to work a bit. 67.Bg6!? Ke5 68.Be8 Kd5 69.Bf7+ Ke5 70.Kd3 (70.Bg8 Ke4 with
counterplay) 70...Kf4 71.Bd5 Ke5 72.Be4 Kd6 73.b6

335
This is an ideal set-up for White.
73...Kd7 74.Bd5 Kd6 75.Ke4 c4 76.b7 Kc7 77.Kd4+–

53.Rxg6 hxg6 54.Bxf5

And Black resigned. 1–0

This game gave me a huge confidence boost and I went on to play some inspiring chess in the rest of the tournament,
qualifying for the Candidates.
However what I didn’t know at the time was that it was to be the start of not only one World Championship cycle,
but two!
My memory is not very sharp, but I think it was at the end of this tournament that all the players were invited to a
meeting with a certain Bob Rice, who proceeded to give vague notions of upcoming changes to the World
Championship cycle and a possible new Interzonal tournament. I remember seeing all the incredulous faces around
336
me and it struck me as unusual.

337
 GAME 13
Gata Kamsky
Vladimir Kramnik
New York 1994
This match was part of the World Championship cycle organized by Bob Rice, an attorney working for Mr.
Kasparov who had quit FIDE (the International Chess Federation) right before his World Championship match vs.
Nigel Short of England in 1993. Mr. Kasparov successfully defended his title, and, seeking legitimacy and
acceptance from the rest of the chess world, decided to start a rival world chess organization that would focus mostly
on chess professionals, the grandmasters. The organization was called the Professional Chess Association, or PCA
for short. In fact this PCA was an echo of an earlier attempt by Mr. Kasparov and Mr. Bessel Kok of Belgium to
organize another rival chess federation, the Grandmaster Association (GMA), but it was short-lived after staging a
number of prestigious GMA Grand Prix tournaments. This split in chess lasted for more than a decade, with FIDE
holding its own World Championship cycle (where I managed to get through the Candidates matches and played the
FIDE World Chess Championship match vs. Anatoly Karpov in 1996) with its own World Champion title. It was a
question of legitimacy, because for everyone it was obvious that Mr. Kasparov, being World Champion prior to the
split, was the best player in the world at that moment. The PCA held its own qualification series, culminating in the
World Championship match between Mr. Kasparov and Mr. Anand at the World Trade Center in New York in 1995,
but decided to reduce costs and only staged the latter half of the cycle, starting with the Interzonal event at
Groningen in 1993. After successfully qualifying from that event as well, I was paired with Mr. Kramnik in my first
PCA Candidates match. Something needs to be said about the organization of this match. Unlike in previous years,
all the Candidates matches were staged together in one location, Trump Tower in New York. The first round was
played inside the giant lobby of the hotel, in close proximity to shops, cafes and large numbers of visitors. Needless
to say, we were all pretty much in shock, given how noisy the place was, being accustomed to the usual silence of
the typical playing hall. After the first round and numerous protests, fearing a backlash, Mr. Bob Rice, as part of the
organizing committee, made a decision to put us in a glass cage in a small room adjacent to a lobby, but on the
fourth or fifth floor. Imagine a room, approximately 2.5 meters by 12 meters, cut diagonally across, leaving only one
half usable by the players. The effect, of course, was tremendous. The air conditioners couldn’t keep the room well
enough ventilated, and if the players wanted fresh air, they had to ask for the door to be open, which would then
invite the noise of the lobby in. Some fifteen to twenty die-hard spectators had to stand outside and peer through the
glass to see what was happening on the boards. The World Championship Candidates tournament of 2018 looked
like a palace compared to the conditions we played under. Apparently both the players and their seconds had figured
out that if they continued to complain, the players would be forced into even worse conditions, so it was decided to
keep the location and just get through the event as fast as possible.

338
1.d4 d5 2.c4 c6

My opponent employs the Slav Defense, just like the previous time during the World Team Championship in
Lucerne. There I avoided the ultra-sharp main lines of the Botvinnik system, as I feared my opponent’s preparation,
but now I had spent approximately a week before the match studying different lines of the Botvinnik to acquaint
myself with the general ideas and structures.

3.Nc3 Nf6 4.Nf3 e6 5.Bg5 dxc4

339
So the Botvinnik it is, quite a popular system at the time. I have lost and won some exciting games in this line, one
of the more memorable losses coming in my game vs. Shirov, from the same above-mentioned World Team
Championship, who sprang quite an impressive piece of preparation on me. However the story did not end there, as I
used the exact same line as Black a few days later to beat my opponent in a Tilburg knockout tournament.
These days, of course, it is impossible to do so, since games are often broadcast online and immediately included in
the ChessBase database, thus becoming known the same day they are played. But in those days, when news of
games didn’t travel so fast, it was quite an adventure!

6.e4 b5 7.e5

Invitation accepted. Now we’re going to see whose preparation was deeper.

7...h6 8.Bh4 g5 9.Nxg5 hxg5 10.Bxg5 Nbd7

Another popular and tricky move is 10...Be7. I’m going to skip the analysis, and will just say that a lot of rather
concrete material needs to be studied here as the structures and positions arising in this line, as for all Botvinnik
positions, are so unbalanced as to make general observations quite irrelevant.

11.exf6

The old main line. The second main line for White was considered to be 11.g3, attempting to delay Black’s ...c6–c5
breakthrough by developing White’s f1-bishop to g2 first. However the positions arising after 11...Rg8 12.h4 Rxg5
are still quite difficult to evaluate — for humans of course, as computers have absolutely no problems.

11...Bb7 12.g3

340
12...c5

12...Qb6 has been played, with the idea of meeting 13.Bg2 with 13...0-0-0 and 14.0-0 with 14...Ne5. It requires long
analysis which would most likely need constant updating, just like Microsoft Windows!

13.d5 Qb6

Black has also tried 13...Bh6 and 13...Nf6 with quite exciting play.

14.Bg2 0-0-0 15.0-0

Hopping along the main line, we rapidly banged out these moves to reach this critical juncture. Now I was curious as
to what my opponent would do.

15...b4!?

341
Logical, and considered to be the main move for Black, attempting to move White’s strong c3-knight which not only
protects d5 but is ideally poised for both an invasion of the queenside and defense in the center. Black tries to hit the
knight before White plays a2–a4, which would make Black’s life complicated.
Other moves tried by Black here are 15...Bh6 and 15...Ne5, but after the text move we have arrived at another
critical juncture, as White has a choice between the move in the game and the paradoxical 16.Rb1!. White’s knight is
tactically protected and he intends to play straightforward moves like Re1 or dxe6.
Many games have been played in this line, even in modern times, where of course every thing has been analyzed
until almost the very end, with computers confirming the suspicion that Black is OK and the whole line is very close
to a forced draw. Fortunately for us it requires so much effort to memorize these lines that no sane human would
want to do this in every game, and would therefore pause and deviate earlier.

16.Na4

Because we didn’t have very strong computer assistance at the time, analyzing these mostly highly tactical positions,
hoping for the best and not to blunder something during the analysis, was quite a task. During my preparation, I was
not sure whether to employ the text move or 16.Rb1, but decided to play 16.Na4, simply because it looked more
convincing in view of a recent game by Salov vs. Illescas (Madrid, 1993) which White had won in spectacular
fashion. I didn’t analyze the particular line, however, not believing that Mr. Kramnik would employ it. I was wrong.

16...Qb5

Black’s other possibility is 16...Qa6 with a lot of forced lines there as well.

17.a3

17...Ne5?!

At the time this was considered to be almost identical to 17...Nb8 which was played in the above-mentioned
predecessor, Salov-Illescas. However modern computers disagree and state that 17...Nb8 is more accurate as the text
move allows the following line. Just a small observation — 17...exd5 has also been played quite a few times, but the
general verdict is that White keeps the advantage there.

18.axb4

342
18.Qe2!? A relatively new move, not much played, but the computer’s first suggestion. 18...Nd3 (18...Qxa4? An
obvious mistake. 19.axb4 Qxb4 20.Qxe5 Bxd5 21.Rxa7+–; 18...Bd6 19.axb4 cxb4 20.f4 Nd3 21.dxe6 Bxg2
22.Kxg2 Qd5+ 23.Rf3 fxe6 24.b3+– with an almost winning position for White.) 19.dxe6 Bxg2 20.Kxg2 Qc6+

21.Kg1
[21.f3 This natural-looking move borders on dubious as it weakens the white king’s protection. 21...Rd5!! 22.e7
(22.h4 Re5! The point. Black is winning.) 22...Bxe7 23.fxe7 (23.Qxe7 Qxa4! 24.b3 Qxb3 25.Bh4 Qc2+ 26.Kh1 Nf4!
27.Rg1 Ne6! 28.Qxa7 Nc7!!–+

A magnificent knight tour that eliminates all potential perpetual checks to the black king and launches a devastating
counterattack.) 23...Rxg5 24.Qe4 Kc7

343
25.Nxc5! The only move to hold the balance. 25...Rxc5 26.axb4 Qxe4 27.fxe4 Re5 28.Rxa7+ Kb6 29.Rd7 Re8
30.b3 Nxb4 31.bxc4 Kc6 32.Ra7 R8xe7 33.Rxf7 Rxf7 34.Rxf7 Rxe4 with a draw.]
21...fxe6 (21...Rd5? 22.e7! Bxe7 23.Qxe7 Rxg5 24.Qxa7+–) 22.axb4 cxb4 23.h4 Rd5

and the position remains quite complicated with chances for both sides.

18...cxb4 19.Qd4!? Nc6 20.dxc6!

A seemingly spectacular positional piece sacrifice which is actually forced, otherwise White has to give up his
central d5-pawn for no compensation, while Black has improved his d7-knight to the c6-square.

344
Does being forced mean it’s bad? The short answer is no, but in modern chess, computers, which are the ultimate
authority on forced lines, give the resulting positions an equal sign! The move had actually been played before, the
general idea being that White gets a really far advanced protected passed pawn on b7 plus two minor pieces for the
queen. However the passed pawn cannot promote on its own, so has to be helped, ideally with White’s dark-squared
bishop supporting it via the a7-square, or the knight somehow hopping to the adjacent a6or d7-squares. Because one
check is all it takes to queen the pawn, or to regain material equality with a serious pawn advantage on the
queenside, White actually hopes to win the game here.

20...Rxd4 21.cxb7+

And now we come to the first critical point of the queen sacrifice line.
With hindsight and the benefit of modern computers, it is easy to say that 21...Kb8 is a much better alternative,
however at the time the game was played the plan employed by my opponent seemed quite logical. Instead of
blocking the passed pawn with the king and always having to be on guard for potential promotion with check, why
not utilize the absent h8-rook, especially since it has nothing to do on the kingside, and make it block the pawn
345
instead!? However computer analysis shows that it is not feasible, and indeed 21...Kb8 holds the balance.

21...Kc7?

21...Kb8! 22.Be3 Several attempts at improvement have been made by Black here, but with no success. Examples:

22...e5! The strongest and best reply for Black.

a) 22...Rd5? 23.Bxa7+! Kxb7 (23...Kxa7 24.Nc5+ Kb8 25.Ra8+ Kc7 26.b8=Q+ Qxb8 27.Na6+ Kb7 28.Bxd5+ exd5
29.Rxb8+ Kxa6 30.Ra1#) 24.Nb6 Kc7 25.Nxc4!!

Taking the pawn that protects the black king is much more important than the exchange. Besides, Black would have
been happy to give up his rook for White’s mighty g2-bishop. Black cannot escape the threats to his king without
giving up a lot of material: 25...Rhh5 (25...Bc5 26.Rfc1 Bxa7 27.Bxd5 Kb8 28.Ra5 Qxd5 29.Rxd5 exd5 30.Ne5+–)
26.Rfc1 Rc5 27.b3!!

346
Such little pawn moves add an especially pleasing aesthetic touch to positions where the side with less material is
dominating the side with a useless material advantage. Black is tied up and can only await White’s decision as to life
and death. 27...Rhf5 (27...Bd6 28.Bxc5 Rxc5 29.Ra7+ Kb8 30.Rca1 Rxc4 31.bxc4 Qe5 32.Rb7+ Kc8 33.Ra8++–)
28.Bxc5 Bxc5 29.Ne3 Rxf6 30.Bf1 Qb6 31.Ra6+–.
b) 22...Bc5 is another try. 23.Nxc5 Qxc5 24.Ra6

24...e5 (If 24...Rhd8 25.h4 Qe5 26.Rfa1! Rd1+ 27.Rxd1 Rxd1+ 28.Kh2 Rd4 29.Bxd4 Qxd4 30.Rc6!+–

347
This motif of a discovered check battery is standard in this position. Black has no resources to defend against it.)
25.h4! It is important to make a luft for the king before assaulting Black’s weak link in the pawn chain on the
queenside, the c4–pawn. 25...c3 26.Rc6 Qa5 27.bxc3 bxc3 28.Rc1

and Black is in big trouble, e.g. 28...Rhd8 29.R1xc3 e4 30.Bf1! Rd1 31.Rb3+–.
23.b3!
[Sacrificing the a4-knight with 23.Nc3 doesn’t work in this situation: 23...bxc3 24.bxc3 Bc5 (24...Bh6!?) 25.Rfb1
(25.cxd4 Bxd4 26.Rfb1 Qc5µ) 25...Rd1+! 26.Rxd1 Bxe3 27.fxe3 Qb2µ and the black king is safer than its white
counterpart.]
23...c3 Forced. Black cannot allow either the a- or c-files to be opened for use by the white rooks.
[23...cxb3??

348
This move is a suicide as it opens the c-file for White’s rooks to get access to the c8-square. 24.Rfc1 Bd6 25.Bxd4
exd4 26.Rc6 Rd8 (26...Be5 27.Nc5+– and the knight reaches the mating squares.) 27.Rac1 Qxb7 28.Rxd6 Rxd6
29.Bxb7 Kxb7 30.Nc5+ Kb6 31.Nxb3+–

White has to play accurately to stop the march of the black pawns, but if he does he wins the game thanks to his own
strong passed pawns on the kingside. 31...a5 32.Ra1 Kb5 (32...d3 33.Kf1 Despite a clear material advantage for
White, these endgames have to be played with care, as the black pawns, especially the connected aand b-passers, can
become quite dangerous if not controlled early. 33...Kb5 34.Rxa5+ Kc4 35.Nd2++–) 33.Rxa5+ Kc4 34.Nd2+ Kd3
35.Nf3 b3 36.Ne5+

349
36...Kc2 (36...Kc3 37.Rc5+!! The only and key winning move.) 37.Rc5+ Kd1 38.Rb5 Kc2 39.Kg2 d3 40.Nc4 Rxf6
41.Ne3+ Kb2 42.h4 Re6 43.Kf3 Kc3 44.h5+–

and the black pawns have been stopped, while White advances his own pawn majority to win the game.]
24.Rfd1!

350
Black’s d4-rook is all that stands between the white pieces, the black king and the promotion of the b7-pawn, but
White needs to keep the e3-bishop alive. 24...Bh6 The only move to hold the game. The following moves are forced
and lead to a draw. 25.Bxd4 (25.Rxd4?! This is playing with fire. Black’s c3-pawn is not too far from queening...
25...exd4 26.Bxd4 c2 27.Nc5 Re8 28.Bf1 Qc6 29.Na6+ Kxb7 30.Bg2 c1=Q+ 31.Rxc1 Bxc1 32.Bxc6+ Kxc6
33.Nxb4+ Kb5µ and White is fighting for survival.) 25...exd4 26.Rxd4 c2

27.Rc4
a) 27.f4? has been tried by some adventurous people, but here Black even wins: 27...Bxf4! 28.Rxf4 Qe5 29.Raf1
(29.Rff1 Qe3+ 30.Kh1 Qxg3 31.h3 Qxb3 32.Nc5 Rxh3+ 33.Bxh3 Qxh3+ 34.Kg1 Qe3+ 35.Kg2 Qxc5+–) 29...Qe3+
30.R4f2 c1=Q 31.Rxc1 Qxc1+ 32.Rf1 Qe3+–+.
b) 27.Rd5? with the idea of Nc5 fails to 27...Qe2 28.Nc5

351
28...Be3!! 29.fxe3 (29.Nd7+ Kc7–+) 29...Qxe3+ 30.Kf1 c1=Q+ 31.Rxc1 Qxc1+ 32.Kf2 Qb2+ 33.Kg1 Qa1+ 34.Kf2
a6!–+ and with his king out of the box, Black is winning.
Back to 27.Rc4.

27...Qxc4!! The correct move, as has been demonstrated in several games. Black’s connected passed pawns cause
White enough concern to agree a draw.
[27...c1=Q+? A bad mistake. White gives up an exchange in reply, but gets an unshakable grip on the position, when
he’ll be able to improve and break through Black’s defenses to win the game. 28.Raxc1 Bxc1 29.Rxc1+–

352
The combination of the passed b- and h-pawns and mating threats to the black king give White a decisive advantage.
29...Qd3 30.h3!! The final nuance! White needs to make his king safe before the rook liftoff. Black is completely
tied up. (30.h4? Qd2 31.Rc6 Qe1+ 32.Kh2 Qxf2 33.Nc5 Rxh4+ 34.gxh4 Qxh4+ 35.Bh3 Qf2+ with a draw.) 30...Qd2
(30...Qxb3 loses on the spot to 31.Nc5+–; 30...Re8 31.Rc6 Qxb3 32.Nc5 Re1+ 33.Kh2+–) 31.Rc6 Qe1+ 32.Kh2
Qxf2 33.Nc5+– Here Black has no perpetual and no defense to the discovered check.]
28.bxc4 Rd8!

The only move to keep the balance. I’ve made an extensive analysis of this exciting endgame, but have not included
it out of space considerations in this book. I would, however, recommend the reader to attempt to analyze this
position for himself, as it will undoubtedly improve his endgame knowledge.

22.Be3 e5

353
23.Nc3! bxc3 24.bxc3 Bc5!?

A novelty which Salov had annotated with a few lines in his game vs. Illescas. I had not seen these annotations and
was annoyed with myself during the game for not having checked the line and now having to work hard with limited
time, while my opponent was blitzing his moves out and enjoying a stroll after each move. In that game, his
opponent retreated his rook to d6: 24...Rd6 25.Rab1 a6 26.Rxb5 axb5 27.Ra1 Rd8 28.Be4 Bh6 29.Bc5+–.

The black pawns are weak and his pieces are dominated. White went on to win in Salov-Illescas, Madrid 1993.

25.cxd4

The rook has to be taken immediately, since 25.Rfb1, as suggested by Salov, gives White no advantage, although the
lines can be incredibly nuanced:

354
25...Rd1+! 26.Rxd1 Bxe3 27.fxe3 Qb6 28.Kh1 Qxe3 29.Ra5 White needs to infiltrate with his rook somewhere.
29...Kb8 30.Rad5 e4 31.R5d4

31...Qe2! The simplest. Black ties up White’s pieces and forces the repetition. Playing for a win with Black is a risky
proposition, as the following examples suggest.
[31...Re8 32.Rxc4 Qe2 33.Rcd4 a5 34.R4d2 Qc4 35.Rd7!? Trying to trick Black with Re7 ideas. 35...Qe2!= Keeping
the white rooks tied together. 36.h3 Rg8 37.g4 Rh8! 38.Kh2 Qe3 39.Rf1 Qxc3 40.Re7 a4

355
41.g5?! Going for an exciting pawn race! (41.Rxe4!? a3 42.Ra4 Rxh3+!! 43.Kg1 Qc5+ 44.Rf2 Qc1+ 45.Rf1= is a
perpetual.) 41...a3 42.g6 a2 43.gxf7 a1=Q 44.Rxa1 Qxa1

and White can give a check on e8 and take the h8-rook with a draw, but even 45.Rxe4!! would be fine. This line
demonstrates exactly how strong his passed pawns are, especially since they are not connected. Here Black has
nothing better than a draw: 45...Kc7 46.Re8 Qb2 47.Rxh8 Qe5+ 48.Kg1 Qd4+ 49.Kh1 Qa1+ 50.Kh2 Qe5+=.]
32.R4d2 Qe3 33.Rd4=

25...Bxd4

356
25...exd4 26.Bf4+ Bd6 27.Bxd6+ Kxd6 28.Rfb1+–

26.Rfb1!

26.Bxd4? Even though it’s possible, I didn’t consider this move as I didn’t want to give up my e3-bishop. 26...exd4
27.Rfb1? This move doesn’t work here. (27.Rxa7 Rb8µ) 27...Qc5 28.Ra6

28...d3! (Not 28...Rb8? 29.Re1+– and White wins again!! This is a perfect example of how complex these positions
are: one wrong move in a seemingly winning position and you can be lost instead!!) 29.Rc6+ Qxc6 30.Bxc6 Kxc6
31.b8=Q Rxb8 32.Rxb8 d2 33.Rd8 c3–+ and one of the pawns queens.

26...Qc5

Now I went into a long think, by some accounts for over an hour, as I felt that there must be something in this
position for White, but what? And where? Black’s d4-bishop is, of course, his pride and joy, blocking my e3-bishop,
defending his king and holding his center and passed c-pawn together. If I don’t do something fast, Black’s c-pawn
will roll down to c3 and c2, bringing about my demise.
357
26...Qd7?? 27.Bxd4 Qxd4 28.Ra4+–

White gets everything he wanted: connected rooks, protected passed pawn and weak black c4-pawn which, when it
falls, will herald Black’s doom.

27.Ra6

A useful tempo which sets up the whole cascading sequence of moves that follows.

27...Rb8

Again played quickly. Since White threatens Rc6+, recapturing the queen, Black has no time to get rid of my e3-
bishop, as the following line shows: 27...Bxe3? 28.Rc6+ Qxc6 29.Bxc6

358
29...Bb6
a) 29...Bg5 30.Bd5 c3 31.Rb3 Bd2 32.Rb4! Kb8 33.Rc4!

and the rook is placed on an ideal square. It keeps the black h8-rook on the back rank because of the Rc8+ threat,
keeps the d2-bishop tied to the defense of the c3-pawn and prevents Black’s a-pawn from getting past the a5-square.
White simply pushes his kingside pawns and wins the game.
b) 29...Kxc6?! 30.b8=Q (30.fxe3?! Rb8 31.Kf2 c3 32.Ke2 c2 33.Rc1 Rxb7 34.Rxc2+ Kd5 and Black might hold the
draw.) 30...Rxb8 31.Rxb8 Bd4 32.Kf1 c3 33.Ke2+– and White obviously wins, since his king blocks Black’s passed
pawns while keeping a decisive material advantage.
30.Bd5!

359
An important nuance. White needs to provoke a black pawn to move, so he can fix them on dark squares. 30...c3
31.Rb3 Ba5 (31...Bd4 32.Rb4+– as demonstrated above.) 32.Be4! and Black is tied up because he cannot dislodge
White’s e4-bishop. 32...Kb8 33.h4 Rd8 34.h5

34...Rd4 (34...Rd6 35.h6 Rxf6 36.h7 Rh6 37.Rb5 Bc7 38.Rc5 a5 39.Rxc3+–) 35.h6! Rxe4 36.h7+– White queens
first with check.

28.Bc1!!

360
And here is the move of my life that makes this the game of my life. Even today I’m very proud that I was able to
find the idea and calculate the resulting sequence of moves almost to the end, all during the game. Since Black’s
novelty 24...Bc5 I had been on my own. Of course it’s quite humbling to see now how the computer relegates my
move to a mere second-best line (the best being the simple 28.Bxd4!!). The beauty of 28.Bc1 lies in the way the e3-
bishop ricochets off the edges of the board to get past its counterpart on d4, when Black’s position simply crumbles.
Here Black went into a long, long think and I vividly recall the shock on my opponent’s face when he realized that
he was simply lost.
28.Bxd4!? is the computer’s first line.

Of course it’s not easy to assess, especially during the game, when time is low, and considering that my opponent
had prepared this line only to finish up in the ensuing mess. 28...exd4 (28...Qxd4? loses on the spot: 29.Rc6+ Kd7
30.Rc8 Rxb7 31.Bh3+ Kd6 32.Rxb7 Qd1+ 33.Bf1+–.) 29.Re1! The point. 29...Rd8 30.Re7+ Kb8 31.Rc6

361
31...Qxc6 ( 31...Qb4 32.h4 d3 33.Rec7+– and White gets in his favorite check.) 32.Bxc6 The race is on. Of course,
with the computer at hand, it’s easy to see that White stops the black pawns and wins the game. 32...c3 (32...d3
33.Bf3 c3 34.Re5! c2 35.Rc5+– Black’s pawns are stopped.) 33.Re5! The only move, but sufficient to stymie the
black pawns. 33...d3 34.Bf3 c2 35.Rc5 a5! Black utilizes his last trump: the a-pawn starts rolling. 36.Kg2 a4

37.Rc3!! The white mega-rook holds all the pawns. Black is lost. 37...a3 38.Rxa3 c1=Q 39.Ra8+ Kc7 40.Rxd8+–
This looks more like a technical win, however if I had played like this, I would not have created the aesthetics that
happened on the board during the game.

28...c3

The most logical and resilient defensive resource. Other moves lose more quickly.
28...Bxf2+

362
29.Kh1 (29.Kf1? would be a mistake: 29...Qd4 30.Ba3 Qd3+ 31.Kxf2 Qc2+! and Black draws, since the white king
cannot go back to the first rank as his rook will be captured with check.) 29...Kd7 Trying to avoid the mating net
does not work: 30.Ba3 Qd4 31.Bf3!

31...e4 ( 31...Ke8 32.Rd6+–; 31...c3 32.Rd1 Qxd1+ 33.Bxd1 Rxb7 34.Bg4+ Ke8 35.Rc6 and in addition to his extra
material, White weaves a mating net around the black king.) 32.Bg4+ Kc7 33.Rd1

363
The Rd7 mate pattern forces Black to give up his queen: 33...Qxd1+ 34.Bxd1 Kxb7 35.Ra4 c3 36.Rxe4 Rd8 37.Bb3
Rd2 38.Re7+ Kc6 39.Rxf7+–.

29.Ba3

During the game, White’s darksquared bishop was the source of some amusement for me. First, the law of physics
about the movement of an object along the circumference of a circle gaining acceleration with the result that it gains
energy and momentum to create the most damage upon impact. Second was the sudden entrance of the movie
Boomerang into my mind. Not sure what psychoanalysts would say about me because of these thoughts, but it was
fun.

29...Qc4 30.Bd6+

The game now enters a sequence of forced moves leading inexorably to its logical conclusion.

30...Kd7
364
30...Kd8 31.Bxb8 Qxa6 32.Bc7++– wins immediately.

31.Bc6+!?

I decided to try to capture the black queen to get a comfortable winning material advantage, but it was not to be.

31...Ke6!

The exclamation mark is for my opponent’s willingness to go along with the king hunt.
a) 31...Kxd6 32.Bb5+ Kc5 33.Bxc4 Kxc4 34.Rc6+ Kd3

This would have posed more problems for me, however White should win without risk. 35.h4 c2 (35...a5 36.h5 a4
37.h6 a3 38.h7 a2 39.Rd1+ Ke2 40.Ra1 Rxb7 41.Rxa2+) 36.Rc1 Rxb7 37.R1xc2 e4 38.Ra2+– and White has a
decisive material advantage.
b) 31...Qxc6 32.Rxc6 Kxc6 33.Bxb8+– is simply lost.

365
32.Bb5!+–

The point, and also the only winning move. The major black pieces are under attack. The position looks very messy,
especially after Black’s next move, and had to be foreseen before 27.Ra6.

32...Bxf2+!

Black had queen and rook under direct attack, yet he sacrifices his bishop!! What nerves!

33.Kxf2

It looks very logical to capture a free bishop, but at the same time it is the only move to win. Playing chicken would
have given Black a draw.
33.Kg2?? Qd5+ 34.Kxf2 Rxb7!= Black is absolutely fine as there is no useful discovered check.

33...Qd4+ 34.Kf1!

The white king escapes the black queen’s checks, and now White has a decisive material advantage in addition to a
mating attack.

34...Qe4!?

Using his last chance.

34...c2 35.Bxb8+ Kf5 36.Re1+– would win for White.

35.Re1!

366
A cold-blooded response. Just like in my game vs. Shirov from Dortmund annotated above, I wanted to show my
opponent that I did not fear him and would take everything he threw at me and survive. It might look like a small
psychological measure, but in the context of a super-important Candidates match, especially the first game, it usually
sets the tone for the rest of the contest, which speaks volumes. A game of chess is like a contact sport, where, after a
certain time, both combatants start to be on a same wavelength, sensing each other’s possible moves and thought
processes. It is therefore vital for one side to be in control of the proceedings, to gain that extra edge in a battle of
otherwise equally skilled opponents. Every little bit counts!

35...Qh1+

a) 35...Qf3+ 36.Kg1 Rxb7 37.Bc6+–

b) 35...Qxb7 36.Rxe5+ Kxf6 37.Bxb8++–

36.Kf2 Qxh2+ 37.Kf3

Black has a last check, then decides he has had enough.

367
37...Rxb7

37...Qh5+ 38.Kg2+–

38.Bxe5+ Rb6 39.Bc4+ Kd7

39...Kf5 40.g4+ Another pleasing aesthetic touch. The little guy moves with check and Black’s mighty queen falls.

40.Rxa7+ Kc8 41.Rc7+

Black finally resigned. 1–0

Since this was the first game of our match, the psychological effect was particularly inspiring for me and depressing
for my opponent. If he couldn’t trust his preparation, what could he play then? Perhaps a less forcing, less tactical
opening? The next game of our match was about to answer this question.

368
 GAME 14
Vladimir Kramnik
Gata Kamsky
New York 1994
This was the second game of our match and was most likely crucial to my overall victory. I am not sure now why I
was unprepared for the opening line that occurred, but I remember that I tried to dig in and defend my inferior
position to the best of my ability. In hindsight, it seems that my opponent, who at that time was famous for his
positional and strategic wins, decided during the game to beat me in attacking fashion to avenge his spectacular loss
in the first game.

1.Nf3 Nf6 2.c4

With this move order, White attempts to prevent me from employing my main defense against 1.d4, the Grünfeld.
Both in those days and in modern times, the defense, with its far-too-extensive theory, is considered to be
unbreakable.

2...e6

And I decided to go for a Nimzolike structure, hoping to throw my opponent off his preparation, but it didn’t work.
Many years later, I would go for a Grünfeld-like structure with 2...g6, e.g. 2...g6 3.Nc3 d5 4.cxd5 Nxd5, a position
which would become a rich field for all sorts of experiments by White, starting with 5.h4, and then different white
queen moves, including the famous 5.Qc2 employed by Topalov in his 2011 Candidates match against me in Kazan.
In general, when a player decides to make the Grünfeld his main opening, he has quite a task in front of him, as there
are so many possible lines and sidelines to prepare, prepare against and memorize that such a choice of opening
leaves little time to study anything else.

3.Nc3 Bb4 4.g3

Just like in modern times, the systems with g3 are very safe for White, offering comfortable play and hope of a small
but lasting advantage.

4...0-0 5.Bg2 c5?!

I continued my bluff, hoping to reach typical blocked Nimzo-style positions by preparing ...d7–d6 and ...e6–e5. If
369
Black achieves such structures he will have equality, but White does not plan on making Black’s life easy.
5...d5 is by far the main line for Black.

6.0-0 Nc6

7.d4

White decides to transpose into a superior version of Nimzo mainlines. Other possible moves include 7.Na4 and
7.d3, going for semiclosed types of positions that do not seem to pose Black serious problems.

7...cxd4

a) 7...b6? This provocative move does not work well for Black, e.g. 8.d5 Na5 9.Ng5 Ba6 10.Nce4

10...Bxc4 (10...Nxe4 11.Nxe4 Nxc4 12.a3 Ba5 13.b4+– and White wins one of the Black’s pinned bishops.) 11.dxe6
fxe6 12.a3+– capturing Black’s b4-bishop.
370
b) 7...d5 seems to be the computer’s first line, however Black must be prepared to go into a structure with an
isolated d5-pawn where White will still be pressing for an advantage.

8.cxd5 exd5 9.Bg5 Be6 10.dxc5 Bxc5 11.Rc1 Be7 12.Nd4 reaches a standard position from the Tarrasch Defense
where Black is slightly worse.

8.Nxd4

We have reached the main tabiya of this line. There are many ways for Black to try to develop his pieces, but
statistics show that White keeps an advantage everywhere, and why shouldn’t he? His strong fianchettoed light-
squared bishop completely paralyzes Black’s queenside and, by playing ...e7–e6, Black further created certain dark-
squared holes in his position, including the obvious one on the d6-square.

8...Qe7

Alternatives for Black were 8...h6, 8...a6 and 8...Ne5, amongst others. We are not going to analyze them, as this
371
tabiya seems to be still quite popular in modern times.

9.Nc2!

Mr. Kramnik demonstrates his high-level preparation. White’s idea is to make Black decide what to do with his
bishop, a decision which will be irrevocable, so Black must take care. Some examples:

9...Bxc3

a) 9...Ba5?! 10.Bf4 (10.Bd2!? A superior move. Black is unable to play the freeing ...d7–d5, and therefore nothing
prevents White from taking over the center with 11.e4 next. As a result, Black will suffer a serious lack of
coordination among his pieces, especially the badly placed a5-bishop. White will also have the flexibility to choose
from several plans, all of which offer superior prospects due to his better development and greater space for
operations. 10...d6 11.e4²) 10...d5

11.e4!? A more forcing move. (11.cxd5 exd5 12.Bg5!? While the pin from White’s g5-bishop looks quite unpleasant
for Black, he is not without defensive resources. 12...Bxc3 13.bxc3 Be6 14.Rb1 Rab8 15.Nd4 h6 16.Bxf6 Qxf6
17.e3² with the small standard advantage for White in such structures.) 11...Bxc3 12.bxc3 e5 13.Bc1 dxe4 14.Ba3
Bg4 15.Qb1!? (Stronger than going into the endgame after 15.Bxe7)

372
15...Qc7 16.Bxf8 Rxf8 17.Re1 Bf3 18.Ne3 Bxg2 19.Kxg2 Nd8 20.Nd5 Nxd5 21.cxd5 f5 22.Qb3 Qd6 (22...Nf7
23.d6±) 23.Rad1 b6 24.Qa4 Nb7 25.Qc6 Rf7 26.Qe8+ Rf8

27.Qe6+! Kh8 28.a4! g6 29.a5!+– White breaks through Black’s incomplete fortress.
b) 9...Bc5 10.e4 d6 11.Rb1 a5 12.a3²

373
The threat of White’s coming b2–b4 push forces Black to put his a-pawn on the vulnerable a4-square, where it will
be very weak in any subsequent endgame, thus forcing Black to be very careful in making his decisions. 12...a4
13.Be3! e5?! immediately is bad. (13...h6 14.f4 The point. White starts to roll his central pawns, especially with h6
being a possible future target for White’s expansion with g3–g4–g5.) 14.Bg5± and the pin is deadly.

10.bxc3 Rd8

Here I felt uneasy, but, at the same time, White must create something quickly before Black completes his
development and White’s doubled c-pawns become a serious liability.

11.Ba3

A natural move. White prevents Black from developing in standard Nimzo fashion with ...b6 and ...Ba6 by taking
the supporting d7-pawn away from the c6-knight.

11...d6
374
Forced. White now has the pleasant choice of several candidate moves.
11...Qe8?! 12.Bd6 b6 13.Bc7 Bb7 14.Bxc6! Bxc6 15.Bxd8 Qxd8 16.Qd4 Qc7 17.Ne3± and Black has insufficient
compensation for the lost exchange.

12.Rb1

I was more worried about 12.Qd2, as Black will have to place his f6-knight very passively on e8, while White has
the two bishops in a semi-open position and possibly Nc2–d4–b5 as an additional resource that will hit both the c6-
and d6-weak links in the black position really hard.
a) 12.Qd2 Qc7 13.Rfd1 Ne8 14.Nd4 Nxd4

15.Qxd4 (15.cxd4 Qxc4 16.Rac1 Qa6 17.Rc3 Bd7∞ and Black slowly unravels while keeping his extra pawn.)
15...e5! The only move that allows Black to complete his development and reach equality. 16.Qd5 Be6 17.Qxb7

17...Qxc4
[17...Qxb7? would be a serious error. 18.Bxb7 Rab8 19.Bd5 Bxd5 20.Rxd5 f6 (20...Rdc8 21.c5 Nc7 22.Rxd6 Nb5
375
23.Ra6 Nxa3 24.Rxa3 Rc7 25.Rd1 g6 26.Rd5 and White has a close to winning position, due to his two extra
pawns.) 21.Rad1 Rbc8 22.c5+–]
18.Qb4 Rac8 19.Qxc4 Bxc4 20.Rd2

20...a5! It is important to prevent White’s a3-bishop from getting to the b4-square, from where it would be able to
both protect the weak c3-pawn and reach the a5-square, harassing Black’s rooks on the back rank. 21.Rb1 Rb8!
22.Rdb2 Rxb2 23.Rxb2 Rc8!= and Black holds the balance as his c4-bishop is extremely strong.

b) 12.Nd4!? is another possibility. White intends to recapture on d4 with his c3-pawn without fearing Black taking
on c4 with his queen. The game might develop along the following lines: 12...Ne5 13.Qb3 a6!? Probably the
strongest reply, leading to very sharp play.
[13...Nfd7 14.Rfd1 Nc5 15.Bxc5 dxc5

376
16.Nf3! A superior idea. White makes sure that Black does not retain his active e5-knight. Now Black will find it
difficult to complete his queenside development without giving White some material concessions. (16.Nb5 Bd7
17.Nd6 Bc6 18.Bxc6 Rxd6 19.Rxd6 Qxd6 20.Bxb7 Rb8 21.Rb1= Qd2 22.Rb2 Qe1+ 23.Kg2 g6 and White cannot
make progress because of his pinned b7-bishop.) 16...Nc6 (16...Nxf3+ 17.Bxf3 Rf8! 18.Qa3!± Further tying up all
the black queenside pawns. White plans to double rooks on the b-file, hitting Black’s c5-pawn.) 17.Ne1!! White
transfers his knight to the ideal d3-square, while opening up his g2-bishop. 17...Ne5 18.Rxd8+ Qxd8 19.Nd3! Nd7
20.Qa3 Qc7 21.Rd1 a6 22.Qc1²

White maintains constant pressure and Black must be on guard at all times against the various white threats. The
position is close to balanced, but Black’s defense requires great accuracy. 22...h6 23.f4 Rb8 24.Qd2

377
White is ready to invade the eighth rank via 25.Nd3–e5 next move, however... 24...Kf8! The only move to hold the
balance. Black brings his king closer to the vital d8-square and threatens to free his queenside by playing 25...b5 on
the next move. (24...b5?? Not suspecting the trap. 25.Ne5!! White is winning as he gets through to Black’s weak
back rank, winning the game due to the pin on the c8-bishop.)]
14.f4 Neg4 15.Nc2 e5

16.f5 (16.h3 Nh6 17.g4 Nfxg4 18.hxg4 Nxg4 19.Qb6 Qh4 20.Rf3 Qh2+ 21.Kf1 Be6 and Black gets certain
counterplay.) 16...b5! The point. Black sacrifices material to get activity for his pieces. 17.h3 Nh6 18.g4 Nfxg4 with
counterplay.

12...Qc7=

378
At this point in the game I started to feel optimistic as I didn’t see how White could get through to my weak d6-
pawn, and started to guess my opponent’s possible plans.

13.Nd4!?

White must play energetically, or he risks getting into an inferior position. The text move allows White to open the
position for his pair of strong bishops, and we have seen in previous games how strong these guys work together!!
13.Ne3?! b6 14.Qd2 Bb7 15.Rfd1 Ne8 16.Bb4 Rac8³ White’s doubled c-pawns will be targeted for capture. In this
semi-closed position the knights are actually superior to the bishops.

13...Nxd4

Taking up the gauntlet and accepting White’s pawn sacrifice.

Anything else would make life much easier for White.

14.cxd4

14.Qxd4 would be a safer way to play for White, but he would be in no position to play for an advantage after the
following forced sequence: 14...e5

379
15.Qd3 (15.Qh4?! allows Black to reroute his f6-knight in an advantageous way: 15...Rb8 16.Bc1 Re8 17.Bg5
Nd7³.) 15...Be6 16.Rxb7 Qa5! A fighting move. The following alternative is inferior.
[16...Qxc4?! 17.Qxc4 Bxc4 18.Rc7! d5 (18...Rac8? would give White a serious advantage. 19.Bxd6! Bxe2 20.Rxc8
Rxc8 21.Re1 Bc4 22.Bxe5 Bxa2 23.Ra1 Bd5 24.Bd4 Bxg2 25.Kxg2 Nd5 26.Rxa7± with serious winning chances.)
19.Rxc4 dxc4 20.Bxa8 Rxa8 21.Rd1± and White’s combination of B+R , controlling the main file and the center,
gives him strong chances for a big advantage.]
17.Rb3 Rac8 18.Bb4 Qxa2 19.Ra3 Qxc4 20.Rxa7 Nd5!

The simplest. Black forces White to transpose into an opposite-colored bishops endgame, where a draw will be
highly likely. 21.Qxc4 Rxc4 22.Bxd5 Bxd5 23.Rd1 Bc6! 24.Re7 f6 25.Ra1 Re8!

380
26.Rxe8+ (26.Rc7 Bb5=) 26...Bxe8 27.Rd1 Rc6! 28.Rxd6 Rxd6 29.Bxd6 Bb5=

14...Qxc4

The pawn must be taken, otherwise Black will be stuck with the inferior structure and inferior pieces.

15.Qd2 Qa6

The black queen leaves of its own accord before being evicted by White’s Rc1.

16.Rb3 Rb8

Preparing to develop the c8-bishop. So far both sides have made natural moves and the game is balanced. Now
White has a wide choice for the development of his initiative. Choice number one would be to focus his pieces on
Black’s queenside, choice number two is to expand in the center and attempt to open central lines for his rooks, and
choice number three is to attempt to utilize his dark-squared bishop on the long diagonal, where it has no
381
counterpart, and go for the attack on the kingside. My opponent decides to go with the last and most dangerous
option, as the following lines will show.

17.e4

White retains the option of developing his f1-rook.

a) 17.Qb2 b6 18.Rc1 Setting up another nasty trick. 18...Ne8! The correct choice. (18...Bb7?

Natural — and a blunder!! 19.Bxd6! Rxd6 20.Ra3 The black queen is trapped, leading to a decisive material
advantage for White.) 19.Bb4 Bb7

20.Ra3 (20.Bxd6?? doesn’t work anymore, as Black’s d6-knight provides an exit for the black queen on b5.
20...Nxd6 21.Ra3 Qb5–+) 20...Qb5 21.Rxa7 Bxg2 22.Kxg2 Rdc8! It is important to contest the open c-file. With the
exchange of the light-squared bishops, White has no advantage at all, and the game is very likely to end in a draw as
both sides have no weaknesses and White’s slightly more active pieces have no way to profit from their activity.
Black will either chase away or exchange White’s active a7-rook, leading to complete equality.

382
23.Bc3 Qd5+ 24.f3 Ra8 25.e4 Qc6 26.Rxa8 (26.Bd2?! Rxa7 27.Rxc6 Rxc6³ In general, the exchange of queen for a
pair of rooks favors the side with the rooks.) 26...Qxa8!

Keeping the rook on c8 to face its counterpart. 27.Qxb6 Qxa2+ 28.Kh3 Qe2 29.Qb7 Attempting to force the black
rook from the back rank allows a draw by perpetual check: 29...Qxf3 30.Qxc8 Qh5+ 31.Kg2 Qe2+ 32.Kg1 Qe3+.

383
33.Kg2 (33.Kf1?! would be playing with fire: 33...Qxc1+ 34.Kf2 Qc2+ 35.Ke3 Qa4, leaving Black with a healthy
extra pawn. 36.Bb4 d5 37.exd5 exd5³) 33...Qe2+=
b) 17.Rc1 This looks quite logical, placing the rook on the open c-file. 17...Bd7 18.e4

18...Ba4 A dangerous move. (18...Rbc8! may objectively be stronger, as the other choice leads to a lot of adventures:
19.e5 Nd5 20.Bf1 Rxc1 21.Qxc1 Qa4 22.Rxb7 dxe5 23.dxe5 Bc6 24.Rb3 Nb6 25.Rb4 Qa5= and Black is ready to
fight for an advantage.) 19.Rbc3 (19.Bf1?! Nxe4! 20.Qe3 b5 and the tactics work in Black’s favor.) 19...h6! In view
of what happens in the game it would be quite prudent for Black to make this king luft, although now White can be
creative and try the following line. (19...b5?? illustrates well why Black needs the luft: 20.Rc6 Rb6 21.e5! and
White’s threats are decisive.

384
21...dxe5 22.dxe5! Nd5 23.Qg5 Re8 24.Bxd5 exd5 25.Qe7!! Ra8 26.e6, winning immediately.) 20.g4!?

Intending to make full use of the c3-rook on the third rank to aid the attack on the black king along the g-file. 20...b5
[20...e5!? Another adventurous idea, and perhaps my favorite move. Black creates a future spot for his f6-knight on
f4. 21.g5 Nh5 22.gxh6 Nf4

385
and Black is just in time, as 23.hxg7? (23.Bf3 g6„) is impossible due to 23...Qe2!, when the tactics work in Black’s
favor: 24.Qe1 Qg4 25.Rg3 Ne2+ 26.Kf1 Nxg3+ 27.hxg3 Rbc8 and suddenly Black starts a decisive counterattack.]
21.Rg3!? (21.Bb4 Rdc8 Starting mass exchanges that will guarantee Black safety from any possible white attack.
21.Rc6 Rb6 22.Rxb6 Qxb6 23.Bb4 Re8 24.e5 dxe5 25.dxe5 Nd5 26.Bxd5 exd5 27.Qxd5 Qd8! leads to a draw, since
28.Bd6? is a blunder: 28...Qg5!µ) 21...Rdc8 22.g5

22...hxg5
[22...Nh5 looks too dangerous for my taste, although Black should hold with precise defense. 23.gxh6 Nxg3 24.Qg5
(24.hxg3 gxh6! 25.Qxh6 b4! Allowing the transfer of the black queen to the center of the battle. 26.Rxc8+ Rxc8
27.Bc1 Qe2 and Black is fine since after 28.Bg5?! Rc2 he is just in time with his own threats to the white king.
29.Be3 Only Black can be better if he brings his queen to g4.) 24...Rxc1+ 25.Bxc1

386
25...Nh5! (25...g6? 26.hxg3+– Winning on the spot. 26...Qc8 27.h7+! Kxh7 28.Qh4+ Kg8 29.Bh6! Qd8 30.Bg5 and
the bishop reaches f6, ending the game.) 26.Qxh5 Qc8 27.Bg5 f6 28.Qg6 (28.hxg7? fxg5–+) 28...Qf8 29.Bh4 Rb7
30.e5 Qf7 31.Qg3 Rc7 32.exf6 Rc1+ 33.Bf1 Bc2 34.fxg7 Kh7 with an extremely unbalanced position.]
23.Qxg5 Ne8 This looks safe for Black.

24.Re1 (24.Rf1 b4 25.Bc1 Bb5 26.Re1 Qxa2 27.Rh3 Rxc1! The safest. Black eliminates White’s potentially
dangerous bishop and fears no white attack, since the black king can reach safety via the e7-square.) 24...Qa5 Black
brings his queen back to the defense with tempo.

17...Bd7

387
18.Re1?!

White declares his aggressive intentions. Now Black has to calculate a lot of lines, though fortunately most of his
moves are the only logical ones, while White keeps more options. 18.e5!? would switch the focus to the center,
which is not what my opponent wanted at this point. 18...dxe5 19.Be7 (19.dxe5? Ba4 20.Qg5 h6 21.Qh4 Bxb3
22.exf6 Qxa3–+) 19...e4!!

A very strong and unusual idea. White cannot really accept the exchange sacrifice as it will allow Black to take
control of the whole center and leave the white d4-pawn very weak. 20.Bxf6
[20.Bxd8 Rxd8 21.Rc1 (21.f3 Bb5 22.Rc1 Bc4 23.Rbc3 exf3 24.Bxf3 b5³) 21...Bc6 22.Bf1 Qa4 23.Rc4 Qa6 24.Rc5
Qa4 25.Rc4 Qa6=

388
and while the black queen looks in trouble, White has no way of capturing it. The position is in a state of dynamic
equilibrium.]
20...gxf6 21.Bxe4 Bc6 22.Bd3 Qa4 23.Qh6 f5

24.Bxf5! and the game is drawn by perpetual: 24...exf5 25.Qg5+ Kf8 26.Qh6+ Ke7 27.Qg5+ Kd7 28.Qxf5+ Kc7
29.Qf4+ Rd6 (29...Kc8? 30.Rc1+–) 30.Qxf7+ Rd7 31.Qf4+=.

18...Ba4

389
The bishop has to go here to provide the black rooks with access to the c2-square. It reminded me of two of my
games vs. Shirov which concluded our short theoretical dispute in a Grünfeld line, played in 1992 and 1993. Feel
free to have a look at them and make your own analysis and conclusions. Not every game can fit into this book!
18...Bc6 is possible, but too passive. White gets an easy game after 19.Bb4!? (or 19.d5 exd5 20.exd5 Bd7 21.Bb2
and White’s monster b2-bishop guarantees him compensation for the sacrificed pawn).

19.Rf3?!

Again White demonstrates his ambitions, but he underestimates Black’s defensive resources, otherwise he would
have settled for the more positional 19. Rc3, which would have kept the game balanced. For example:
19.Rc3 Rbc8 20.Rec1 Rxc3 21.Qxc3 Bc6 22.Bf1

22...Qa4 (22...Nxe4?? 23.Bxa6 Nxc3 24.Rxc3+–; 22...Bb5 23.Bg2„) 23.f3 h6 24.Qb4 Qxb4 25.Bxb4 and White’s
powerful pair of bishops and strong pawn center make the game very close to equal.

19...Rbc8
390
A very reasonable move, not fearing possible exchange sacrifices on f6 because White’s a3-bishop is so far away.
To keep the balance, though, White should have considered the sacrifice.

20.Bf1?

This certainly took me by surprise. Apparently my opponent was worried about a possible ... Rc2 invasion and
therefore decided to get Black’s a4-bishop away from protecting the c2-square. However the b5-square turns out to
be vitally important and White has lost a crucial tempo.
a) 20.d5?! Rc2 21.Qg5 h6 22.Qh4 Re2!³ Black has the advantage as White needs to spend more tempi protecting his
weak back rank. 23.Rxe2 Qxe2 24.h3 Qd1+! (24...Qxe4?! Not the strongest. White now opens the position. 25.Rf4
Qb1+ 26.Kh2 g5 27.Qxh6 gxf4 28.Qxf6 fxg3+ 29.fxg3 Re8 30.Bxd6 Qf5 31.Qh4 with compensation) 25.Kh2 Qd4!

Very strong centralization which also takes away the long diagonal from White’s a3-bishop. 26.dxe6 fxe6 27.e5!
(27.Re3 Bc6) 27...dxe5 28.Be7 Qxh4 29.gxh4 Re8 30.Bxf6 gxf6 31.Rxf6 Kg7 32.Rf3 b5 33.Rc3 Kf6µ
b) 20.Rxf6!? gxf6 21.e5!!
391
During the game I felt that White should equalize here because of the dark squares, but it was not easy to calculate
everything. (21.Qh6 Qa5! and White has no time to capture on f6. 22.Rf1 Qg5–+) 21...dxe5 22.Qh6 (22.Be7? Rxd4
23.Qh6 Rg4!–+ This rook lift is decisive. Black avoids the mating net and starts to strike at White’s own weak back
rank.) 22...Rxd4

23.Be4!! The only move to keep the balance. (23.Qxf6 Rg4 24.Qxe5 Qb5! Preventing White from setting up a
battery along the long diagonal. White’s attack is now deflected and Black has a material advantage which, however,
will not be easy to convert due to his weak king and slight lack of coordination between his rooks. 25.Qe3 Qb6! The
main idea is to centralize the queen, taking over the long diagonal. 26.Qe2 Rg6 27.Qe5 Qc7 28.Qa1 Bc6–+; 23.h3?
is too slow. 23...Qa5 24.Re2 Qd8 and Black has successfully defended his f6-pawn.) 23...Rxe4 24.Rxe4 Bc2

392
25.Be7!! Again the only move. (25.Rg4+? Bg6 26.Be7 Qc6! 27.Bxf6 Qc1+ 28.Qxc1 Rxc1+ 29.Kg2 e4 30.Rg5 Rc2
31.Kg1 b6 32.g4 h6 33.Rb5 Rxa2 34.Rb3 Kh7 35.Rc3 e3 36.fxe3 Be4–+ Black gets his king out of the box,
obtaining a winning advantage.) 25...Bxe4 26.Bxf6 Rc1+ 27.Qxc1 Qd6

Black now has to be careful, but both kings are weak, which keeps the game balanced. 28.Qc8+ Qf8 29.Qc7 Bc6
30.Qxe5 h6 31.Qf4 Kh7= and Black successfully keeps his king out of danger.

20...Bb5 21.Bh3

393
The point of White’s idea: he wants to break open the position for his bishops by combining d4−d5 and a Rxf6
sacrifice. I felt that I had no choice but get on with my counterplay on the queenside.

21...Qa4 22.d5?

The losing move. It starts a forcing sequence which culminates in a win for Black, but the calculation had to be
precise as the black king gets into a mating net. In hindsight, White should have gone over to defensive mode by
playing 22.Rb3, but it’s hard to believe that the position, with a space advantage and strong pair of bishops for White
(especially the dark-squared bishop) and the black king lacking any piece defense except the f6-knight, could be
better for Black.
a) 22.Rc3? d5–+
b) 22.Rb3 b6 23.f3 Rc7 24.Bb4 Rdc8 25.Ra3 Rc2 26.Rxa4 Rxd2 27.Rxa7 Rxd4 28.a3³ White has good chances to
draw this endgame.

22...Rc2!

I spent a long time calculating the final line and verifying it several times, as it required a temporary queen sacrifice,
followed by quiet moves.

23.Qe3

23.Qg5?! h6 24.Qh4 Re2 25.Rb1 Rxe4–+ makes things too simple for Black.

23...exd5 24.e5 d4!!

394
Black needs to deflect White’s queen from the e2-square.

25.Qg5 Re2

Unfortunately the alternative, 25...h6, is perhaps even simpler. It wouldn’t have required any sacrifices, but then
chess fans would have been robbed of the beauty that unfolds in a few moves: 26.Qh4 Re2 e.g. 27.Rf1 Re4 28.g4
Bxf1 29.exf6 Qd1 30.Bxf1 Qxf3 31.h3 Re1–+.

26.exf6

White only needs two more moves to mate the black king, and in order to achieve his goal, he sacrifices almost his
entire army.

26...Rxe1+ 27.Bf1 Rxf1+ 28.Kg2 Rg1+! 29.Kh3

29.Kxg1? Qd1+ 30.Kg2 Qf1#

29...Bd7+ 30.Kh4 g6 31.Qh6

This is the position that White had been aiming for. It seems that mate is unavoidable and the discovered check is
useless, however...

31...d3+! 32.Rf4

395
32...Qxf4+!!

The lightning strike that reveals the true state of the position. I recall that my opponent was visibly perturbed that I
had found this resource, and the rest of the moves were blitzed out quickly with some of the other players gathering
round to watch the game.

33.Qxf4

Forced.
33.gxf4 Rg4+ 34.Kh3 Rg5+ 35.Kh4 Rh5+ recaptures the queen.

33...Rh1 34.g4

34...h6!!
396
This little pawn move, which I had to foresee before making my 22nd move twelve moves ago, decides the game! I
was again astonished by fate, or whatever higher forces exist in this world, showing me the correct path to create
such a tactical masterpiece. Sometimes it makes me wonder... Back to the game, the threat of ...g6–g5 check doesn’t
give White too many options and he cannot capture on h6 as ...Rxh2 will score a direct hit.

35.Kh3 g5

In addition to the white king standing in a mating net, White has to pay attention to the far-off passed pawn on d3
that is close to becoming a queen. White’s f6-pawn is actually providing excellent protection for the black king.

36.Qd4 d2!

The final touch. Black opens the diagonal for the bishop to create a lethal mating threat.

37.Qxd2 Rg1 38.f3 Bb5 0–1

This game made the score 2–0 against an opponent whom I considered one of my most dangerous adversaries along
with Mr. Anand and Mr. Gelfand. Having scored the wins in such spectacular fashion showed me that I was in very
good tactical form despite my lacklustre opening preparation. I felt very confident of clinching the match — and did!
The following three games were drawn, although my opponent missed some chances, and after again suffering in the
sixth game, I managed to snatch a win from a much worse position, winning the match 4½-1½.

397
 GAME 15
Gata Kamsky
Viswanathan Anand
Sanghi Nagar 1994
A lot could be said about this match, in fact there would be enough material to write a whole book about my FIDE
and PCA Candidates matches, but the purpose of this book is to acquaint the reader with the possibilities of chess,
the richness of its ideas and its sometimes impossible beauty, especially in the endgame. To survive that far, a
modern top GM has to spend many hours studying opening theory and be on top of his game in the middlegame, and
the first two games of the match were absolutely thrilling affairs. I am especially fond of the first game, which
featured a spectacular middlegame run by my king to join in a direct attack against White’s king while most of the
pieces were still on the board. It was very tempting to include it amongst the games here, but I felt that would be
asking too much. The interested reader can find it in a chess database.

1.e4

At this point in the match I was trailing by two points with three games left. My team, which consisted of my father,
GM Dzindzihashvili and a local Indian cook provided by the organizers, was not in great spirits. Personally, I was
not overly concerned about the result, but was more worried about a sudden argument between my father and GM
Shabalov (who was part of our preparation team, and who spent a large part of 1994 in Brooklyn helping me train
for the Candidates) which resulted in a physical altercation between them and Alexander Shabalov taking the next
flight home to the USA. In other times, my negative score in the match would have pressurised my father into doing
something, which in turn would have usually resulted in great psychological pressure on me to do something, but the
distraction of the incident with GM Shabalov was on my father’s mind, since he suddenly felt defensive and a great
need to protect himself from a possible legal case. Dzindzi and I felt great sympathy for Alexander, but of course we
could do nothing against the unpredictable character that was my father. One typical thing he would do was, while
teaching the Indian cook how to prepare a Russian meal, run upstairs to the second floor, where Dzindzi and I were
trying to come up with ideas to keep my chances alive, and suggest about ten moves in as many seconds. Because
my father’s chess ability was actually around first category, it was impossible to bluff him with obvious lines. Of the
ten moves he suggested, one or two would be quite relevant to the position, and so when we analyzed those in
addition to the moves we were looking at, and were unable to reach a clear verdict, my father would triumphantly
declare, “I told you about this move!”. This would go on for at least a couple of hours every day. Fortunately,
because there were so many people around, including local Indian guards and constant press attention, my father was
quite busy and Dzindzi and I had enough time to prepare something.
Some of the lines below were taken from analysis made back in 1994. It is quite something for me to return and see
how I viewed chess at that time.

1...c5 2.Nf3

The previous two games in the match saw me play 1.d4 and score only ½/2, where the ½ really should have been
zero, but by some miracle I was able to save the game in the ending. Given the score in the match, a change of
opening was logical. It was decided that whatever my opponent might play, I would pick a fight on the board.

2...Nc6 3.d4 cxd4 4.Nxd4 g6

398
One of the lines we expected in our preparation, however given how intuitive these positions can be, and how we
were assisted in those times by such mediocre chess programs, sifting through games in the database was a tiring
and messy job that often led to conflicting results. We decided not to go into Maroczy Bind positions arising after
5.c4, even though I often played them and with very good results, but to accept the invitation to the Dragon and
follow the main lines.

5.Nc3 Bg7 6.Be3 Nf6 7.Bc4 Qa5

Black attempts to dissuade White from castling long and aims to reach a dynamic position with risks and
possibilities to play for three results. Leading by two points in the match, of course Black wants a quiet positional
game, but developing the queen too early in the opening is never the best thing to do and Black will have to pay a
price for denying White’s monarch his queenside castling.

8.0-0 0-0 9.Bb3 d6 10.h3

An important little pawn move. The g4-square is absolutely critical for White’s expansion in the center. All the
399
white pieces will be developed logically and Black will be faced with the task of how to develop his rooks and the
problem of his wayward queen.

10...Bd7 11.f4

11...Rac8

Quite logically Black finishes his development and waits for White to declare his plans.
a) 11...Qh5 would be an attempt to reach safety in the endgame, but White can decline the exchange. 12.Qd3 The
main move. White connects his rooks and improves his queen. Black’s queen on h5 doesn’t have many useful
functions, except to support the standard ...b7–b5. (12.Nf3!? b5 13.a3 a5 14.Qd3 a4 15.Bd5 e6 16.Bxc6 Bxc6
17.Nd4 Bb7 18.f5

Slightly better for White (according to my old notes): Garcia, Gi.Zamora,J New York 1994, 1–0 (25). I believe that
still holds true today. Black is in no danger of being checkmated, and his two bishops provide enough counterplay to
400
hope for equality. 12.Qxh5 Nxh5 13.Rad1 Nxd4 14.Bxd4 Bc6 and it will be extremely difficult for White to win this
game.) 12...b5 13.a3!² I’m not sure that modern teenagers, or even those born in the 1990s, have any idea of what it
was like to study from the multi-volume series of books on the opening like the Encyclopedia of Chess Openings,
a.k.a. ECO. Such books contained knowledge, games, analysis by leading grandmasters and possible mistakes in
lines and evaluations. Everyone who hoped to become a GM had to study these books, and study them deeply,
hoping to discover these mistakes so he could use them to his advantage in the next important tournament game. In
addition, many top GMs required their seconds to bring such books to top-level tournaments to study or refresh their
memory. Of course every player, in addition to studying such books, had to keep scores of notes in jealously
guarded notebooks with careful hand-written lines and comments, which added tremendously to the weight of one’s
baggage!!

13...a6 The verdict here is that White is indeed slightly better due to his superior development, better piece
coordination and potential breakthroughs. I’ll leave it here so the interested reader can perhaps learn the intricacies
of the opening for himself.
b) 11...Nxd4!? Another standard idea for Black is to exchange his c6-knight and put the light-squared bishop on the
c6-square, allowing the transfer of the f6-knight to the ideal c5-square. 12.Bxd4 Bc6 13.Nd5!

401
This seems to be main move that leaves White with some advantage. The idea is for White to press ahead on the
kingside with the f4–f5 pawn push. Since Black is forced sooner or later to exchange White’s strong centralized d5-
knight, White will recapture on d5 with his e4-pawn, creating a superior pawn structure and leaving Black with a
weak, backward, e7-pawn. In return, however, Black will get a formidable outpost on the e5-square for his minor
pieces. The game will be quite double-edged — a perfect opening choice for the player who finds himself in a must-
win situation with either color!!

12.Qf3

12.Nf3!?

Quite an interesting idea, exploiting the c6-knight’s lack of useful squares. In fact one of the standard squares for
this knight has been taken away by its own queen. White also keeps the overall exchanges to a minimum and plans
to play a slow positional game. However my opponent’s great understanding of the game, combined with his ability
to calculate at lightning speed, would have made me feel uncomfortable playing such a position against him,
especially given the score in the match. Instead, I decided to focus on one target alone, the black king, and go after
402
it. 12...Rcd8 (12...Qh5 13.Qd3 b5 14.a3 b4 15.axb4 Nxb4 16.Qd2 Nc6 17.Bc4!? Na5 18.Be2 Nc4 19.Bxc4 Rxc4
20.Rxa7± Garcia,G-Oblitas,C Capablanca Memorial 1992, 1–0 (38).) 13.Qe1 Bc8 14.Rd1 e6 15.Qh4 Qh5 16.g4
Qxh4 17.Nxh4².

12...Qh5?!

At this point I was out of my book, as I didn’t know this particular line well. I did feel that White’s moves were
quite logical and natural and therefore easy to guess and make. Black, on the other hand, has a hard choice. One is to
exchange the knights on d4 and simplify the position, but then White’s bishop on d4 becomes strong. The other
choice is to go after white’s b3-bishop with ...Na5. Given the tension of the match, and probably the fact that Black
wasn’t quite ready for this particular line in this game, and so feared my potential preparation, he decides to ask me
if I’m ready to open my king up with a move like g2−g4, or admit to losing a tempo by moving my queen from the
ideal f3-square.

13.Nxc6!N

I came up with this move during the game. Apparently it was a TN — a theoretical novelty — at the time. It is quite
incredible to me that for so many years a lot of people would doubt my ability to find moves like this, on my own,
during a game. Most would even claim that I had prepared them at home, probably due to the stereotypical image of
me as a nerd in big glasses who showed very little emotion during a game. Who knew what crazy ideas I had when
playing a game, even to this day!
a) 13.Qf2!? b5!

403
Black manages to play this standard Dragon queenside pawn push which signals the beginning of strong
counterplay. 14.Ndxb5 Nxe4 15.Nxe4 Qxb5= Given how strong Black’s central pawn formation is, it seems very
likely that he has equalized.
b) 13.Rad1 Na5 14.g4 Nxg4∞
c) 13.g4? loses on the spot: 13...Bxg4 14.hxg4 Nxg4 15.Rf2 Nxd4–+.

13...Bxc6

a) 13...bxc6?! is a move that would never enter the mind of a top player for several reasons. One, Black’s previous
move, ....Rac8, becomes a loss of tempo; two, it would change the whole character of the position, from the simple
...b7–b5–b4 push against White’s c3-knight and placing the d7-bishop on the c6-square followed by ...Nf6–d7, to a
more complex structure where Black would have to switch his rooks to the semi-open b-file, and go for ...a7–a5–a4,
... Bd7–e6 and ...Nf6–d7. Given the awkward positioning of Black’s h5-queen this would be more complicated, and
White’s game would become much easier.
b) 13...Rxc6?!

404
Apparently not something that Black should do. 14.Bxa7! White simply grabs a pawn! (14.g4?! But not this.
14...Bxg4! 15.hxg4 Nxg4 16.Qg3 Bxc3 17.bxc3 Rxc3 18.Rf3 Nh2! 19.Rf2 (19.Qxh2?? Qxf3 20.Bf2 Qg4+ 21.Kf1
Rf3 is just losing for White.) 19...Ng4 20.Rf3 Nh2 with a draw.) 14...Qxf3 15.Rxf3 b6 16.e5 Ra8 17.exf6 Bxf6
18.Bd5 Rxa7 19.Bxc6 Bxc6 20.Rd3±

and although Black has an absolutely wonderful bishop pair, the lack of material or concrete compensation for the
missing exchange should eventually spell trouble for him.

14.g4!²

14.Bxa7 Qxf3 15.gxf3 Nh5 with compensation.

14...Qa5 15.Rad1

405
I remember that I was pretty happy with the outcome of the opening here. By wasting a lot of time moving his queen
Black has actually helped me to advance my pawn chain on the kingside. I felt very optimistic about my chances.

15...b5

As mentioned before, Black has to play this move and the following standard minority attack to create pressure on
White’s center and the queenside. Quite interestingly, the computer’s first impression after this move is that Black
has a slight advantage! However the game is far too complicated to make definite evaluations at this point, and at the
time I was faced with the choice of which pawn to move first, f4 or g4.
15...Nd7!? The old computer engine that I used at the time, the MG2 (I can’t remember its full name!) preferred this
move. It seems to fall within the top three lines of modern engines as well. 16.f5!ƒ Both sides are going full steam
ahead with their plans, White with his direct attack on the king and Black with the strategic placement of his pieces
on their best squares, along with paying attention to the structure. These plans are going to conflict at one point and
result in one mess of a tactical battle.

16.g5

The correct choice.

a) 16.f5? b4

406
Black actually finishes up better for a number of reasons. 17.Nd5
[17.g5? Forcing matters is usually a bad idea. Here is just another example. 17...bxc3 18.gxf6 Bxf6 19.fxg6 hxg6

20.Rxd6!? White’s last chance to mix things up before he finds himself in a strategically lost position. (20.Bh6?!
cxb2 21.Bxf8 Kxf8 22.Rxd6 Qg5+!µ Black completely owns the board.) 20...Qe5!!–+ Black wins thanks to White’s
ruined position on the queenside, in the center and on the kingside. (20...exd6?? 21.Qxf6+– Without Black’s dark-
squared defensive bishop, White’s attack becomes unstoppable. 21...Qe5 22.Bxf7+ Rxf7 ™ 23.Qxf7+ Kh8 24.Qf2
Kh7 25.Bd4 Qg5+ 26.Kh2 Rc7 27.Bxc3+–)]
17...Nxd5

407
Modern computer programs are not afraid to take the b2-pawn in this position. (My old notes gave 17...Bxd5
18.exd5 Nd7 as good for Black, but the position after 19.Bg5 Rce8 20.Rb1∞ is actually a mess.) 18.exd5 Bb5 19.Rf2
Bxb2 20.Qe4 Bf6 21.Bd4

21...Bxd4 (21...Bd7 This move is quite possible as well, since White’s attack is completely stopped, and his central
d5-pawn is blocking his own light-squared bishop, making it quite useless. White must start looking for ways to
equalize.) 22.Qxd4 Rc3 23.Kh2 f6 24.Re1 Rf7 25.Re4 and while Black’s game is clearly superior, White can hope
for a draw.
b) 16.a3 b4 17.axb4 Qxb4 and Black has succeeded in opening queenside files for his rooks and creating pressure on
White’s Achilles heel, the e4-pawn.

16...Nd7

16...Nh5 was the other possibility. The knight will be stuck there for an unknown period of time, which is a
significant psychological factor, and while the d7-knight has obvious functions, the h5-knight is solid but has no
future. However sometimes you can afford one ‘inactive’ piece if this piece can annoy the opponent in making his
408
plan work. On h5, the knight definitely limits White’s power on the kingside, and also controls some squares that
could be useful to White.
After analysis, my preference goes to 16...Nh5!.

17.Nd5 White’s best reply.

a) 17.f5 Bxc3! It could be quite difficult to give up such a defender, even if it was objectively necessary. Black has
no bad pieces, no obvious weak spots or pawns in his position, and therefore doesn’t deserve to lose. In fact, he is
well placed to start attacking White’s weaknesses created by his kingside pawn advance. 18.bxc3 Qxc3 19.Rf2

19...Ng7!! A brilliant move. Other options were unclear. (19...Qe5 20.fxg6 hxg6 21.Bxf7+ Kh7 22.Bd4 Qxe4
23.Qxe4 Bxe4 24.Re1 d5 25.c3

409
and now a sample continuation would be 25...a5 26.a3 Ng3 27.Bxd5!? Rxf2 28.Kxf2 Nh1+ 29.Rxh1 Rf8+ 30.Kg3
Bxd5 31.Re1 Rf3+ 32.Kg4 e6 Attack has met its equal in defense, and the result is a draw; 19...a5 20.fxg6 hxg6 and
now White has nothing better than to go for a forced temporary queen sacrifice: 21.Bxf7+ Kh7

22.Qxh5+! gxh5 23.g6+ Kg7 24.Bd4+ Qxd4 25.Rxd4 Be8 resulting in a position that is better for Black because of
White’s ruined pawns. Luckily for White, a draw is close: 26.e5! dxe5 27.Rd5 Bxf7 28.gxf7 Rxf7 29.Rg2+ Kf6
30.Rxb5 a4 31.Rf2+ Kg6 32.Rg2+ Kf6 33.Rf2+=) 20.fxg6 hxg6 21.Bxf7+ Kh7µ

410
and Black successfully defends against White’s attack. White has to watch out for ...Nf5.
b) 17.Ne2

17...b4! Black simply tries to trade White’s best attacking piece, the ‘Spanish’ b3-bishop, putting a stop to White’s
ambitions on the kingside. (17...Bxb2 18.f5 b4 19.fxg6 hxg6 20.Bxf7+ Kh7

411
21.Rd5! Bxd5 22.exd5 Be5 23.Qe4 Rxf7 24.Rxf7+ Ng7 25.Nf4 Bxf4 26.Bxf4 Qxa2 and Black is just in time to
counter White’s drawing threats. 27.Rxg7+ Kxg7 28.Qxe7+ Kh8 29.Qf6+ Kh7 30.Qf7+ Kh8=) 18.Nd4 Ba4 19.Rf2
Bxb3 20.axb3 Now the game can go several ways, but in every case Black has sufficient counterplay, for example
20...e5, the most logical and standard Sicilian break in the center.

21.Nf5 (21.fxe5 dxe5 22.Nf5 gxf5 23.exf5 Nf4 24.f6 Bh8 Black’s extra piece is temporarily locked out of the game,
but eventually it will have its say. White doesn’t have sufficient attacking resources.) 21...exf4 (21...Nxf4?
22.Ne7++–; 21...gxf5 22.Qxh5∞) 22.Ne7+ Kh8 23.Bxf4 Nxf4 24.Nxc8 Qxg5+ 25.Kf1 Qb5+ 26.Ke1 Rxc8 27.Qxf4
Re8!

412
and it is White who has to be careful not to slip into an inferior position.
17...Bxd5 18.Rxd5 Bxb2 19.a4 a6 20.axb5 axb5 The advance of White’s kingside pawns has been stopped, but he
has managed to keep his strong light-squared b3-bishop alive, which gives him hope of some sort of compensation
for the missing pawn. However Black’s advantage is rather obvious if White plays passively.

21.Rb1!! The only move to keep the balance. (21.f5 Be5µ; 21.e5? dxe5 22.f5 Nf4µ) 21...Rc3! (21...Bc3 22.Rxb5=)
22.f5! Just in time, as Black’s b2-bishop is temporarily cut off from reaching its ideal central outpost on the e5-
square. 22...Ng7 23.f6 exf6 24.gxf6 Ne6 25.Kg2 Qa3 26.Rbd1∞

413
and White’s pair of bishops keeps his chances alive for all three results in this complicated middlegame.

17.f5

There is no other move. It’s all or nothing for White now; he cannot go anywhere but forward. I vaguely recollect
that at this point, my opponent went into a deep think. There are quite a few interesting choices for Black here.

17...Bxc3!?

A brave and also practical decision by Black. White will be unable to use the weak dark squares around Black’s
king, at least for now, but if the attack doesn’t happen, White will be stuck with a terrible pawn structure.
a) 17...Ne5? would be an absolutely terrible beginner’s mistake. While knights are best positioned in the center, this
position is all about the bishops!! 18.Qf4! After the coming f5–f6, locking out Black ‘s bishop, White will continue
his attack with advantage. 17...b4 is another choice, but it leaves White with strong pieces and a continuing attack.
18.Nd5 Bxd5 19.Rxd5 Qc7 20.h4!‚ White throws everything including the kitchen sink into the attack. All he has to
do is to open up his b3-bishop...
414
b) 17...gxf5!? An extremely interesting fighting idea which is also the computer’s first choice. Black radically
prevents the f5–f6 push at the cost of weakening his king’s defensive pawn shield, but disrupts White’s pawn chain!
18.Qh5™

The only possible continuation for White, but also the best. 18...Bxc3! (18...Ne5? 19.exf5 is just lost for Black.) It is
imperative for Black to exchange this knight. He cannot afford to give up his light-squared bishop, but at the same
time must prevent White’s Nc3–d5 as well as Rf1–f4–h4 and g5–g6 break. It’s one thing to say this from behind a
soothing computer program that finds the best defense, unfazed by White’s attack, but an entirely different story
with this position on the board in a really important game, as Black gives up the only defender of his king!! Apart
from 19.Rxf5!, when Black is again left with several choices, but only one correct way to to defend, White has two
options:
b1) 19.g6?! This was my original intention during the game, however it doesn’t work. 19...Nf6 20.Qh6

20...Bxb2! An incredible position where White’s entire army is participating in the attack on the black king, yet it is
only good enough for equality.

415
[20...Bxe4? 21.bxc3! (21.Kf2 f4–+) 21...d5! (21...Qxc3 22.Kf2 f4 23.gxf7+ Kh8 24.Qxf4 Rc4!? 25.Bxc4 Qxc4µ)
22.gxh7+ 22...Nxh7 (22...Kh8?? 23.Bd4 and White’s d4-bishop is an unstoppable monster. 23...Qa6 24.Bxd5 Bxd5
25.Rxf5 Be4 26.Rxf6+–) 23.Bxd5 (23.Kf2 Kh8 24.Rg1 Rg8–+) 23...Bxd5 24.Rxd5 (24.Rxf5?? Rc6–+) 24...f6!
Black still has to be precise to avoid falling into a worse endgame.

25.Kf2 Kh8 26.Rg1 Rg8 27.Rd7! Qxc3 28.Rxe7 Qxc2+ 29.Kf3 f4 30.Rgg7 Qd1+ 31.Kxf4 Rc4+ 32.Kg3 Qd6+
33.Kg2 Rc2+ 34.Bf2 Qd5+ 35.Kg1 Qd1+ 36.Kg2 Qd5+= with a perpetual.; 20...Qb4? 21.Rf4!!± The threat of
transferring this rook to the h-file is too great to ignore. Black is in great danger.]
21.Bg5! and now Black can go for a draw with 21...hxg6 or try for more.
[21.Rxf5? hxg6 22.Qxg6+ Kh8 23.Rh5+ Nxh5 24.Qxh5+ Kg8 25.Kf2 Bg7!

26.Qg4! A tricky attempt, however Black is winning. (26.Rg1 Qc3–+) 26...Rc7!–+ (26...Qc3? allows 27.Bd4)]
21...hxg6 (21...Qb6+!? 22.Be3 Qc7 23.Bg5 d5

416
The point. Black ‘s queen gets access to the g3-square. 24.Bxf6 Qg3+ 25.Kh1 Qxg6 26.Qxg6+ hxg6 27.Bxb2 dxe4
28.Rf4! Rfd8 29.Rh4 Rxd1+ 30.Kh2 e5 31.Bxe5 Kf8 32.Rh8+ Ke7 33.Rxc8 Bd7∞ With an approximate material
and positional balance as White’s bishop pair is compensated for by Black’s strong central connected passed pawns.)
22.Qxg6+ Kh8

23.Bxf6+ Bxf6 24.Qh5+ Kg8 25.Rxf5 Qb6+ 26.Kh2 Be5+ 27.Rxe5 Qf2+ 28.Kh1 dxe5= with a draw by perpetual
Qh6–g6–h6.
b2) 19.bxc3? Qxc3 20.Rd3 Qe5 21.Rf4 Bxe4 22.Rh4 Qg3+ 23.Kf1 f4–+
19...e6! Scary, but sufficient for equality.
[19...Ne5? loses at once:

417
20.Rxe5 Bxe5 21.g6 hxg6 22.Qxg6+ Kh8 23.Qh6+ Kg8 24.Kf2+–.; 19...Bg7? Trying to save the piece just doesn’t
work. 20.g6 hxg6 21.Qxg6 e6 (21...d5 22.Qg4 and Black has no defense to the coming Bh6: 22...Qc7 23.Bh6 Qb6+
24.Kh1+–) 22.Qg3! Nf6 (22...Qb6 23.Bxb6 exf5 24.Bd4 Ne5 25.exf5+–) 23.Rg5 Qd8 24.Rxg7+ Kh8 25.Rxd6 Nxe4
(25...Qe7 26.Qh4++–) 26.Bd4!!

The final touch; all of White’s pieces have joined the attack. Queens don’t matter; mating the king is the thing!
26...Nxg3 27.Rxf7+ Kg8 28.Rg7+ Kh8 29.Re7+ Kg8 30.Bxe6+ Rf7 31.Rxd8+ Rxd8 32.Bxf7+ Kf8 33.Bc5 Rd1+
34.Re1+ Kxf7 35.Rxd1+–.]
20.Bxe6!

418
(20.Rxf7? Rxf7 21.Bxe6 Rcf8 22.Rf1 Ne5–+) 20...Ne5
[20...fxe6?! 21.g6 h6 (21...hxg6 22.Qxg6+ Bg7 23.Bh6 Qb6+ 24.Kf1!!+– A suggestion by my trainer, GM Roman
Dzindzi, who was present at the match.) 22.Qxh6 Rxf5 23.Qh7+ Kf8 24.exf5 Bg7

25.fxe6 (25.Bh6? Qb6+ 26.Kf1 Bg2+! turns the tables on White. Black ‘s attack is decisive. 25.Rf1 Nf6 26.fxe6
Qe1!! and Black simply wins.) 25...Ne5 26.e7+ Kxe7 27.Qxg7+ Ke6

419
Black has miraculously survived the attack with his extra piece. White has to come up with counterplay to
compensate and can do so after 28.Qh6! (28.Rf1?! Bf3 29.Qh6 Kd7–+) 28...Rg8 29.g7+ Kf7 30.Qf4+ Ke6 (
30...Ke8? 31.Qf6±) 31.Qh6+ Kf7=]
21.Bxf7+! It is White’s turn to be accurate and steer the game towards the draw. (21.Rxe5? fxe6 22.g6 Rc7!–+)
21...Rxf7 22.Rxf7 Nxf7 23.g6 hxg6 24.Qxg6+

24...Kf8 (24...Bg7 25.Bd4 Ne5 26.Qe6+ Kh7 27.Qf5+ Ng6 28.Bxg7 Kxg7 29.Qxc8 Qb6+ 30.Kf1 b4 31.Qg4 Qe3
32.Rxd6 Bxe4 33.h4=; 24...Kh8? 25.Qh5+ Kg8 26.Qg4++–) 25.Rf1 Be8 26.Qxd6+ Kg8 27.Qg6+ Bg7!? (27...Kf8=)
28.Bd4 Ne5 29.Qe6+ Kh7 30.Qxc8 Nf3+ 31.Rxf3 Qe1+ 32.Kh2 Bxd4 33.Qc7+ Bg7 34.Rg3 Bg6 35.Qf4=

420
After all the tactical complications both kings are naked and prone to perpetual check, while there’s an approximate
material balance on the board.

18.bxc3

18...Ne5

18...Qxc3?? 19.fxg6 hxg6 20.Bxf7++– loses immediately for Black.

19.Qf4 Nc4

Black follows through with his plan to close out White’s b3-bishop. The following lines illustrate how carefully
Black has to play to avoid an early mate.
19...Qxc3?! 20.Kh2 (20.Bd4? Qxh3µ) 20...Nc4 21.Bd4 Qa5 22.Qh4+– with the idea Qh6; White gets to the black
king.
421
20.Bd4 e5?!

While Black doesn’t lose after this move, he allows White to transpose into a superior endgame where he will be left
with no counterplay.
20...Qc7!∞ Black prepares ...e7–e5 on the next move and White has to respond. 21.Qh4 e5 22.fxe6 fxe6 23.Bxc4
bxc4 24.Bf6 d5

25.e5 (25.exd5 Bxd5= White cannot make progress.) 25...Rb8 Black has counterplay, so isn’t worse, e.g. 26.Qg4
Rfe8 27.h4 Rb2 28.Rd2 Ba4 29.Rff2 Rb1+ 30.Kh2 Qf7 31.Rg2 Bc6 32.Qd4 Re1 33.Rde2 Rxe2 34.Rxe2 Qc7
35.Qc5 Qb6 36.Qd6 Kf7! and White cannot make progress.

21.fxe6 fxe6

22.Bf6?!

422
Returning the favor.
22.Bxc4! is a better move. 22...bxc4 ™ (22...Rxf4 23.Bxe6++–) 23.Qxd6 Qxg5+ 24.Kh2

24...Rfe8! The only defense. Despite the gigantic holes on the dark squares around the black king, White has no
winning blow and must settle for a slightly better endgame.
[24...e5? 25.Qe6+ Kg7 26.Bxe5+ Kh6 27.Bf6!ƒ Black is forced to give up the exchange, leading to an almost
decisive advantage for White. 27...Qh5 (27...Rxf6 28.Qxf6 Qxf6 29.Rxf6 Bxe4 30.Rf7 followed by Rd4 should be
an easy win for White.) 28.Rd2 Rcd8 29.Rg2 Bd7 30.Qe7, winning material with a continuing attack.]
25.Rf8+!?

A unique position where White doesn’t have a decisive knock-out blow. This move leads to a slightly better
endgame. 25...Rxf8 26.Qxe6+ Rf7 27.Qxc8+ Rf8 28.Qe6+ Rf7 29.Rf1 Qe7™ 30.Qxc4

423
30...Qd6+! (30...Qc7+ 31.e5 Be8±) 31.e5 Qd5 32.Qxd5 Bxd5 33.Rb1 Bxa2 34.Ra1 Bd5 35.Ra6 Re7 36.Kg3 Kf7
37.h4 Rc7 38.Kf4 Bc4 39.Rd6 Be6 40.Kg5 a5 41.Ra6

and White can still torture Black for a long time to come. Current World Champion Magnus Carlsen is an obvious
example of one adept in such situations, a genius in technical positions who can conjure up wins from nothing, and
here White has a whole extra pawn!!

22...Qc7?!

22...d5 Scary, but Black has just enough resources to defend his position. 23.Bxc4 bxc4

424
24.Qd6 (24.exd5 Bxd5 25.h4 Rf7 26.h5 Qc7 27.Qh4 gxh5 28.Rf2 Rb8 29.Kf1 Rb2 30.Qxh5 Qg3 Black has covered
all points of entry into his camp and now White has to be careful not to lose.) 24...Rfe8! A key defensive move.
Black’s position is safe now because of White’s f6-bishop, which blocks access for its own pieces!! 25.Qg3 dxe4
26.h4 With the idea of opening the h-file for additional lines of attack. 26...e5 27.h5 Qc5+ 28.Kh2 e3 29.hxg6 Rc7!
30.Rd4! e2 31.Re1 hxg6 32.Rh4 Rd7! and Black holds.

23.Bxc4 bxc4 24.Rxd6 Bxe4 25.Rxe6!

Around here I was happy that I was getting at least a little advantage and was able to focus on the plan of how to
render Black’s defense difficult.

25...Qxf4 26.Rxf4² Bf5?!

White’s doubled c2-pawn isn’t going anywhere, so Black decides to save time by not taking it. Right now, Black
wants to exchange all the rooks and get his king out of the corner and on to e6, when it’s an immediate draw. To
play this for a win, White must keep both his rooks on the board and come up with a plan.
425
26...Bxc2 27.h4 Rc7 28.Ra6 Re8 29.Kf2 Bf5 30.Bd4 Ree7 31.Kg3 Kf7 32.Bf2 Red7 33.Rd4 Rxd4 34.Bxd4 Bc8
35.Rf6+ Ke7 36.Kf4 Ke8 37.Ke5±

Black still has to suffer, due to the white king being far advanced and one of his weak pawns about to fall, without
being able to exchange the rooks.

27.Ra6

Hereabouts my opponent realized that, despite the presence of opposite-colored bishops, the position had a lot of
play left in it as the difference in the power of both bishops and rooks is clearly in White’s favor and Black has to
watch for some of his pawns falling without compensation. During the game I thought the position was winning for
me, however with modern computer defenses, Black is just laughing. First of all he can simply put his rooks on the
d- and e-files to prevent White’s king from advancing.

27...Rf7

a) 27...Bxh3!? was an interesting choice. Black gives up two pawns, but activates his king. 28.Rh4 Bf5 29.Rxa7 Rf7
426
30.Rxf7 Kxf7 31.Rxh7+ Ke6 32.Ra7 Bxc2 33.a4 Kd5 34.Bd4 and again, Black still has to fight for a draw; his king
is not active.
b) 27...Rc7! The best defense. 28.Be5 Rd7 29.Kh2 Re8 30.Bd4 Bxc2 31.a4 Bf5 32.Kg3 Rc7 33.a5 Ree7 34.Rf2 Bc8
35.Rd6 Red7 36.Rdf6 Rf7 37.h4 a6 38.Rb6 Rxf2 39.Kxf2 Kf7 40.Ke3±

and despite equal material and opposite-colored bishops, White’s pressure persists.

28.h4 Bxc2

Black decides that since he’s going to have to defend passively, he might as well go ahead and take the pawn. I
agree.

29.Rd4

An inaccuracy, though neither of us realized it during the game.

29.Kf2!? was necessary to prevent Black’s pawn break. 29...h5 30.gxh6 Kh7 31.Kg3 Rcf8 32.Bg5±
427
29...Bf5

29...h5!? An interesting computer suggestion. Black gets rid of his weakness on h7. 30.gxh6! (30.Rdd6 Re8 31.a4
Rb7 32.Kf2 Bf5 33.Rd8 Rxd8 34.Bxd8 Kf7 35.Bf6

35...Bc8! 36.Rc6 Bd7 37.Rxc4 Ke6 with reasonable chances to hold.) 30...Rcf8! 31.Rf4 Kh7 32.Kf2 Kxh6 33.Kg3
Kh7 34.Be5±

30.a4!!

The beginning of a clear plan for White. The idea is to use both rooks and the f6-bishop to create an outside passed
pawn on a6 that will not just deprive Black’s rook of the key b7-square, but also make that square available for
future use as an outpost. A pawn march in such endgames is really quite standard, but nevertheless very effective.

30...Rb7 31.Kf2 Re8 32.Kf3

428
Black can only shuffle his pieces around and watch White improve his position.

32...Kf7?!

I still don’t like this move. Black gives up control of the c6-square and his king’s presence on the seventh rank only
helps White with potential invasions with check!
32...Rc8!? 33.Kf4 Kf7 ( 33...Kf8 34.Rd5 Re8 35.Rda5 Re4+ 36.Kf3 Rd7 37.Bd4 Rxh4 38.Rxa7±) 34.Rd5 Re8
35.Rda5 Re4+ 36.Kf3 Rd7 37.Rxa7 Rxh4 38.Bd4 Re7 39.Rxe7+ Kxe7 40.Re5+ Kd6 41.a5 Bc8 42.Rb5 and White
keeps pressure here as well.

33.Rc6!

I would prefer active defense here as Black. Passivity will get him nowhere.

33...Bd3?!

33...Re1 was a better defense, for example 34.a5 Rf1+ 35.Kg3 Rg1+ 36.Kf4 Rf1+ 37.Ke5 Re1+ 38.Kd6 Re6+
39.Kc5 a6 40.Rb6 Rc7+ 41.Kb4 Re1 42.Rd8 Rb1+ 43.Ka3 Ra1+ 44.Kb2

429
44...Rb1+ (44...Rxa5? 45.Rh8) 45.Ka2 Rb7! 46.Bd4 Rb5 47.Rxb7+ Rxb7 48.Ra8 Rb5 49.Rxa6±.

34.a5

34...Re6??

The decisive blunder.

34...Kg8! was necessary, moving the king off the seventh rank to avoid checks. Black could then prepare the
infiltration of his rooks into the enemy camp to inflict perpetual check. 35.Rcd6! White will exchange one pair of
rooks and maintain great pressure. (35.a6 Rb1 36.Rc7 Rf1+ 37.Kg2 Rfe1! and White has no more than a draw.)

35.a6!+–

In his distaste for passive defense my opponent probably overlooked that he cannot capture on c6.

430
35...Rbe7

35...Rxc6 36.axb7 Rb6 37.Rd7+ Ke6 38.Rxh7 Kf5 39.Re7!+– The key idea. The black king is trapped in a mating
net. 39...Rxb7 40.Re5#

36.Rxe6 Rxe6 37.Rd7+ Ke8 38.Rd8+!

A trick as old as chess itself: White repeats moves to make the time control and calm his nerves when given a
winning opportunity.

38...Kf7 39.Rd7+ Ke8 40.Rxa7

The game is won. White has an outside passed pawn and the black king is cut off on the back rank. He cannot
defend against the decisive march of the white monarch into his camp.

40...Be4+

40...h5 41.Re7++–

41.Ke3!

431
Played for aesthetic reasons. Other moves win as well.

41...Bb7+ 42.Kd4 Bxa6 43.Kd5 Rb6 44.Kc5 Re6 45.Rxh7

1–0

Black resigned as he’s completely tied up. White will capture on g6 and push his connected passers on the kingside.
45.Rxh7

45...Re4 (45...Kf8 46.Rh8+ Kf7 47.Ra8 Zugzwang. Black loses his bishop. 47...Bb7 48.Ra7 Rc6+ 49.Kb5 Rc7
50.Kb6+–) 46.h5! gxh5 47.g6 Rg4 48.Re7+ Kf8 49.g7+ Kg8 50.Re8+ Kf7 51.Rf8+ Ke6 52.Bd4+–

432
My first victory in the match. I was absolutely elated. This game was important not only because it marked my first
win over my personal nemesis (I had an absolutely terrible score vs. Anand before this win), but also because I was
given hope for the first time in the match. Like a lot of top players, I knew that my opponent had an unstable psyche
at that moment and any loss would inflict devastating damage on his self-confidence, which would, in turn, make
him susceptible to further losses.

433
 GAME 16
Viswanathan Anand
Gata Kamsky
Sanghi Nagar 1994
Having scored a huge win in the previous game, but still behind overall, I felt it necessary to keep on the offensive,
even though I had the black pieces. The problem was that at this point, my second, Dzindzi, was the only one left in
the team, and we had no alternative to our main weapon vs. 1.e4, as the Marshall Attack would be expected and
prepared for by my opponent’s team. My father kept running the whole day from kitchen to second floor, suggesting
ideas. The opening used in the tiebreak rapid game, the Scandinavian Defense, was deemed insufficient for a serious
game. Finally, after my father started listing all the openings for Black, we decided to prepare the Archangel line.
We didn’t have much time, but felt that the element of surprise and the character of the opening suited the match
situation perfectly. I felt ready to go and play risky chess, knowing that Vishy was especially vulnerable, both
psychologically and strengthwise, after the loss. Even so, I didn’t expect an easy fight.

1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.0-0 b5 6.Bb3 Bb7

Probably the first critical position of the game. Without preparation it is difficult for White to enter the complicated
main lines, however, if he doesn’t go into a main line, he allows Black to equalize and get a comfortable game.

7.d3

The main move is, of course, 7.c3, with so much theory that it is pointless to go into the details.

7...Bc5

Now White has to decide where to develop his pieces. At the time of the game, this line of the Archangel wasn’t
played much. Many years later, Onischuk would come up with his own version of the line and then, in the 2011
World Cup, I would lose spectacularly to Peter Svidler with White, earning him a Game of the Year award, so,
despite the strong reputation of the Spanish ‘torture’ as a quiet positional opening, it’s anything but boring!

8.a4

White has not yet decided on his scheme of development, so he makes a useful move. In hindsight it is easy to say
that he should have played 8.Nc3, with the plan of putting this knight on d5, followed by Be3, with a very small but
434
comfortable plus because of his well-placed ‘Spanish’ bishop on b3. But it’s one thing to advise it and quite another
to go for such exchanges when you’re one point up in the match and playing White.

8...0-0 9.Bg5

White decides to develop more aggressively, forcing Black to decide how to deal with the pin. There are two ways to
solve the problem One is to return my c5-bishop to the e7-square and tranpose into a typical Spanish Chigorin
structure after ...d7–d6, ...Nc6–a5 and ...c7–c5, where White’s chances would be preferable, since Black has lost a
tempo developing his bishop first to c5, and then retreating it to e7. The other is the approach in the game. Since
there were only two games left in the match, I needed no invitation to demonstrate my aggressive intentions.

9...h6 10.Bh4

White, of course, had the option of taking on f6.

a) 10.Bxf6 Qxf6 11.Nc3 Nd4

435
12.Nxd4 (12.axb5 Nxb3 13.cxb3 axb5 A typical pawn sacrifice in this line. Black gets the two bishops and White’s
extra pawn is doubled. Since the position is still closed, the bishops aren’t operating at full power, and the game is
considered to be balanced. 14.Rxa8 Rxa8 15.Nxb5 c6 16.Nc3 d6 17.Qc2 Bb4 with approximate equality, since
Black keeps control of the a-file, the only one open.) 12...Bxd4 13.axb5 Bxc3

14.bxa6 (14.bxc3 axb5 is completely equal.) 14...Bxb2 15.Ra2 Rxa6 16.Rxb2 Bc6 and despite White’s better bishop
on b3, the game is equal. Black continues with ...Kh8, removing all his pieces from the range of the white bishop.
b) 10.axb5 axb5 11.Rxa8 Bxa8 12.Bxf6 Qxf6 13.Nc3 b4 14.Nd5 Qd6!

15.c3 (15.Ne3 Na5 16.Ba2 b3!? Another typical idea in these structures. Black sacrifices a pawn to block White’s
Spanish bishop and ruin his pawn structure. Black, however, has to play precisely to gain full equality, otherwise
White’s material advantage will have the final say. 17.Bxb3 Bxe3 18.fxe3 Nxb3 19.cxb3 Qb6 20.Nxe5 Qxe3+
21.Kh1 Qd4 22.Nc4 d5 23.exd5 Bxd5 24.Qd2 Rb8³

436
and Black’s bishop, combined with the powerfully centralized queen, gives a dominating performance.) 15...Na5!
16.Ba2 b3! With active play for Black.

10...g5?!

Using the fact that White is unlikely to risk a standard Nxg5 sacrifice, however it was more prudent to play 10...d6
first.
10...d6 11.Nc3 g5 12.Nxg5 hxg5 13.Bxg5 Nd4 14.Nd5! Bxd5 15.Bxd5 c6 16.Ba2

437
White wants Kh1 and f4−f5. A sample line from my post-game analysis: 16...Kg7 17.Kh1 Ne6 18.f4 Nxg5 19.fxg5
Nh7 20.h4 f6!? 21.g6 f5! 22.g3 Kxg6 23.Rxf5 Rxf5 24.exf5+ Kg7 25.Qg4+ Kh8 26.Bf7 Qe7 27.Qh5 Rf8 28.Be6 b4
29.g4 Qe8 30.Qxe8 Rxe8 31.Kg2

and the endgame continues to be just as sharp as the middlegame.

11.Bg3

11.Nxg5?! was still a viable alternative, but here White puts himself in a situation where he’ll be either with the
shield or on the shield. A draw is not a likely result with this line. 11...hxg5 12.Bxg5

438
Black has a choice now, but there is only one correct path: 12...Nd4!.
a) 12...Kg7 13.c3 (13.Qf3? Nd4 14.Qg3! Ne2+–+) 13...Rh8 14.axb5 axb5 15.Rxa8 Bxa8 16.Nd2 Bb7 17.Nf3 Kf8
18.Nh4 Rg8 19.Qc1

and White keeps the initiative as his minor pieces are very effective in disorganizing Black’s counterplay.
b) 12...Be7? would be a major mistake, allowing White to open files against the exposed black king. 13.f4‚ with a
very dangerous attack.
13.Nd2 (13.Ba2 Ne6 14.Bh6 Re8 15.Qf3 d5 16.Nd2 Bf8 17.Qg3+ Ng7 18.f4 Nfh5 19.Qf3 exf4³) 13...Kg7

439
The correct move order. (13...Re8 14.Kh1 Bf8 15.f4 with a dangerous attack for White.) 14.Kh1
[14.c3 Ne6 15.Bxe6 fxe6 16.Nf3

16...Qe8! and with Black’s pieces all looking at the white king, Black starts his counterattack. (16...d6 17.d4 Bb6 is
playable, but not as good.) 17.Nxe5 d6 18.Ng4 Nxg4 19.Qxg4 Qg6 20.b4 Bb6 21.a5 Ba7µ with a clear advantage for
Black.]
14...Rh8 Setting a little trap for White.

440
15.axb5
[15.f4?? Rxh2+!! and it is White’s king that will get mated: 16.Kxh2 Qh8+ 17.Kg3 exf4+ 18.Rxf4 (18.Bxf4 Rg8–+)
18...Nh5+ 19.Kg4 Nxf4 20.Bxf4 Rg8–+.]
15...axb5 16.Rxa8 Bxa8 17.Ba2 Ne6 18.Bxe6 dxe6

19.Nb3 (19.f4 Be3–+ successfully defending against White’s attack, while keeping the extra material.) 19...Be7
20.f4 Nd7 21.Qg4 Kf8µ

11...d6

I remember feeling quite comfortable in this position, however, objectively, White has the better game.
In accordance with the old principle that action on the flank should be met with a counter in the center, White should
have started the skirmish in the center, trying to exploit Black’s open king. I was counting on White’s loss of
confidence to try to challenge me there.

441
12.Nc3?!

Too timid. Black fully equalizes and gets enough positional support to play for an advantage.
12.c3! Challenging Black’s concept of playing on the whole board. Black has played a lot of pawn moves: ...a6,
...b5, ...h6 and ...g5, which can’t be good!! 12...Na5 A more natural and standard response in this type of position.
a) 12...Nh5? Just losing, thanks to the standard little trick. 13.Nxe5! Nxg3 14.Nxc6 Bxc6 15.hxg3+–.
b) 12...Bb6 13.Na3 b4 14.Nc4 Ba7 15.Ne3± White has a big advantage due to the superior placement of his minor
pieces.
c) 12...b4 13.Nbd2 bxc3 14.bxc3 Rb8 15.a5 Bc8 16.Re1 Be6 17.Ba4!

A standard and very strong reaction. White deflects Black’s c6-knight to break through in the center with devastating
effect. 17...Nxa5 18.d4 exd4 19.Nxd4! Very strong; Black’s king is getting more uncomfortable by the move. (If
19.cxd4? Bb4 and Black is fine.)

442
13.Ba2 b4!

14.Nbd2! White continues with the correct strategy of developing and improving his position before committing
himself to a central break when Black would likely be punished for his over-aggressive strategy. (If 14.d4?! exd4
15.cxd4 Ba7³ then Black is just better.)

12...b4 13.Nd5 Na5!

Forcing White to exchange his central d5-knight.

14.Nxf6+

14.Ba2? Nxd5 15.Bxd5 Bxd5 16.exd5 f5 17.d4 Bb6 18.dxe5 f4 19.Qd3 Qd7 and Black captures White’s g3-bishop,
leading to a big material advantage with little compensation for it.

14...Qxf6 15.Ba2

443
15.Bd5? Bxd5 16.exd5 Qg7!? 17.h3 f5 18.d4 exd4 19.Qd3 f4 20.Bh2 Qh7 The extra material and superior minor
pieces give Black a solid advantage.

15...b3?!

Making rushed decisions, trying to capitalize on White’s defeatist attitude. I wanted to play ...f7–f5 as quickly as
possible, but didn’t realize that White will have such strong central counterplay.
15...Qg6!? deserved serious consideration at least.

16.c3 bxc3 17.bxc3 Bb6! avoiding White’s d3–d4 breakthrough with gain of tempo, and planning to play ...Kg7, to
meet d4 with ...f6. 18.Rb1 Kg7 19.Re1 Rae8 Black is ready for his thematic ...f7–f5 push.

White has two choices, either to meet it with exf5, or to try to set Black problems on the opposite wing. 20.Rb2!?
(20.h3? f5µ; 20.Rxb6!? A positional exchange sacrifice, in the spirit of the famous ex-World Champion Tigran
Petrosian, whom I admired greatly when I was growing up. In return, White obtains pressure against the weak black
pawns on the queenside, and any potential black kingside attack will have a much less powerful effect. 20...cxb6
21.Qb1 b5 22.axb5 axb5 23.Qxb5 Ra8 Switching the focus of the battle to the queenside. The position remains
444
balanced. 20.d4? Bxe4 21.dxe5 Bxb1 22.Bxb1 f5 23.exd6 Rxe1+ 24.Nxe1 cxd6 25.Bxd6 Rd8 26.Be5+ Kf8µ 27.Qf3
Nc4 28.Bd4 Qe6!–+) 20...f5 21.exf5 Qxf5 22.d4!

Just in time. Black has to give up one of his attacking bishops to make progress, but then the game fizzles out into a
less exciting phase. 22...Bxf3 23.gxf3 exd4 24.Bxd6! Rxe1+ (24...cxd6? 25.Rxe8 Rxe8 26.Rxb6+–) 25.Qxe1 cxd6
26.Rxb6 Qxf3 27.cxd4 (27.Qe7+ Kh8µ) 27...Qg4+ 28.Kf1 Qh3+ 29.Kg1 Qg4+= with a perpetual.

16.Bxb3

Forced.
16.cxb3? Nc6

A standard reaction; White’s a2-bishop is just awful.

16...Nxb3 17.cxb3

445
17...Qe6!

17...h5?? 18.Qc1!! and White suddenly wins the house as Black has too many unprotected pieces: 18...h4 (18...g4
19.Nxe5+–) 19.Bxe5 dxe5 20.Qxc5+–.

18.Rc1 f5?!

Again feeling the need to rush before White strikes with d3–d4. The way I had played so far was based on intuition
and psychology. What I needed was to start to calculate some lines and realize that Black has to be careful not to fall
into an inferior position.
18...Bb6! was preferable.

I was afraid to give White an additional tempo, but the position is quite balanced because of Black’s strong light-
squared bishop. 19.b4
[19.Re1 f5 20.exf5 Qxf5 21.d4 Bxf3 22.Qxf3 Qxf3 23.gxf3 Bxd4 24.Rxc7 Rxf3 25.Rd7 Raf8!
446
26.Re2 (26.Rxd6 Rxg3+ 27.hxg3 Bxf2+ 28.Kf1 Bc5+–+) 26...Bc5 27.Rc2 R3f5 28.Rb7 a5 29.Rb5 Bb4µ and while
neither side has a material advantage White’s structural deficiencies provide Black with sufficient hopes for a win.)
19...Rac8! Another strong defensive move. Black invites White to create more weaknesses by playing a4–a5. 20.a5
Ba7

21.d4! (21.Re1? Time is of the essence here. This careless move allows Black his ...f7−f5 break in ideal
circumstances. 21...f5 22.exf5 Qxf5 23.d4 Bxf3 24.Qxf3 Qxf3 25.gxf3 Bxd4 26.Kg2 c6 27.Rc2 Kf7 28.Rec1 c5
29.bxc5 Rxc5 30.b4 Rxc2 31.Rxc2 Ke6 and Black should win, for example 32.Rc6 Ra8 33.Rc7 h5 34.h4 gxh4
35.Bxh4 Rb8 36.Rc6 Rxb4 37.Rxa6 Bc5 38.Bd8 Rb2 39.Bb6 Rxf2+ 40.Kg3 h4+ 41.Kg4 Kd5 42.Bxc5 Kxc5 43.Ra8
h3 44.Kg3 h2 45.Rh8 Ra2–+) 21...exd4 22.Re1 c5 23.e5! The point. Black is not allowed to have connected passed
pawns.

447
23...Bxf3 (23...dxe5 24.Rxe5 Qd7 25.bxc5 f6 26.Re1 Bxc5 27.b4 Ba7 28.Qd3

and White has serious counterplay: 28...Qh7 (28...Kg7? 29.Ne5!!+–) 29.Qb3+ Qf7 30.Qd3 Qh7=]
24.Qxf3 dxe5 25.Rxe5 cxb4 26.Rd1 Qc6

448
27.Rd5! f5 28.Bd6! Rcd8 29.Qb3 Kh7 30.R5xd4 Bxd4 31.Bxf8 Bxf2+ 32.Kxf2 Rxf8 33.Qxb4, again with equality.

19.exf5?!

a) 19.Bxe5? fxe4 20.dxe4 g4 (20...Rxf3?? 21.gxf3! Qxe5 22.b4 Bd4 23.Rxc7 Bc8 24.Qd2 Be6 25.Rd1 and White
has both a material advantage and the initiative.) 21.Nd2 Qxe5 22.b4 Bb6 23.a5

23...Bd4 (23...Bxf2+ 24.Rxf2 Rxf2 25.Qxg4+ Kh8 26.Kxf2 Qxb2 27.Qf4! Qxc1 28.Qxh6²) 24.Rxc7 Rf7 25.Nc4 The
start of a nice line. 25...Qxe4 26.Rxb7

449
26...Rxf2! 27.Rxf2 Bxf2+ 28.Kxf2 Rf8+ 29.Kg1 Qxb7 30.Qxg4+ Qg7 31.Qe6+ Qf7 32.Qxf7+ Rxf7 33.b5! axb5
34.Nxd6 Rd7 35.Nxb5 Rd5 36.Nc3 Rxa5 37.Kf2 with a drawn endgame. As we can see. the game is still very
balanced, even after my rushed pawn break.
b) 19.b4! would have posed me serious problems. 19...Bxb4™ 20.Rxc7 Rab8

21.exf5
[21.Nxg5?! The move that we both missed. It seems no better than the ordinary 21.exf5, but has plenty of shock
value! The position is equal, despite Black’s impressive pair of bishops. 21...hxg5 22.Qh5 Rf7 23.Qxg5+ Kf8
24.Rxf7+

450
24...Kxf7 (24...Qxf7 25.f4! allows White to create direct threats against the vulnerable black king, even if Black
seems to escape: 25...Bd2 26.Qh6+ Kg8!? 27.Qxd6 Rc8 28.fxe5 Be3+ 29.Bf2 Bxf2+ 30.Kxf2! fxe4+ 31.Kg1 Qd5
32.Qg6+ Kh8=) 25.exf5 (25.f4 Rg8=) 25...Qf6 26.Qh5+ Kf8 27.Bh4 Qf7

28.Qh8+ Qg8 29.Qf6+ Qf7 30.Qh8+= with a draw by perpetual check.]

21...Rxf5 22.Nd2! Ba5 23.Rc2 Bb6 24.Ne4

451
and Black has to prove that he has enough compensation for the pawn, now that White has managed to reroute his
knight to a far better central location.

19...Qd5?!

Personally I thought this was a brilliant move, tying up White’s minor pieces on the kingside, however the reality is
far different, so the question mark is for the objective strength of the move and the exclamation mark is for the
audacity of the concept.
19...Rxf5!

20.Nd2
[20.b4? White needs to reroute his knight immediately. 20...Bxf3 21.gxf3 Bd4 22.Rxc7 Raf8 23.Qc1 Rxf3 24.Qc4
Qxc4 25.dxc4 h5 26.Rd1

452
26...R8f5! A very mysterious move, the point of which will be revealed later. (26...h4 seems simpler, but requires
more studylike moves later! 27.Rxd4 exd4 28.Bxd6 Re8 29.Re7 Rxe7 30.Bxe7

30...Rf4!! 31.b5 axb5 32.cxb5 d3 33.Kf1 Rd4 34.Ke1 Re4+ 35.Kd2 Rxe7 36.Kxd3 — or 36.a5 Re5 — 36...Kf7–+
and Black is just in time to stop White’s pawn avalanche on the queenside.) 27.Rxd4 exd4 28.b5 d3 29.Rd7 d5!–+
The point behind Black’s mysterious 26th move.]
20...Bb6!!

453
I missed this quiet move completely. 21.b4 (21.Nc4?? Qd5–+; 21.Kh1 Raf8 22.f3?! g4! 23.fxg4?! Rxf1+ 24.Nxf1
Qd5 25.Rc2 Rf2!–+ and Black cuts through to White’s king.) 21...Rf7 22.Ne4 Bd4 23.Rc2 Rb8!³ and Black keeps a
comfortable advantage.

20.d4!

A well-timed break!! 20.b4 would only lead to an equal game: 20...Bxb4 21.Rxc7 Ba5 (21...Rxf5? 22.d4 e4 23.Ne1
e3 24.Nf3! and Black has problems (24.Qh5? doesn’t work: 24...Rf6™ 25.fxe3 Bxe1 26.Rxb7 Rxf1+ 27.Kxf1
Qxb7µ) ) 22.Re7 Bd8

454
23.Rxb7
[23.Re6!? Rxf5 24.Rxh6 Kg7 25.Rh5 Rxf3 (25...Bf6? 26.Ne1 Re8 27.Rh3!!±

White’s rook run is remarkable and ends with White getting an impregnable fortress and decisive material
advantage!) 26.gxf3 Qxf3 27.Qxf3 Bxf3 28.Rh3

455
28...Bg4 29.Kg2 Rb8 30.Rc1 Bd7 31.d4 exd4 32.Bxd6 Bxh3+ 33.Kxh3 Rxb2 34.Be5+ Kf7 with a highly probable
draw.]
23...Qxb7 24.d4! Qd5 25.dxe5 Qxd1 26.Rxd1 dxe5 27.Nxe5 Rxf5

28.Ng4 Ra5 29.Nxh6+ Kg7 30.Ng4 Rxa4 31.h3 Be7 32.Be5+ Kf7 33.Bc3 and this endgame seems to be quite equal.

20...Bxd4

20...exd4!? 21.b4 Bxb4 22.Qxd4 Bc5 23.Qxd5+ Bxd5 24.b4 Bxb4 25.Rxc7∞

21.Rxc7

456
21...Rf7

The only move that allows Black to hope for more than just fighting for survival.
21...Rxf5 22.Qc2! (22.Rxb7!? Qxb7 23.Nxd4 exd4 24.Qxd4 with compensation.) 22...Raf8 (22...Rxf3? is impossible
due 23.Qg6+ with mate to follow.) 23.Rd1 R8f7 24.Rxb7 Qxb7 25.Nxd4 exd4 26.Rxd4 d5 27.h4 and the situation
has changed drastically.

White’s king is now very safe, while Black’s pawn structure is in ruins and his king is wide open to threats. In
addition, White has a pawn majority on the queenside. White is better.

22.Qc2?!

White kindly gives Black the tempo. He would have lost if he had captured on f7 first:
22.Rxf7 Kxf7

457
23.Qe2!? (23.Ne1?? Bxf2+; 23.Nxg5+? hxg5 24.Qh5+ Kf8 25.Qh8+ Ke7 26.Qg7+ Ke8 27.Qh8+ Kd7 28.Qg7+
Kc8–+) 23...Rc8 (23...Rf8 24.b4 Kg8 25.Qc2 would have posed serious problems for Black, since he cannot
recapture the f5-pawn so easily.) 24.Re1!!² This little move is absolutely critical in preventing Black from
attempting to exchange queens via ...Qe4. Black is worse, as he cannot improve his position any further, while
White has many improving moves and two extra pawns!
[24.Rd1 g4 25.Ne1 Qe4 26.Qd2! Qxf5

leads to very unclear situations: 27.Bf4!? (27.Qxh6 Bd5! 28.b4 Bb3 29.Rc1 Rxc1 30.Qxc1 Bxa4 31.Qc7+ Kg8
32.Qxd6 Qb1 33.Qe6+

458
and the black king can’t escape a perpetual check: 33...Kf8 34.Qf6+ Ke8 35.Qe6+ Kd8 36.Qd5+ Kc8 37.Qe6+ Kb7
38.Qe7+ Kc6 39.Qe8+ Kd5 40.Qd8+ Kc4 41.Qg8+ Kxb4 42.Qf8+=

42...Kb3?? 43.Qa3++–) 27...Bd5! 28.Be3 Bxb3 29.Ra1 Qe4

459
The only move, otherwise Black is worse. Now the game is balanced yet again.]

22...Rxc7 23.Qxc7

23...Rf8 24.Qc2

24.Qd7 Bxb2 25.Qe6+ Qxe6 26.fxe6 Re8µ

24...Bc5!

460
Black is going for the only plan left to him in view of the limited amount of pieces. Exchanging into an endgame
where he’ll have two bishops and a rook will give him a comfortable game, especially given White’s locked-in g3-
bishop.

25.Re1!

Preventing ...Qe4.
25.Rd1 Qe4 26.Qc4+ Qxc4 27.bxc4 Bxf3 (27...Rxf5 28.Ne1 h5 29.Nd3 Ba7 30.h4³) 28.gxf3 Rb8 29.Rb1 Rb3µ

25...Bb4!

Black insists!!
25...e4 26.Rd1 Qxf5 27.Nd4

26.Rd1

Had to be played.
26.Re2? Rc8 27.Qb1 e4 (27...Qxb3µ) 28.Ne1 Bd2–+ and Black wins.

26...Qe4

461
27.Qxe4

27.Qc4+ was a trickier choice, but White wanted to prevent the exchange on f3. 27...Qxc4 (27...Kg7??

28.Bxe5+!! dxe5 29.Qe6! and White is suddenly on top. 29...Rf7 30.f6+ Kg8 31.h4! g4 32.Rd7 Bd5 33.Rxd5 gxf3
34.Rd8+ Bf8 35.Rxf8+ Kxf8 36.Qc8#) 28.bxc4

462
28...Bxf3 (28...Rxf5 29.Ne1³) 29.gxf3 Rxf5 30.Kg2 h5 31.h4 gxh4 32.Bxh4 Rf4 33.Be7 Rd4 34.Rxd4 exd4 35.f4
Kf7 36.Bg5 d3 37.Kf3 Ke6 38.Ke3 d2 39.Ke2 Kf5 40.f3=

And the result of this crazy line? It’s still equal!!!

27...Bxe4 28.Ne1?

Natural, but I didn’t realize it was a serious mistake. White had a very tactical opportunity to try to equalize after
28.Nd4, but the best computer line is something neither of us saw, obviously!
a) 28.Nd2 Bxf5µ
b) 28.Nd4! Rc8 (28...Kf7 29.f3 Bd5 30.Ne6 Bxe6 31.fxe6+ Kxe6 32.Rc1 Rb8 33.Bf2=) 29.f3 Bd5 30.Kf1! a5

463
31.Be1!! He had to see this move, which was not easy. 31...exd4 32.Bxb4 axb4 33.Rxd4 Rc5 34.Rxb4 Kf7 35.Rb6=
and there is still some fighting left in the position.

28...Bxf5

A careless move.
28...Rc8!µ While it looks like a simple transposition into the game, White doesn’t have the option of placing his
rook on the c-file.

29.f3?!

a) 29.Nd3? Bxd3 30.Rxd3 Rc8 31.h4

31...Kf7!? ( 31...Rc2 32.hxg5 hxg5 33.Kh2 Rxb2 34.Kh3 Kf7 35.Kg4 Kg6 36.Rf3 Rd2 37.Rf8 Rd4+ 38.Kh3 a5µ)
32.hxg5 hxg5 33.Kh2 Ke6 34.Kh3 Rc2 35.Kg4 Bd2 36.b4 Rxb2 37.b5 axb5 38.axb5 d5 39.f3 d4! 40.b6 Ba5–+
b) 29.Rc1! is much stronger. White doesn’t need to fear the rook exchange as he has sufficient time to activate his
464
pieces before Black regains material equality. 29...Kf7 30.f3

and now if 30...Rc8?! 31.Rxc8 Bxc8 White should hold: 32.Nd3 Ba5 33.b4 Bd8 34.Bf2 Bd7 35.b3 Be6 36.b5 axb5
37.axb5 Bxb3 38.Nb4 Be6=.

29...Rc8µ

We have reached the second critical point of the game. Now that Black has emerged from the middlegame with a
comfortable advantage, it was time to think about how to improve his position. I had started to play quickly and on
impulse, trying to keep my opponent pinned down in time pressure. This was bad strategy, especially against
someone like Mr. Anand, who has been famous for lightning fast and correct decisions throughout his career.

30.h4

30.Nd3 Bxd3 31.Rxd3 Rc1+ 32.Kf2 Bc5+ 33.Ke2 Rg1µ

30...Be6?!
465
The first minor mistake, allowing White to activate his knight. It was much better to bring the king into the game to
prepare for the central d-pawn push: 30...Kf7 31.hxg5 hxg5 32.Kf1

32...Be6! Black’s king is already on f7 compared to the game line.

[32...Ke6 An obvious move, but not the strongest, after which the game enters wild endgame complications!! 33.Bf2
d5 34.Be3 g4 35.Rc1 Following the theme of the game that a rook exchange favors the defensive side. However,
even here, Black can pose serious problems. 35...Rxc1 (35...Rh8?! 36.Rc6+ Kd7 37.Rh6 Rxh6 38.Bxh6³ Black loses
several tempi compared to the other line.) 36.Bxc1

36...Bb1!! Black finds the Achilles heel in White’s fortress. The b3/a4 pawn chain will be under attack, while taking
on g4 will result in Black’s central connected passers being stronger than all the white pawns taken together. (36...d4
37.Ke2 with counterplay) 37.Ke2 Ba2 38.Nd3 (38.Nc2?! An ingenious attempt to save the queenside, but it will fail
for another reason. 38...Be7 39.Na1™ Remarkable! Both Black and White consciously put their minor pieces out of
play, just like in a study!! Of course not 39.b4 Bb3 — 39...Bb1! closing the exit for White’s knight. 40.Bd2 d4 41.b4
d3+! 42.Ke1
466
Best, but White’s a1-knight is still a major factor. 42...Ba2 and Black has successfully imprisoned White’s a1-knight
in the corner and has great winning chances.)
38...Be7

39.Be3! Transposing into a drawn opposite-colored bishops endgame. (39.b4 Bc4 with some pressure still.)
39...gxf3+ 40.gxf3 Bxb3 (40...d4 41.Nc1!!=) 41.Nc5+ Bxc5 42.Bxc5 Bxa4=]
33.Nd3 Bxb3 34.Ra1 Ba5 35.Be1 Bd8 36.Ra3 Bd5

467
37.Bd2 (37.b4 Rc2 38.Rc3 Ra2 and the combination of the rook with two powerful bishops causes White severe
coordination problems. Black has a big advantage.) 37...g4 38.fxg4 Rc2 39.Ke1 (39.Ke2? Bg5–+) 39...Bxg2 40.Rc3
Rxc3 41.Bxc3 Ke6³

and despite material equality, Black’s connected passed pawns supported by the two bishops make all the difference.
A human player has almost no chance of defending this position against another human of similar strength.

31.Nd3

31.Rd3 Rc1 32.Kf1 Bf5 33.Rd5 Bc2µ

31...Bxb3

468
32.Ra1

32.Rc1 Rxc1+ 33.Nxc1 Bc5+ 34.Kf1 Bxa4–+

32...Bd2?

An error in a position very difficult to evaluate in standard terms, as sometimes Black wants to keep his pair of
bishops and sometimes doesn’t.
The important thing to keep in mind is that Black wants to centralize his king and pressurize White’s weak
queenside pawns, while denying White’s d3-knight squares where it can attack Black’s pawns.
32...Ba5 33.hxg5 ( 33.Be1 Bb6+ 34.Bf2 Bxf2+ 35.Nxf2 gxh4) 33...hxg5 34.Be1 Bxe1 ( 34...Bb6+?! 35.Bf2, forcing
Black to exchange bishops. 35...Bd8? 36.a5!=) 35.Nxe1 a5!!µ Key move!

Black keeps the white queenside pawns disconnected and under threat. His plan is to centralize the king, when one
of the white pawns will fall after ...Rb8 and ...Bc4 due to White’s misplaced e1-knight.
469
33.hxg5

33.Be1 Be3+ 34.Kh2 is another defensive idea for White. He wants to put his bishop on c3, where it will apply
pressure on the e5-pawn, prevent ...d6–d5 and also allow the white d3-knight access to the b4-square.

33...hxg5

34.Ra3?!

A return inaccuracy.
a) 34.a5 seems necessary before exchanging the dark-squared bishops. That way Black will not have ...a6–a5, fixing
the white a4-pawn.
b) 34.Be1 Bxe1 35.Nxe1 a5! 36.Ra3 Rb8 37.Kf2 Kf7 38.Ke3 Bc4 39.Rc3 d5–+

34...Bc2??

An absolutely terrible move, spoiling all my efforts in the entire game so far!!!
34...Rb8!

470
was simple and absolutely necessary, keeping White’s rook on a3 and creating the threat of ...Bc4. I admit that I
didn’t see ...a6–a5, but as you can see from the line above, ...a6–a5 is not even necessary. One move is all it took to
change the game from almost winning to a draw. 35.Be1 Bxe1 36.Nxe1 Kf7 37.Kf2 Bc4–+.

35.Be1!

Of course. White eliminates Black’s pair of bishops and now all his pieces become active. 35.Kf2 Bxd3 36.Rxd3
Rc2 37.Kg1 Bb4 38.b3 Kf7µ

35...Be3+ 36.Bf2 Bd2 37.Be1 Bxe1 38.Nxe1

471
38...Bf5?=

Mistakes do not come alone.

38...Kf7! was much stronger of course. 39.Kf2 Ke6 40.b4 Best. (40.Ke2 Bg6 41.Kd2 Rb8 42.Kc3 a5 43.Rb3 Rxb3+
44.Kxb3 Kd5 45.Kc3 Kc5

and the endgame is winning because White’s e1-knight is completely cut off. 46.g3 d5 47.Ng2 d4+ 48.Kd2 Kb4
49.f4 Be4–+) 40...Bf5 41.b5 axb5 42.axb5 Rb8 43.Rb3

472
and White has good chances to hold: 43...Rb6!? 44.g3 d5 (44...Kd5 45.f4 Kc4 46.Rb2 gxf4 47.gxf4 exf4 48.Kf3 Kc3
49.Ra2 Rxb5 50.Kxf4= and White should hold the draw.) 45.f4= with a likely draw.

39.Rb3!

Now White can hit some of Black’s pawns in return. The game is completely equal. I was running short of time and
was very frustrated with such a change in the position.
39.b4 Rc1 40.Kf2 Rb1 41.Nd3 Kf7µ

39...Kf7 40.Rb6 Ke6

Following the classical rule to centralize the king. White is equal, but the game is not finished.

41.a5

Vishy is precise, putting his pawns on dark squares to avoid getting hit. His rook is not only active, it also ties up the
473
black central passed pawn on d6.
41.Rxa6 Rb8 42.g4 Bg6 43.Kf2 Rxb2+ 44.Ke3 Ra2µ

41...Rc1 42.Kf2 Rb1

With the only idea left, to reroute my f5-bishop to a better location, the b5-square! Now a forced sequence of moves
follows that should have resulted in a drawn position.

43.g4 Bg6 44.Ng2

44.Ke2 Be8 45.Kd2 Bb5µ

44...Bd3 45.Ne3 Bb5 46.Nf5 Rxb2+

47.Ke1

Objectively the game is completely equal. My hopes were based on my slightly better-placed pieces.
47.Ke3?? Re2#

47...Kd5

Going for the last chance. Black needs to activate his king or there’s nothing left to hope for.
47...Re2+ 48.Kd1 Rf2 49.Rxd6+ Kf7 50.Rb6! (50.Rd5 Rxf3 51.Rxe5³) 50...Bd3 51.Nd6+ Ke7 52.Ne4 Rxf3
53.Nxg5 Rg3 54.Re6+ Kd7 55.Rxe5 Rxg4 The annihilation which has just occured has resulted in a drawn game.

48.Rxd6+

48.Nxd6 Kd4 49.Nxb5+ axb5

474
50.Kd1! (50.a6? Ke3 51.Kd1 Kxf3 52.a7 Ra2 53.Rxb5 Rxa7 54.Rxe5 Kxg4 55.Ke2 Rf7 56.Rd5 Kh4–+ Transposing
into a well-known theoretical win.) 50...Ke3 51.Rf6!=

48...Kc5 49.Rg6

It was quite possible for White to play Re6, but Rg6 would have to be played anyway: 49.Re6 Re2+ 50.Kd1

50...Bd3 51.Re7 ( 51.Rg6 Kb4 52.Rxg5 Kc3 53.Rg6 Ra2 54.Rc6+ Bc4 55.Ke1!=) 51...Kb4 52.Nd4 Ra2 53.Nc6+
Kc3 54.Ke1 e4 55.fxe4 Re2+ 56.Kf1 Rxe4+ 57.Kf2 Rxg4µ

49...Re2+ 50.Kd1 Rf2

475
51.Rxg5

51.Nd6 was a good alternative: 51...Ba4+ 52.Ke1 Rxf3 53.Ne4+ Kb4 54.Nxg5 Rf4 55.Rxa6 Rxg4 56.Nf7 Bb5
57.Rb6 Rg1+ 58.Kd2 Rg2+ 59.Ke3 Re2+ 60.Kf3 e4+ 61.Kg3= An example of another way for White to equalize.

51...Rxf3 52.Rg8??

An incredible error that was played quickly. Practically any other move would have drawn on the spot.
a) 52.Nh4 Rd3+ 53.Kc2 Kd4 54.Nf5+ Ke4 55.Rg6 Ra3 56.Nd6+ Kf4 57.Nxb5 axb5 58.a6=
b) 52.Rg7! was the easiest: 52...e4 53.Re7= Bd3 54.Kd2 Kc4 55.Nd6+ Kd5 56.Nf5 Rf2+ 57.Ke3 Rf3+ 58.Kd2=.

52...e4

Suddenly White has no good defense against the black pawn reaching the e3-square, and from there, e2... My
opponent went into a deep think here, but it was too late.
476
53.Kd2

53.Rc8+ Kb4 54.Kd2 Rd3+ 55.Ke1 e3 and Black wins, e.g. 56.Rc1 Rd2 57.Rb1+ Kc3 58.Rc1+ Kb2 59.Rc5 Rf2
60.Rxb5+ axb5 61.Nxe3 Rf7–+.

53...Rd3+ 54.Ke1 e3 55.Rc8+ Kb4 56.Rc2

56.g5 e2 57.Rc1 Rf3 58.Rb1+ Kxa5–+

56...Rb3 57.Rc1 Rb2

The game is lost for White. Around here it was adjourned, but brief analysis showed that Black is completely
winning in all lines.

58.Nd4

58.Nxe3 Re2+–+

58...Rg2 59.Nc6+

59.Nxb5 Rg1+ 60.Ke2 Rxc1–+

59...Kb3 60.Nd4+ Kb4

60...Kb2 61.Nxb5 Kxc1 62.Nc7

61.Nc6+ Ka4 62.Ra1+ Kb3

477
63.Nd4+

63.Rb1+ Kc4 64.Ne5+ Kc5 65.Nf3 Rf2–+

63...Kc4 64.Nxb5 Rg1+ 65.Ke2 Rxa1 66.Nc7 Rxa5 67.Kxe3 Ra1

67...Ra1 68.g5 a5 69.Kf2 Rd1 70.Ne6 a4 71.g6 Rd7–+

0–1

Quite a shocking game, with the evaluations changing so much, but the match was now tied, with me playing White
in the last game which history shows was a draw.
After my comeback to tie the match 4–4, my opponent was unable to reclaim his usual form for the tie-breaks, where
I played above my usual level. I employed the the Dzindzi Scandinavian Defense (1.e4 d5 2.exd5 N f6) successfully
in the first of the two tiebreak games, reversing the situation compared to the classical match, as Anand now had to
478
beat me to level the score. In the second game I played a Torre/London System and had an overwhelming position
by move eleven. The match ended in an unusual way when my opponent resigned before move twenty in a position
where he could have provided some resistance, although only to fight for a draw.
The match was and remains a very memorable experience, both in terms of the exotic location and the intensity of
the chess fighting. Of course it was deeply disappointing when, during the PCA Candidates matches in Linares
several months later, my opponent claimed that the match actually ended in a draw, but he only lost on rapid
tiebreaks. Our next match happened on the heels of my FIDE Candidates match victory vs. Salov, but a mere couple
weeks were not enough for me to get the required rest and prepare for Mr. Anand. He had several months to prepare
for me after his PCA Candidates in the fall of 1994, and in our match in Las Palmas, I was not able to pose Mr.
Anand any serious problems. Of course, there was a lot more going on behind the scenes than I am able to relate in
this book, since it is about my memorable games and not politics, but it is a fact of life that chess is no longer just a
meeting of two minds, but of politics, with all the positive and negative implications.

479
 GAME 17
Gata Kamsky
Valery Salov
Sanghi Nagar 1995
After defeating V. Anand in the FIDE quarter-final matches in the summer of 1994, I and the three other winners of
their respective matches had to return to Sanghi Nagar for the semi-final matches, as the organizer had signed up for
both events in advance. The timing of this match was important, because the PCA Candidates semi-finals ended in
September 1994, which meant that V. Anand had had at least several months’ advantage in preparaion for his match
vs. me, as Kasparov & Co. sent me an ultimatum either to accept the scheduling of the PCA finals two weeks after
the FIDE semi-finals or refuse and be disqualified from their PCA World Championship system, even though Bob
Rice himself had approached me, I think right after my match with Kramnik, and asked me to ‘focus’ on their PCA
championship. I was never a fan of Mr. Kasparov, and it was mutual.
Even nowadays Mr. Kasparov is trying to rewrite chess history and omit my achievements, but as I get older, I get
more philosophical about such people. People like Kasparov, who feel a constant need to be in the spotlight, must be
feeling very unhappy with themselves.
But back to the game. My opponent was already a famous grandmaster when I was just starting to learn chess in St.
Petersburg. For many years, there were rumors about Mr. Salov’s mental stability, constantly disappearing from
chess events, only to keep returning. The current FIDE Candidates event was his swansong.
In the first game, where I had the white pieces, I went for the sharp Sicilian line with 9.a3 that was utilized by my
former coach, GM Shabalov. Black achieved a serious advantage in the opening, but closer to the middlegame, one
of my opponent’s longtime faults, bad time management, had its say and he blundered in tactics close to the time
control, leaving me with a decisive material advantage, which, however, I almost failed to convert. The second game
of the match saw me employ a novelty on the black side of the Sämisch King’s Indian. White could have posed me
problems at some stage of the game, but spent too much time and energy trying to figure out the optimal
development of his pieces, giving me a chance to strike for a perpetual and lead the game to a draw. Now came this
third game of the match which was important in deciding the overall outcome.

1.d4

Choosing a more solid approach to the game, after the disaster in the first game. I felt that my opponent had to
demonstrate some surprise in order to try to even the score and to have a chance of overall victory.

1...d5 2.c4 e6 3.Nc3 Be7

480
The first important moment. In my match vs. van der Sterren in the last sixteen of the FIDE Candidates in January
1994 I chose the Exchange line of the Queen’s Gambit Declined and generally achieved positions to my liking.
White usually develops his king’s knight to e2, followed by preparing to grab the center with f2–f3 and e3–e4. If
Black attempts to stop White with some sort of ...c6–c5 counter, White just exchanges that pawn and goes for a
small advantage in the shape of Black’s isolated d-pawn. It seems that my opponent wanted to avoid these positions,
hence this waiting developmental move, as 4.e4 would not bring White any advantage.

4.Nf3

4.e4?! dxe4 5.Nxe4 Nf6³ and Black is already slightly better as he gets a lead in development and White’s d4-pawn
could become a weakness.

4...Nf6 5.Bg5

At this point I was thoroughly surprised by my opponent’s choice of opening and was frantically trying to recall
anything and everything that I could possibly have seen in this position. I briefly recalled some games by GM
Yusupov that I had studied in my childhood, and, of course, those of the famous incomplete first Karpov-Kasparov
match in 1984. And then finally I remembered that Lasker had come up with the system named after him in this line.
I also recalled Karpov’s reaction when faced with the now-famous Kasparov gambit in the Sicilian, where, after a
long think, Karpov just exchanged most of the pieces, leading the game to a draw. So I decided to play logical
moves and keep the game clear of complications, unless absolutely necessary.

5...h6 6.Bh4 Nbd7

481
This move meant that my opponent didn’t want to go for Lasker’s system; he had something else in mind. Could it
be the plan that involves ...c7–c5? Given that White is still developing his kingside, I didn’t like the prospect at all,
but felt that White should at least not be worse.

7.e3 0-0 8.Rc1

The second nuance in such positions: White generally avoids the early development of his king’s bishop so as not to
give Black the tempo of taking on c4. At the same time he brings the rook to the potentially useful c-file and
prepares the b1-square for the f1-bishop, where it will not block the rook from participating in the action.

8...c6

Of course in modern chess everyone plays 8....c5 with very sharp play, but at the time of the match this position did
not arise very often, so it was not easy to analyze with the terrible engines of the time. Black’s plan is to finish the
development of his queenside by means of ...dxc4, ...b5 and ...Bb7, similar to the Slav Meran.

9.Bd3 dxc4 10.Bxc4 b5 11.Bd3 a6 12.a4!?

482
I felt that I should at least try to prevent Black’s ...c6–c5, as after 12.0-0 c5 Black has an easy time.

12...bxa4!?

The position arising after the typical Meran move 12...Bb7 is not easy to evaluate, but let’s have a look at some
possibilities: 13.0-0 b4
[13...Qb6?! If Black attempts to prepare the ...c6–c5 break, he’s in for a surprise: 14.Ne5! Very strong. White
eliminates a key black defender, the d7-knight, which was helping Black with a potential ...c6–c5 break.

14...Rfd8 15.Qe2 b4 16.Nxd7 Rxd7 17.Bxf6 Bxf6 18.Ne4 Be7 19.Nc5 Bxc5 (19...Rdd8 20.Be4±) 20.Rxc5± is a big
advantage for White, since Black is stuck with his weakness on c6, which also blocks the b7-bishop. Technically,
such positions should be won for White as Black lacks any source of counterplay against White’s possible kingside
attack.]
14.Ne4

483
Now complications arise which turn out to be quite favorable for White.
14...Nxe4 15.Bxe7 Nxf2! 16.Bh7+! A very important move!! White wins a tempo, nudging the black king further
away from the center, which might count in the endgame if the white king is rushing to the queenside to collect
pawns.
[16.Bxd8?! Nxd1 17.Rfxd1 Rfxd8 18.Be4

18...Rac8 (18...Nf6 19.Bxc6 Bxc6 20.Rxc6 (…Ne5) 20...Ng4 21.e4² and White’s game is more comfortable, thanks
to his control over the open c-file, although Black has strong defensive chances. For example 19.Nd2² Surprisingly,
despite being a pawn down, White is better, as the difference in bishops is too great and White will eventually
recapture either Black’s weak a6- or b4-pawn. 19...g5 (19...Ba8 20.Nb3 c5 21.Bxa8 Rxa8 22.Nxc5 Nxc5 23.Rxc5
Rd5 24.Rdc1± with a difficult rook endgame for Black) 20.Nb3 Kg7 21.Bf3! and Black has to play carefully to
avoid ending up with a difficult defense, e.g. 21...f5 22.g3 e5?! 23.Na5 Ba8 24.Bxc6 exd4 25.Bxa8 Rxa8 26.Nc6
Rdc8 27.exd4²]

484
16...Kh8™ (16...Kxh7?? 17.Qc2++–) 17.Bxd8 Nxd1 18.Rfxd1 Rfxd8 19.Be4 Nf6 20.Bxc6 Bxc6 21.Rxc6 Ng4
22.Re1 Rdc8 (22...Rac8 23.Rxa6 Rc2 24.h3 Nf6 25.Ne5!² The tempo won with the bishop check on move sixteen
finally plays a role! White gets the e5-square for his knight, threatening Black’s f7-pawn and reaching the critical
c6-square.) 23.Rc5! Rxc5 24.dxc5 Rc8 25.h3 Nf6 26.Rc1 Kg8

And again that extra tempo is felt as Black is forced to spend one whole move to bring his king back. White has the
better chances, thanks to the passed c-pawn. (26...Ne4?! 27.Rc4 Nxc5 28.Nd4+–) 27.a5! Nd5 28.Kf2 f6 29.Rc4 Kf7
30.Ne1! Ne7 31.Nd3 Nc6 32.Nxb4 Nxb4 33.Rxb4 Rxc5 34.Rb7+ Kg6 35.b4²

485
and White keeps an advantage in the rook endgame.

13.Nxa4 Qa5+

Black had been blitzing out these moves and I was looking for a trap, but couldn’t see it. White doesn’t have much
choice here either.

14.Nd2 Bb4

14...c5?! 15.Bxf6 Bxf6 16.Nxc5 e5 17.0-0 exd4 18.Ncb3! Qb6 19.Nc4 Qa7 20.exd4± gives White a healthy extra
pawn with superior development.

15.Nc3 c5 16.Nb3 Qd8

16...Qb6

486
This move is quite dangerous for Black. 17.dxc5 Bxc5 (17...Nxc5? loses on the spot: 18.Nxc5 Bxc5 19.Bxf6 gxf6
20.Qg4+ Kh8 21.Qf4 Kg7 22.0-0 Be7 23.Na4 Qd6 24.Qg4+ Kh8 25.Qe4 f5 26.Qxa8+–) 18.Nxc5 Nxc5 19.Bxf6
Nxd3+ (19...gxf6? allows White a dangerous initiative: 20.Qg4+ Kh8 21.Bb1 Rg8 22.Qh4±) 20.Qxd3 gxf6 21.0-0
Bb7 22.Na4

And now Black manages to keep equality thanks to a series of only moves. 22...Qd8! The best defense.
a) 22...Qb5 23.Qxb5 axb5 24.Nc5 Bd5 25.Nd7 Rfc8 26.Nxf6+ Kg7 27.Nxd5 exd5 28.g3±
b) 22...Qa5? 23.Nc5 Bc6 24.Nxe6 Bb5 25.Qf5 fxe6 26.Qg6+ Kh8 27.Qxh6+ Kg8 28.Qg6+ Kh8 29.Rc5 f5 30.Rfc1
Ra7 31.Rc8 Qb4

487
32.e4! With the idea of bringing White’s second rook into the kingside attack via the third rank. (32.h3? is too slow
after 32...Qe7∞) 32...Rd7 (32...Qe7 33.R1c3+–) 33.Qh6+ Kg8 34.R8c3 Rg7 35.Qxe6+ Rff7 36.Rc8+ Kh7
37.R1c3+–
23.Qe2 Qd5 24.f4 Rad8 25.Nc5 Bc6 26.Rfd1 Qf5 27.b4 Bb5 28.Qf3 Rxd1+ 29.Qxd1 e5 30.Qd5 Kg7² and Black still
faces an unpleasant task in proving equality.

17.0-0

17...cxd4

Forced, otherwise White tags Black’s b4-bishop via the Nc3–a2 maneuver.
17...Bb7?! 18.Na2! and White picks up the bishop pair. Black’s dark-squared bishop was especially important, since
it was both defending the kingside and protecting the d6-square.

488
18.Nxd4 Bb7 19.Be4!

At this point I was quite happy with my position. I felt that White had obtained a small advantage, thanks to all
Black’s queen and ...Bb4 moves. White now has access to the c6-square with his d4-knight, and if he manages to
plant it there he will get a mighty grip on the center and deny Black’s rooks lots of squares.

19...Qb8

An alternative was to play 19...Qb6 when it’s only thanks to the strong engines of today that we can find a delicate
line that leads to a balanced position.
19...Qb6 20.Na4 Qa7

21.Bc6! During the game I thought this was a dangerous move for Black, and indeed it is. Black is forced to
calculate the only line that leads to equality.
a) 21.Nc6 Bxc6 22.Bxc6 Rad8 23.Qf3 (23.Qe2 Ne5= In view of Black’s ...Rd2 threat, White is forced to give up one
of his bishops, and with it any hope of the advantage.) 23...Ne5 24.Qg3 Ng6!=
489
After this move Black manages to exchange White’s dark-squared bishop, which leads to an equal opposite-colored
bishops middlegame.
b) 21.Rc7!? Nxe4 22.Rxd7 a5! A ver y important little pawn move, not so much to protect Black’s b4-bishop as to
gain access to the a6-square for his cornered queen on a7. While Black’s king looks vulnerable and White has his
mega-active d7-rook, White doesn’t have much else and the position is roughly equal. 23.f3 Nd2 24.Re1 Rfc8
25.Nb5 Qa6 26.Nc7 Qc6 27.Nxa8 Rxa8

(27...Bxa8 28.Ra7∞ ) 28.Nc5! A very cool move, working on the concept of overload. (28.Rxd2 Bxd2 29.Qxd2
Qxa4=; 28.Rd4 e5=) 28...Ba6! A quiet move in turn by Black that also exploits the overloaded c5-knight. 29.Be7
Nc4 30.Rd8+ (30.Re2?? Bxc5–+) 30...Rxd8 31.Qxd8+ Kh7 32.Nd7 Bxe1 33.Nf8+ Kg8= and despite White’s pieces
weaving a mating net around Black’s king, White has nothing better than to go for the perpetual.
21...Rac8

490
22.Qf3 (22.Bxf6 Nxf6 23.Qf3, hoping for an exchange of lightsquared bishops and to plant the d4-knight on c6, runs
into 23...Nd5! 24.Rc4 Ne7!=) 22...Ne5 23.Qg3! The point; Black has no time to capture White’s c6-bishop.
23...Ned7 (23...Ng6? 24.Bxf6 gxf6 25.Bxb7 Qxb7 26.Nxe6± wins a pawn.) 24.Bxd7 Nxd7 25.Rc7 Rxc7 26.Qxc7

Now it looks as though White has finally achieved something and managed to tie up Black’s entire queenside.
However... 26...Nb8!! We have reached an absolutely unbelievable position. It is definitely not recommended for
anyone with claustrophobia to attempt this position with Black!! Despite White’s seemingly impressive
achievement, Black is fine, and in fact can hope to obtain the better position, since he has a strong pair of bishops,
especially the light-squared one. White’s back rank weakness is also important enough to prevent him from
achieving anything tangible. 27.Qg3 This seems to be the best. White threatens Bf6, forcing Black to return his
knight to the d7-square. (27.Be7 If White could just exchange his dark-squared bishop, he would get a large
advantage, but alas... 27...Rc8–+ and Black is winning. 27.Nb6 Ba5 28.Bd8 Bxb6 29.Qxb6 Qa8!

491
and Black has just achieved the advantage. It is almost as though Black is playing in praying mantis kung fu style on
the chess board. With purely defensive moves he has invited White to over-extend and, if he does, he will fall into
the numerous traps that Black has prepared. 27.Rd1 Rc8 28.Qb6 Bc5! 29.Qxa7 Bxa7³) 27...Nd7

28.Qc7= and White has to repeat the position, since other moves lead to an inferior game. (28.Rc1?! Rc8!³ Inviting
White to make the last aggressive mistake in the game... 29.Rc7?? Qb8!µ and Black takes over.)

492
20.Nc6

At this point, quite unwittingly, I deviated from game 12 of the Fischer-Spassky World Championship match!!!
20.Bg3!? Quite possibly Bobby’s move is stronger than my over-the-board inspiration. 20...Qa7 21.Nc6 Bxc6
22.Bxc6 Rac8 23.Na4 Rfd8

24.Bf3!?
[24.Qe2 would be met by... 24...Nc5! 25.Nxc5 Rxc6 26.Ne4 Rxc1 27.Nxf6+ gxf6 28.Rxc1 Qb7!

493
It is important not only to take the long diagonal, but also to prepare the exchange of rooks on the c-file before
White unleashes his deadly kingside attack. 29.h3 White needs to create a luft for his king, however it gives Black
sufficient time to mount an adequate defense. (29.Rc7 Qb5 30.Qf3 Qf5=; 29.Qg4+ Kh7 30.Rc7 Rc8! The point; now
Black defends ever y thing and equalizes.) 29...Bf8! 30.Qg4+ (30.Rc7 Qb5 31.Qf3 Bg7=) 30...Bg7=]
24...a5!?

25.Rc6?! Perhaps White’s only inaccuracy in the above mentioned Fischer-Spassky game, after which Black
manages to equalize completely. It was important to retain control of the queenside light squares, when White would
keep chances for an advantage with his strong pair of bishops in the endgame. (25.Rxc8!? Rxc8 26.Qd3 Nc5
27.Qc4! Rd8 28.Nxc5 Rc8 29.Qa6 Qxc5 30.Rd1 Qc2 31.Qb7 Preventing Ne4. 31...Rf8 32.Be5! Qf5 33.Bd4²

494
and White has achieved a certain measure of success. Black has to play very carefully to keep the balance.)
25...Rxc6 26.Bxc6² Fischer-Spassky, Reykjavik 1972.

20...Bxc6

20...Qd6!?

I dare say that neither of us could possibly have seen this move at the board. The complications arising from the
lines here extend far into the endgame, which I’m not going to cover, since it’ll take too much space!! I suggest that
the reader practice his ability to analyze and calculate possible lines for both sides to a definitive conclusion, as some
of the lines are quite aesthetic!! I’ll give one sample line just to show how complicated this can be:
21.Qxd6 Bxd6 22.Bf3 Rac8 (22...Bxc6 23.Bxc6 Rac8) 23.Bg3 Bxc6 24.Bxd6 Rfe8 25.Be2 Bb7

495
26.Nb1!±
Back to 20...Bxc6, after which Black had an absolutely amazing resource...

21.Bxc6 Ra7 22.Bg3 Ne5!

As I learned later, even this move — and some later ones — had been played before!!! This move is actually forced,
as alternatives give White an undisputed advantage.
a) 22...e5?! 23.Nd5 Bd6 24.Bh4 Nxd5 25.Bxd5±
b) 22...Qb6?? This even loses on the spot. 23.Na4 Qa5 24.Qd4+– Black’s a7-rook has just run out of squares.

23.Qd4!?

When I saw this move, I felt pretty good. Only here do I finally deviate from previous games, and Black faces a
difficult choice. At this moment my opponent finally stopped blitzing out his moves and went into a long think.

496
a) 23.Qe2 Bd6 24.Rfd1 Bc7 25.Ne4 Neg4!=
b) 23.Ne4 Nxe4 24.Bxe4 Bd6

25.Rc3!? A very strong and original idea. White intends to seize control of the b-file from Black and then try to hit
Black’s a6-pawn with his rooks, focusing the game on the queenside. Black has to defend carefully as White keeps a
small, comfortable advantage which, however, can easily grow into a more serious one at the first inaccuracy by
Black. (25.Qe2?! Nd7 26.Bxd6= was seen in Stempin-Bozek, 1987, which resulted in a draw.) 25...Rd8 26.Rb3
(26.Qc2 Rc7 27.Rxc7 Qxc7 28.Qxc7 Bxc7 29.Rc1 Bb8 30.f3 f5 31.Bb7 a5 32.Kf1 Bd6 33.Be1 Kf7 34.Bxa5 Rb8
35.Bc7 Bxc7 36.Rxc7+ Kf6=

and Black holds the balance despite being a pawn down.; 26.Qe2 Qb5!=) 26...Qc7 27.Qe2 Qc4!= (27...a5? 28.Rd1±)

23...Bd6

23...Nxc6? is obviously bad for tactical reasons: 24.Bxb8 Nxd4 25.Bxa7 Bxc3 26.Bxd4±.

24.Ne4
497
Natural and strong. White continues to open the position for his bishop pair and attempts to pin Black’s e5-knight to
the d6-bishop, causing Black serious discomfort.

24...Nxc6?

I’d say this was the decisive mistake. Black had probably considered that in structures with doubled f-pawns, he
would keep the balance and be safe, but this is not the case.
24...Nxe4! 25.Bxe4 Rd8 26.Rfd1

26...Rc7!! Probably the move we both missed. This only defense by the computer allows Black to liquidate into an
equal endgame. (26...Rad7? 27.Bxe5 Bxe5 28.Qxd7 Rxd7 29.Rxd7!±

498
is the line we both saw, and Black looks to be in trouble as the two rooks are much stronger than the queen in this
position, coupled with the fact that the attacker’s chances are much higher in the presence of opposite-colored
bishops, as his bishop has no counterpart. Just to illustrate how dangerous the black position is, here is a sample line.
29...g6 30.Rb7 Qd8 31.g3 h5 32.h4 a5 33.Kg2 Kg7 34.Rc6! Qa8 35.Rc5 Bd6 36.Rg5+– Kh6 37.Bf3 Qf8 38.Rxa5+–
with a technical win for White.) 27.Rc2!? Rxc2 28.Bxc2

28...Rc8 (28...Bc7 29.Qc3 Rxd1+ 30.Bxd1 Bd6 31.Be2 a5 32.h3² with some small chances for White still, due to the
fact that it’s not easy for Black to exchange the dark-squared bishops, thus allowing White the bishop pair.) 29.Bb1
Bc7! Preparing to meet 30.Qe4 with ...Ng6, exchanging the dark-squared bishop and aiming for 30...Rd8. 30.Rc1
Bd6 31.Rxc8+ Qxc8 32.Qd1 a5 33.h3 Qc7=

25.Nxf6+ gxf6

499
Of course Black is now in serious trouble.

26.Rxc6±

White’s advantage is obvious. He has superior pieces, both major and minor, and the black king is wide open thanks
to the ruined kingside pawn structure. The game is not lost, but compared to the situation a couple of moves ago,
Black has just made his life a lot more dangerous.
26.Qxd6? would be a bad mistake: 26...Qxd6 27.Bxd6 Rc8².

26...Be5

Black now attempts to set up a defense around the e5-bishop. It’s not easy to see how to break through, but at this
point I remembered my fantastic game vs. Kramnik (see earlier) and the motif from there. The reader can guess my
follow-up thought!!
26...Bxg3 27.hxg3 [27.fxg3!? f5 28.g4!

500
28...Rd8 29.Qf6 fxg4 30.Qxh6 (30.Rc4?! Kh7™) 30...Qb5 31.Qh4 Qxc6 32.Qxd8+ Kh7 33.Qh4+ Kg8 34.Qxg4+±]
27...f5

28.Rd1!± An important move. White keeps control of the main open files and prepares for the infiltration of Black’s
kingside. White’s potential g3–g4 pawn break will create additional mating threats. 28...Rb7 29.Qf6 Kh7 30.Rcd6
Rc7 31.Rd8 Qb4!

Tricky defense. White now has to watch out for tactical motifs. 32.f3! The simplest. White prepares Kh2 followed
by R1d4–h4. ( 32.Kh2!! Qg4!=; 32.R1d6?? Qe1+ 33.Kh2 Rc1!–+) 32...Qc5 33.Kh2

501
33...f4 ( 33...Rc8 34.R8d7 Rc7 35.Qxe6+– Prosaic but simple. White captures the crucial e6-pawn, after which
Black’s entire kingside pawn structure falls. The position becomes a technical win.) 34.g4 fxe3 35.g5! hxg5 36.Rh1!
and mate follows.

27.Qg4+

First White wins a tempo to improve the position of her white majesty.

27...Kh7?!

27...Kh8 was Black’s best defense, but White would still keep his advantage and chances for a win. 28.Qh4!
(Despite what other commentators have said, 28.Qh5? throws away the advantage: 28...Kh7 29.Rc4 Bxg3 30.hxg3
Qb5=.) 28...Kg7 (28...Kh7?! would transpose into the game.) 29.f4

502
Just as in the game, this is the only way for White to make progress. All the white pieces are participating in the
attack on the black king. 29...Bxb2 In such a bad position, at least take a pawn! 30.f5 Be5™
[30...Qb5? 31.Bf4 Rh8 32.Qg4+ Kh7

33.Bxh6! Rg8 (33...Kxh6 34.Rf3+– and mate follows.) 34.Qh3 Rd8 35.Bf4+ Kg7 36.Qg3+ Kh7 37.Qh4+! The
reason behind this stairwell line of checks will soon become clear. 37...Kg8 38.Bh6 Kh7 39.Rc4!+– The rook is
protected and comes swinging over to the kingside, deciding the game.]
31.fxe6 fxe6 32.Rxe6 Bxg3 33.hxg3 Raf7!

503
34.Rf4! h5 ( 34...Qb5 35.e4 Kh7 36.e5+– and once again the push of the e-pawn decides the game.) 35.Qxh5 Qb1+
36.Rf1 Qg6 37.Qf3 a5 38.Re4! f5

39.Rh4 A fantastic rook transfer!! White has not only prevented Black’s a5-pawn from moving, but also made the h-
file safe for his king and dangerous for the black king! In addition, potential pawn breaks become lethal. Black’s
defensive task looks increasingly unenviable with each passing move. 39...Re8 40.Qf4 Re4 41.Qb8 Rxh4 42.gxh4
Qe6 43.Rf3+–.
After the text move White goes for a direct attack.

28.Qe4+ Kg7

504
29.f4!

First I need to redeploy my dark-squared bishop to a better location, and what location could possibly be better than
c3!! Black’s monarch will have to face the deathly stare of the c3-bishop and pray that he can survive.

29...Bc7?

The final mistake, after which Black will be unable to save the game.
a) 29...Bxb2!? was still possible, with the idea of exchanging White’s g3-bishop as soon as possible. 30.f5 (30.Rb1
Rb7 31.Rxa6 Rb4; 30.Qc2?! looks tricky, but is nothing in fact: 30...Rb7 31.f5 Be5=) 30...Be5

31.fxe6 ( 31.Rb1? exf5 32.Qxf5 Rb7=) 31...Bxg3 (31...fxe6 32.Rb1 Qe8 33.Bxe5 fxe5 34.Qxe5+ Rf6 35.Rbb6 Qd8
36.Rd6 Rd7 37.Rb8+–) 32.Qg4+ (32.e7 Bxh2+ 33.Kh1 Re8 34.Qg4++–) 32...Kh7 33.hxg3 fxe6 34.Rxe6

505
34...Qd8 (34...Rg7?? 35.Qe4+ Kg8 36.Rfxf6 Rxf6 37.Rxf6 and Black is lost: 37...Qxg3? 38.Qe6++–) 35.Qf5+ Kg7
36.Rf4 Kh8 37.Rh4 Rh7 38.Rxa6+– and as we have observed in similar lines above, Black has to wait passively for
White to slowly improve his position. In a lot of lines Black is simply put into zugzwang, running out of moves and
forced to give up more material to avoid suffocation.

38...Qd7 39.Re6 Qg7 40.Rg4 Qf7 41.Rg6+–.

b) 29...Bd6 seems to be best. Black attempts to prevent White from playing f5 at all costs. White now has two
attractive options at his disposal, however neither of them provides the win, only chances of a good advantage.
30.Bh4 (30.Be1!? Rc8 31.Rxc8 Qxc8 32.Bc3 Be7 33.Rf3 Rd7 34.Rg3+ Kh8 35.Qf3 Qf8 36.h3 Rd5 37.Kh2 Rf5!
38.e4 Rb5 39.Qf2!±

506
All White needs to do is infiltrate Black’s camp, however Black has covered all his weaknesses and it’s not easy for
White to find a finishing blow.) 30...Be7 31.f5 Rc8 32.Qg4+ Kh7 33.fxe6 (33.Rxe6?? Rg8–+) 33...Rxc6 34.Qe4+
Kg7 35.Qxc6

35...Qxb2 (35...fxe6 36.Qxe6 Qxb2 37.Bg3!± White goes for the slow grinding attack, thanks to the unstoppable
presence of his powerful queen on e6. Black will be forced to give up material to stop White’s mating threats, and
whether it will be enough to equalize remains to be seen.) 36.exf7 Kxf7 37.Qc4+ Kg7 38.Qg4+ Kf7 39.Bg3 Qb5
40.Qe4 Kg7 41.h3 a5 42.Rf5 Qe8 43.Rf3 Qg6 44.Qd4!±

507
Disrupting coordination among the black pieces is vital. Black’s a7-rook is overloaded, trying to protect both the a-
pawn and the seventh rank, while Black’s queen has to protect the e7-bishop, the open black king and the h6-pawn.
Something has to give. 44...Ra8 45.Bf4 Kh8 46.Rg3± Winning a pawn with good chances of winning the game.

30.Be1!

White’s bishop is heading to its final destination.

30...Qb5

30...Qxb2? 31.Bc3 Qb8 32.Rf3+– The mating attack is unstoppable. 32...Rd8 33.Rg3+ Kf8 34.Rxe6! fxe6 35.Qh7+–

31.Rf3!

It’s never too late to make a mistake.

31.Bc3?! immediately gives Black some hope. One of the important chess rules of thumb is that ‘the threat is is
508
stronger than the execution’!!! 31...Ba5 32.Bd4 Rc7 33.Rxc7 Bxc7 34.Rf3 Bd8

31...Rd8

a) 31...Qxb2 32.Bc3 Qc1+ 33.Rf1 Qa3 34.Qd4+–

b) 31...Ba5 32.Rg3+ Kh8 33.Qd4!+–

The point behind 31.Rf3. Black cannot defend everything with just his queen and bishop. 33...Bd8 34.Rc8 Rd7
35.Rxd8! Rxd4 36.Rxf8+ Kh7 37.Rxf7+ Kh8 38.Rxf6+–

32.Rg3+ Kh8

32...Kf8 33.Bb4+ Ke8 34.Rxe6+!+–

33.h3?

Here I relaxed too much, trying to find the most unpleasant move for my opponent in his typical time trouble. With
the correct defense, Black could possibly have defended his position. Seeing Black’s last move gave me the thought
that perhaps Black wanted to try to trade queens on d5 and that gave me an idea for the trap.
The correct move was 33.Qf3!+–

509
not allowing Black to play ...Rg8 and at the same time creating the deadly threat of Qg4–g7 mate. 33...Qf5 (33...Ba5
34.Qg4+–; 33...Rg8 34.Bc3 Bd8 35.Rxg8+ Kxg8 36.Rc8 Qd7 37.Qg4+ Kf8 38.Bb4++–) 34.h3! And only now this
luft. Black is lost. 34...Rg8 35.Rxg8+ Kxg8 36.Bc3 e5 37.Qd5 Kg7 38.Qc5 Qe4 39.Kh2

39...Bd8 ( 39...Bb8 40.Rc8+–) 40.fxe5 fxe5 41.Bxe5+ Kh7 42.Qxa7 Qxe5+ 43.Kg1+– with a decisive material
advantage.

33...Qd5?

Black falls into the trap.

a) 33...Rd1 34.Rxc7 Rxe1+ 35.Kh2+–
b) 33...Rg8! was the only move, and would most likely have shocked me. It was the second time in this match that I
got lazy with my calculations and just played moves that felt good. 34.Rxg8+ Kxg8 35.Bc3 Bd8 36.Bd4 Rd7 37.Rc8
Kg7 38.Qf3

510
I would most likely have blitzed out these moves as they seem most natural. 38...Qf5 Again the only move. (
38...Qd5?? 39.Qg3+ Kh7 40.Bxf6+–) 39.Qa8 ( 39.e4 Qg6 40.Bc3 h5 41.Kf1 Be7± but the black position is hard to
break through.) 39...Qd5 40.Qxa6±

In practice this should be a technical win, since White has an extra pawn, superior kingside pawn structure and much
better pieces, but it would still require a lot of work. 40...Bc7 41.Re8! Rd8 (41...e5 42.Qe2+–) 42.Rxd8 Qxd8
43.Qe2 Bb6 44.Bc3!

511
White must keep the bishops on board to win this — slowly. As the black king remains very vulnerable, White
should improve his king position to f2, accompanied by slowly pushing his kingside pawn chain. Let’s see how the
game could continue: 44...Qe7 45.Kh2 Bc5 46.Qe1! Bb6 47.Qg3+ Kh7 48.Qf3 Kg7 49.b4 Qd8 50.Kg3 Kh7 51.Qh5
Kg7 52.Qe2 Kf8 53.b5 Kg7 54.Qg4+ Kh7 55.Qh5 Kg7 56.Qf3! Ba7 57.Qg4+ Kh7 58.Qe2 Bb6 59.Qd2 Qe8
60.Qd3+ Kg7

61.Kh4! Qb8 62.Kh5! White’s king feels very comfortable here, where it aids the attack and is out of the way of
Black’s potential checks. 62...Qc7 63.Bb4 Ba7 (63...Qc1 64.Qd6+–) 64.g4 Bb6 65.Qd6!

512
White transposes into a winning bishop endgame. 65...Qxd6 66.Bxd6 Kh7 67.e4+– Kg7 68.f5 Kh7 69.Ba3 Kg7
70.Be7+– Black is in zugzwang. If the bishop moves, the b5-pawn will reach the promotion square, and if the king
moves, either the f6- or h6-pawn will fall.

34.Qc2!+–

The white queen moves back with tempo, preparing both Be1–c3 and the e3–e4 push.

34...Bd6?!

a) 34...Qd1 35.Qc3+–
b) 34...Rd7 35.Bc3 Bd8 36.Bd4 Ra8 37.Rc5 Qd6 38.Rh5 Qf8 39.Qc5 Be7 40.Bxf6+ Bxf6 41.Qxf8+ Rxf8 42.Rxh6#
c) 34...Bb8 35.Bc3

35.e4
513
Black resigned. 1–0

I went on to win the match 5½-1½, then we flew directly to Las Palmas, where my PCA Candidates match with
Anand was to take place about two weeks later. We decided to create a new weapon to compensate for the lack of
serious preparation, and thus the idea of playing the Zaitsev Variation of the Spanish game was born, along with an
invitation to GM Alexander Ivanov of the USA to join the team, which he thankfully accepted. The Zaitsev held for
a couple of black games, but was then ultimately dismantled in large part thanks to my inexperience with it and
insufficient time for preparation. With White we could do nothing against Yusupov’s Open Spanish and Vishy won
the match in convincing fashion. Still, if we had had at least a month to prepare, compared to several months of
preparation by Anand, who knows what we could have prepared and how it might have changed the outcome. From
that time on I became bitter towards Mr. Kasparov and the dictatorship-like organization of his ‘professional’
association world championship system. Needless to say, in 2000, when Kasparov lost his title to his former student
and helper Kramnik, the whole PCA came crumbling down. The schism that had been created in chess continued
until the reunification match in 2006.

514
 GAME 18
Gata Kamsky
Anatoly Karpov
Dos Hermanas 1995
After winning our final FIDE Candidates matches, Karpov and I were waiting for FIDE to announce the location and
prize fund for our FIDE World Championship match. In the meantime we played in some tournaments together,
such as the one in Dos Hermanas, Spain. Chess was very popular in Spain in those days, and while towns like
Linares and Dos Hermanas were small, they managed to organize tournaments that rivaled modern traditions like
Wijk aan Zee and Dortmund. It was obvious that both of us were trying to hide some of our preparation before the
match, though the Caro-Kann remained Karpov’s main defense in the 1996 World Championship match. Afterwards
I was quite happy with the quality of the game, until the last few moves where I felt that I missed a simpler win.

1.e4 c6 2.d4 d5

My opponent has been playing Caro-Kann his whole life, so it was expected.

3.e5

The Advance Variation, championed by former world title contender, English GM Nigel Short. With its help he
managed to win his Candidates match vs. Karpov and go on for his title shot vs. Kasparov in 1993.

3...Bf5 4.Nf3 e6 5.Be2 c5

This and other lines were very topical and heavily analyzed by the top players at the time. It remains one of White’s
main weapons today, although Black rarely plays 5...c5 any more, as it requires really good memory.

6.Be3 Nd7

This move is not considered one of the top choices any more.
The main line, which continues for many moves, and which we are not going to examine here, starts with 6...Qb6.
At the end, according to the computer and numerous games, Black holds the balance in a slightly worse endgame!!
Very few people want to go there, although there are some stalwarts who keep playing the line.

7.c4
515
In general 7.0-0 is considered to be White’s main move, however these often transpose into each other and, keeping
in mind my opponent’s well-known time management problems, I wanted to keep surprising him.

7...dxc4

But this is rare, as 7...cxd4 has been tried here more often.
7...cxd4 8.Nxd4 Bxb1 9.Rxb1 Bb4+

10.Kf1 (10.Bd2 Bxd2+ 11.Qxd2 Nxe5 12.cxd5 Qxd5 13.0-0 Ne7 14.Qb4 Qd7= Black is completely fine here.)
10...Ne7 11.Qa4 a5!? A computer suggestion.
[11...Qa5 is a standard move here that leads to an equal endgame: 12.Qxa5 Bxa5 13.b4 Bc7!

516
Keeping the b6-square for the d7-knight. 14.f4 dxc4 15.Kf2 g5 16.g3 gxf4 17.gxf4 a6! 18.Bxc4 (18.b5!? Nb6
19.bxa6 Rxa6 — and not 19...bxa6 20.Nxe6± — 20.Nb5 Kd7! 21.Rhc1 Ned5 22.Nxc7 Kxc7 23.Bxb6+ Rxb6
24.Rxb6 Nxb6 25.Bxc4 Nxc4 26.Rxc4+ Kd7= with a draw in sight.) 18...Nb6

19.Rhc1 (19.Be2?! Nbd5 20.Bf3 Nxe3 21.Kxe3 Bb6³ Black manages to take over the initiative in the endgame.)
19...Nxc4 20.Rxc4 Nd5 21.Rbc1 Kd7 22.Kf3 b5 23.Rc6 Rhc8 24.a3 Ne7 25.R6c2 Bd8 26.Ke4 Rxc2 27.Rxc2 Nd5=]
12.a3 Bc5

517
13.Nxe6 fxe6 14.Bxc5 0-0 15.Bd6 Nb6 16.Qb5 Nxc4 17.Bxc4 dxc4 18.g3 Nf5

The point. White cannot capture Black’s f8-rook because he gets mated!! 19.Kg2 (19.Bxf8? Qd3+ 20.Kg2 Qe4+
21.f3 Ne3+ 22.Kf2 Rxf8 23.f4 Ng4+ 24.Kg1 Qe3+ 25.Kg2 Qe2+ 26.Kh3 Ne3–+) 19...Nxd6 20.exd6 Qxd6 21.Qxc4
Rac8 22.Qe4 Qd5 with equality.

8.0-0

Transposing into the main lines, however the alternatives were quite interesting:
a) 8.Nc3 Ne7 9.Bxc4 Nc6 (9...a6 10.0-0 is yet another transposition.) 10.d5

518
10...exd5
[10...Nb6?! happened in Anand vs. Ivanchuk 1995 PCA rapid games. 11.Bg5!N (White was quite surprised and lost
the game after 11.Bb3 exd5 12.Nxd5 Be7=.) 11...Be7 (11...f6? 12.exf6 gxf6 13.dxc6 Qxd1+ 14.Rxd1 Nxc4 15.cxb7
Rb8 16.Bxf6 Rg8 17.b3+–; 11...Qc8? 12.dxc6 Nxc4 13.cxb7 Qxb7 14.Qa4++– The point.) 12.Bxe7 Kxe7!?

This recapture might have been the idea behind Vassily’s new move (10...Nb6?!), however White keeps the
advantage by simple means. (12...Nxe7 13.Bb5+±) 13.d6+ Kf8 14.Bb5±]
11.Bxd5

519
11...Be7 (11...Nb6?! 12.Be4²) 12.0-0 0-0 13.Qb3 Nb6 14.Be4 Bxe4 15.Nxe4 c4=
b) 8.Na3!?

8...c3! 9.Bg5 f6 10.exf6 gxf6 11.Bc1 cxb2 12.Bxb2 Qa5+ 13.Nd2 cxd4 14.Nac4 Qc7 15.Bxd4 Ne7 16.Ne3 Bg6
17.Rc1 Nc6∞

520
This position is quite difficult to evaluate, due to its unbalanced nature. Black has an extra pawn, but his king and
central pawns are weak, yet White’s pieces are not that well placed either. It is most likely to be quite equal at the
end.

8...a6

As is typical of this line in CaroKann, both sides can develop in many possible ways which often interchange and
transpose into one another. Such is the case with 8...a6. It’s a very useful move because it takes away the b5-square
from White’s c4-bishop, thus protecting Black against checks along the a4–e8 diagonal.

9.Bxc4

The pawn must be recaptured, as Black was ready to play ...b7–b5, protecting his extra c4-pawn.
White did have a curious alternative here: 9.d5!? This move seems a bit premature, especially in a view of White’s
underdeveloped queenside, however Black has to play precisely in order not to fall into a worse position. 9...Ne7!

521
[9...exd5?! 10.Qxd5 Qc7 11.Ng5! Nb6 12.Qf3 Nh6 13.Nc3

This move sets a trap for Black. 13...Bg4 (13...Be7? 14.Nxf7! Kxf7 15.Bxh6+–) 14.Qe4 Bf5 15.Qf4± White has the
advantage here in a view of his massive lead in development and Black’s horribly misplaced knights. Black’s extra
c-pawn is doubled and not felt anyway, but the open lines to the black king render him vulnerable to White’s
offensive.]
10.Nc3

The strongest move for White. (10.dxe6?! Bxe6 11.Qc1 b5³ gives Black a comfortable game; 10.d6? Ng6µ allows
Black to capture White’s too far advanced central pawns.) The position is very dynamic, with White trying to open
the center, where his superior development will cause serious problems for Black, for example: 10...exd5
[10...Ng6?! 11.Ng5 h6 12.Nxf7 Kxf7 13.dxe6+ Bxe6 14.f4! with a strong initiative for the sacrificed material;
10...b5?! Too slow. 11.Nh4! White identifies Black’s f5-bishop as his best defensive piece, so sets about eliminating
it. White’s central pawns create a certain pressure in the center and once Black’s light-squared bishop is gone,
White’s light-squared bishop will exert influence on the light squares. 11...Nxe5 12.d6! N7g6 13.Nxf5 exf5 14.Bxc5
Rc8 15.Ba3 Rc6 16.f4± forces Black to give up a piece, or face a mating attack.
522
16...Bxd6 17.fxe5 Qb6+ 18.Kh1 Bxa3 19.bxa3 0-0 20.Rxf5 Rd8 21.Qf1± Black does not have sufficient
compensation for the missing minor piece.]
11.Nxd5 Nxe5!? (11...b5?! 12.a4!±) 12.Nxe5 Qxd5 13.Nxc4 Qxd1 14.Rfxd1 Nc6 15.Bf3

Despite the game heading towards an ending it looks like White keeps the initiative, however Black manages to
neutralize White’s threats.
15...Rd8 This move follows the guideline which says: ‘when in trouble, go for liquidation’. 16.b3 Be6 17.Bxc6+
bxc6 18.Na5 Be7 19.Rxd8+ Bxd8 20.Nxc6 c4 21.Rd1 Bf6 22.Bd4 0-0 23.Bxf6 gxf6 24.bxc4 Bxc4= Black
successfully equalizes.

9...Ne7

Black needs to complete his kingside development or risk not being able to survive the middlegame.
9...cxd4?!

523
This attempt to go for White’s e5-pawn has generally been considered dubious, as Black uses too much time and
lags in development.
10.Nxd4 (10.Qxd4?! would justify Black’s last move. 10...Bc5 11.Qc3 Rc8 12.Bb3 Ne7 13.Rd1 Qc7³ and suddenly
it is Black who is better developed and has fewer weaknesses.)
10...Nxe5

11.Qa4+ This move leads to great complications.

[11.Be2!? A standard retreat in this structure. 11...Nf6! 12.Nc3 (12.Nxf5?! Too early. White needs his central d4-
knight for the initiative on the queenside. 12...Qxd1 13.Rxd1 exf5 14.Nc3 Rd8 15.Rxd8+ Kxd8 16.Rd1+ Kc7=)
12...Bg6=

524
The position seems ripe with possibilities. We will not delve into it any further. Let’s leave that for future analysts.]
11...b5 12.Nxb5 (12.Bxb5+?? axb5 13.Qxb5+ Nd7 14.Nxf5 exf5 15.Re1 Be7–+) 12...axb5 13.Bxb5+ Ke7 14.Qb4+
Kf6 15.Qf4! h6 16.Bd4 Bd6 17.Qe3 g5! 18.Nc3 Ne7

and Black should be able to survive in this crazy position. 19.g3 Bc7 (19...g4? 20.f4 gxf3 21.Rxf3±) 20.f4 gxf4
21.gxf4 Kg6!

525
The point. White cannot recapture with his d4-bishop as it would leave him with pieces that are suddenly out of
touch with the kingside, where the black king will be safer than its white counterpart! 22.fxe5 Kh7 23.Kh1 Rg8
24.Bd3!

White must guard against possible checks along the a8–h1 diagonal.
24...Ra5! The position is unclear, for example 25.Be4 Qb8 26.Rae1 Rg6 27.b3 Kg8 28.Rg1 Bxe5 29.Bf3 Bxd4
30.Qxd4 Rxg1+ 31.Rxg1+ Bg6 32.Ne4 Nd5 33.Nf6+ Nxf6 34.Qxf6 Rf5 35.Rxg6+ fxg6 36.Qxg6+ Kf8 37.Qxh6+
Ke7 38.Qh4+= and attack, defense, counterattack and counter-defense are all equal.

10.Nc3

526
We have reached the main tabiya of this sub-line. Black has several possibilities here.

10...b5?!

Black goes for an aggressive attempt to create a favorable structure and fix White’s d4/e5 pawn chain, thus creating
a strong potential post on d5 for his minor pieces. Now it was my turn to think.
a) 10...Nc6 This alternative seems more logical. 11.Be2! This paradoxical retreat is probably White’s best choice.
The idea is that after mass exchanges in the center, White’s c4-bishop will be stuck looking at the e6-pawn, while
from e2 it can regroup to the f3-square, posing Black significant problems on the queenside. Also, White prepares a
more favorable version of his thematic d4–d5 push.
[11.d5!? The only possibility I considered during the game. This standard push and the accompanying sacrifice of
the e5-pawn are ways for White to break open the position to exploit his lead in development. Black has a number of
alternatives at his disposal. 11...Ncxe5 (11...Ndxe5? 12.Bb3! Nxf3+ 13.Qxf3 Ne5 14.Qg3 c4 15.Ba4+ b5 16.Qxe5
bxa4 17.dxe6 fxe6 18.Rad1 Qb8 19.Qd4 Bd3 20.Rxd3 cxd3 21.Qxa4+ Kf7 22.Qd7+ Be7 23.Ne4 h6 24.Bc5±;
11...exd5 12.Bxd5±) 12.Be2° White’s standard bishop retreat.

527
This position is rich in analytical possibilities. White clearly has sufficient compensation, but whether he has more,
only future analysis will tell. We’ll leave that to the reader. (Sadly, after 12.dxe6? fxe6! 13.Nxe5 Nxe5 14.Qh5+ g6
15.Qe2 Bg7 16.Rad1 Qh4!

Black seems to be fine, for example 17.f4 Ng4 18.Bb5+ axb5 19.Qxb5+ Kf8 20.Bxc5+ Kg8 21.h3 Nf6 22.Qxb7 Re8
23.Rf3! h6 24.Bf2 Qh5 and Black is looking forward to his counterattack.)]
11...Be7 (11...cxd4 is more solid, but White gets a pleasant advantage, e.g. 12.Nxd4 Nxd4 13.Qxd4 Bc5 14.Qf4
Bxe3 15.Qxe3 Qb6

16.Qxb6! Nxb6 17.Rad1 Ke7 18.Rd6 Nc8 19.Rd2 Rd8 20.Rxd8 Kxd8 21.g4 Bg6 22.h4 h6 23.f4 Ne7 24.h5 Bh7
25.Rd1+ Kc7 26.Bf3 Nc6 27.Kf2± with a very comfortable advantage for White in the resulting endgame, as he has
the superior pawn structure and better pieces.) 12.d5!

528
12...exd5 (12...Ncxe5? 13.Nxe5 Nxe5 14.f4 Nd7 15.g4 Bg6 16.d6! Bf6 17.f5 Be5 18.fxg6 hxg6 19.Bf4 Qh4 20.Bxe5
Nxe5 21.Rf2+–) 13.Nxd5 Ndxe5 14.Nxe5 Nxe5 15.b4!± Suddenly White strikes on the queenside, preventing Black
from castling. This move also allows White to get his pawn back by tactical means. 15...Nc6
[15...Bc2?! 16.Qxc2 Qxd5 17.Rfd1! Qe6 18.Qa4+!

18...Kf8 (18...Nc6 19.b5+–) 19.bxc5+– with a decisive advantage for White, since Black’s piece coordination is
completely busted, leaving White in charge thanks to his strong pair of bishops hitting Black’s weak queenside
pawns.]
16.Nxe7 Qxd1 17.Raxd1 Kxe7

529
18.g4! A computer line of course. The level of finesse required to stabilize White’s advantage is quite high. Every
move counts! 18...Be6 19.Bxc5+ Kf6 20.a4± Black’s c6-knight is a key defender, since it takes control of the crucial
d4- and e5-squares from White’s dark-squared bishop.
Once the c6-knight is gone, Black will be hard-pressed to defend both his king and b7-pawn.
b) 10...cxd4

This move is quite relevant here. White can recapture in three ways, but only one maintains his initiative. 11.Nxd4!
[11.Qxd4?! Nc6 12.Qf4 h6!? (12...Qc7 is possible as well) 13.Rad1 Qc7 14.Be2! g5 15.Qa4 Rc8 16.Rd2 Nc5
17.Bxc5 Bxc5 18.Ne4 Bxe4 19.Qxe4 Rd8! and in this sharp position, both sides have chances; 11.Bxd4? the worst
capture. 11...Nc6³]
11...Nxe5 12.Be2! Do you notice a pattern here? White invites Black to capture on e5, so that later White can go f2–
f4–f5 with tempo! 12...N5c6 13.Bf3 Nxd4 (13...Rc8? 14.Qa4 b5 15.Qxa6 Nxd4 16.Bxd4 Qxd4 17.Nxb5+–) 14.Bxd4
Nc6 15.Qa4

530
15...f6! Black really needs to get his king out of the center, and this move allows him to achieve two goals at the
same time. One, it provides the black king with a haven on the f7-square, and two, Black’s f6-pawn blunts the
activity of White’s dangerous d4-bishop on the a1–h8 diagonal. (15...Rc8 16.Rad1 Qc7 17.Rfe1 Be7 18.Bxg7 Rg8
19.Bh6 Rg6 20.Be3 Kf8 21.Bh5±) 16.Rfd1 Kf7 17.Ne2!

The game becomes very complex, with White maneuvering his pieces on to optimal squares, trying to exploit his
lead in development, while trying to get at Black’s vulnerable queenside pawns.
Back to 10...b5.

531
11.Bd3?!

While the computer does not judge this move too harshly and says that White keeps an advantage, I felt that the
improvement I discovered after the game, and used to win my game vs. Adianto the following year, was better.
11.Be2! This standard retreat is strongest, just as we have seen before! 11...b4? 12.Na4 Nd5 13.Bg5 Qb8 14.Rc1

and after just fourteen moves, Black is almost lost. The game continued 14...h6 15.Bh4 Qa7 16.dxc5 Bxc5 17.Nxc5
Nxc5 18.Qd4 Rc8 19.Rc4!

532
and Black lost his knight thanks to the cross-pin: 19...g5 20.Bg3 0-0 21.Rfc1 b3 22.axb3 Nxb3 23.Qxa7 Nxc1
24.Bf1+– and Black resigned shortly in Kamsky-Adianto, Donner Memorial 1996, my last tournament before my
early retirement which lasted until 2004!!

11...c4!?

Consistent.
11...cxd4 This seems plausible as well, though not quite consistent with Black’s expansion on the queenside a move
earlier, because it gives White too many ways to open the position before Black completes his kingside
development. 12.Bxd4 Bxd3 13.Qxd3

13...Nc6 This turns out to be an inferior position for the knight. (13...Nf5 is the most solid choice, but White’s
position remains pleasant. 14.Rfd1 Be7 15.Ne4

533
A logical continuation. 15...0-0 16.Be3 Nxe3 17.Qxe3 Qc7 18.Rac1 Qb7 19.Nd6 Bxd6 20.exd6 Rac8 21.Ne5! Nxe5
22.Qxe5 Rfd8

23.Rc5!!± An extremely important move, as it allows White to keep a significant advantage thanks to his strong
passed d6-pawn. Black will be hard-pressed to prevent his queenside pawns from falling.)
14.Rac1! (14.Rad1?! Be7 15.Ne4 Nb4 16.Qa3 Nd5 17.Nd6+ Bxd6 18.exd6 0-0² and Black survives.) 14...Nxd4
15.Nxd4! Bc5

534
16.Rfd1! The only way to fight for the advantage. (16.Nf3?! 0-0 17.Ne4 Be7 18.Rfd1 Nb6 19.Nd4!

White’s knight returns to go forward this time. 19...Qe8 20.Nc6 Rc8 21.Nd6 Bxd6 22.Qxd6 Kh8 23.Na7 Rxc1
24.Rxc1 Qb8 25.b3² with a very small advantage for White.) 16...Qb6 (16...0-0?? 17.Nc6+–; 16...Nxe5 17.Qe4 Bxd4
18.Rxd4 Qb8 19.Re1 Ng6 20.Qc6+ Kf8 21.Rd7 Kg8 22.Qf3!±) 17.Ne4

535
17...Bxd4
[17...Nxe5?! 18.Qc3 Be7 (18...Bxd4? 19.Rxd4 is disastrous for Black, as White is threatening Rd6. If 19...Nc4 then
for example 20.b3 f5 21.Ng5 h6 22.Nf3 Nd6 23.Ne5 Ne4 24.Qe3 Qb7 25.Rd7 Qxd7 26.Nxd7 Kxd7 27.Qd4++–)
19.Qg3!

19...Ng6 20.Nc6 0-0 21.Nxe7+ Nxe7 22.Nf6+ Kh8 23.Nd7+– and White should win, due to his material advantage
and superior pieces.]
18.Qxd4 Qxd4 19.Rxd4!

536
19...g5 (19...Nxe5 20.Rc5 Nd7 21.Rc7+–) 20.Nd6+ Kf8 21.Rc7!+– (21.Nxf7±) 21...Nxe5 22.f4 gxf4 23.Nxf7 Nxf7
24.Rxf4 Kg7 25.Rcxf7+ Kg8 26.R4f6 h5 27.Re7 Rc8 28.h4+–

with a winning rook endgame.

12.Bc2

Here I was comforted by the thought that the game suddenly appeared to have taken on the outlines of the Spanish
game!!
a) 12.Be4?! b4! 13.Bxa8 bxc3µ would be a clear mistake.
b) 12.Bxf5! This is much stronger! 12...Nxf5

537
13.a4!² This move should be played immediately! White needs to undermine Black’s queenside pawn chain before
he completes his piece development and gets a grip on the center. Black’s queenside pawn formation would then be
an advantage, not a defect as it is now. The following main line is quite instructive.
[13.d5!? Bc5! (13...Nxe3? Exchanging a good knight for a bad bishop is usually not a great idea in any situation.
14.fxe3 Bc5 15.dxe6 fxe6 16.Nd4 Qb6 17.Qh5+ g6 18.Qh4!

White prevents the black king from moving anywhere away from the center. The holes on d6 and f6 guarantee that
White’s attack will be irresistible.) 14.Bf4! 0-0 15.a4 Rb8

538
16.Ne4!?² (16.Ng5 Qe8∞; 16.Qe2!? White maintains the tension in the center and seems to have somewhat better
middlegame prospects. However, with precise play, Black may equalize: 16...Nd4 17.Nxd4 Bxd4 18.axb5 axb5
19.Rfd1 Qb6 20.Rd2 Bxc3 21.bxc3 b4 22.Be3 Qb5 23.dxe6 Nxe5 24.cxb4 Qxb4 25.exf7+ Rxf7 26.Bd4 Nd3=)]
13...b4 (13...Rb8 14.axb5 axb5 15.d5 Bc5 16.dxe6 fxe6 17.Bxc5 Nxc5 18.Qxd8+ Kxd8

19.Nxb5!± Boom! White is just in time. Black is much worse.) 14.Ne4

539
14...Rc8 15.Bg5 Qa5 16.g4! Nh6 17.h3 Ng8 18.Qe2 Qd5 19.Rfc1 a5 Trying to patch up the holes on the queenside
in order to protect the c4-pawn, without which Black is just lost. 20.Nfd2 Nb6 21.Rc2!

Black’s position is a disaster, for example 21...h6 (21...b3 22.Nxb3 cxb3 23.Rxc8+ Nxc8 24.Rc1 Nb6 25.Qa6+–)
22.Bf4 b3 23.Nxb3 cxb3 24.Rxc8+ Nxc8 25.Qa6 Nge7 26.Rc1 f5 27.Rxc8+ Nxc8 28.Qxc8+ Kf7

540
29.gxf5! Kg8 (29...Qxe4? 30.Qxe6#) 30.Qxe6+ Qxe6 31.fxe6+– White finally breaks through. The advantage of
having his entire army at the ready is more important than allowing Black access to the d5-square.

12...Rc8!

At this moment I was not so sure about my position and started to use a lot of time. Basically, White’s d4–d5 break
in the center becomes less effective as Black makes more developing moves. Black is also ready for any breaks on
the queenside and has very few moves left before he completes his kingside development. The question is: what to
do? Generally, when you have no plan and no idea, it is useful to ask yourself what your opponent’s strongest piece
is. In this case it should be evident that the f5-bishop is clearly Black’s best, as it defends the light squares and
protects Black from any possible kingside attack. That gave me the idea that perhaps I should NOT play on the
queenside, but in the center and on the kingside, where White has the pawn advantage, and more importantly, space
for maneuvering!

13.Re1
541
A useful waiting move to check my opponent’s intentions.
a) 13.Bg5?! This would force Black to exchange his f5-bishop, but leave White with a bad bishop on g5, taking the
square from his f3-knight. Still, this move serves a certain purpose. 13...Bxc2 14.Qxc2 Qc7 15.Rfe1 Nb6

16.a4 b4 17.a5! Nbd5 18.Bxe7 Bxe7 19.Nxd5 exd5 20.e6 f6 21.Qf5 Qd6 22.Nh4 g6 23.Qh3! Rg8 24.g3∞

and a very unbalanced position appears, where both sides can be optimistic about the possible outcome of the game,
since both possess some kind of trump: White’s strong e6-pawn completely destroys the coordination between the
black rooks, while Black has a very compact pawn structure and a protected passed pawn on c4, which, along with
White’s potentially weak a5-pawn, could be crucial in the endgame.
b) 13.Ng5!?

542
This seems to be the most logical, however Black has sufficient defensive resources. 13...Bxc2 (13...b4 14.Nce4 Nd5
15.Qh5 g6 16.Qf3 h6 17.Nxf7 Kxf7 18.g4ƒ) 14.Qxc2 Nf5 15.Nge4 Be7 16.g4!

The game becomes complex here too. Black has a choice of plausible moves that will determine the structure for the
rest of the game. 16...Nh4 (16...Nxe3 17.fxe3 b4 18.Na4 0-0 19.Rf3 The position seems roughly balanced.) 17.f4
This idea looks attractive, with the idea of playing f4–f5 next, however Black refutes it tactically. 17...h5! 18.f5
(18.h3? hxg4 19.hxg4 f5!µ) 18...hxg4 19.fxe6 fxe6 20.Qe2 Nf3+ 21.Rxf3 gxf3 22.Qxf3 Nb6! 23.Qg4 Rc6!µ.

543
Protecting the crucial e6-pawn and giving Black good chances to take over the game, thanks to his material
advantage.

13...b4!?

Black wants to deploy his e7-knight to the d5-square immediately.

13...Nb6 14.a4! b4

15.a5!? The most accurate.

[15.Ne4 Nc6 (15...Bxe4? 16.Bxe4 a5 17.Bb7±) 16.a5 Nd5 17.Bg5 Qd7 18.Ba4∞ and another unbalanced position
arises, with chances for both sides.]
15...Bxc2 16.Qxc2

544
16...Nbd5 (16...b3? 17.Qe2 Nbd5 18.Ne4 Ng6 19.Nfd2 Bb4 20.Rec1 c3 21.Nxb3±) 17.Ne4 Nf5 18.Bg5 Qd7 19.g4
Nh6 20.Bxh6 gxh6 21.Nf6+ Nxf6 22.exf6 Bd6 23.d5‚

with a huge initiative for White.

14.Ne4 Nd5∞

545
And now a more or less forced sequence starts which I evaluated as better for White.

15.Bg5 Qa5?!

A bit too optimistic.

a) 15...Qc7!

16.Qe2
a1) 16.Ba4!? h6 17.Bh4 N5b6 18.d5!

546
The point. 18...Nxa4 19.Qxa4 Qb7 (19...exd5?? 20.Nd6++–) 20.dxe6 fxe6 (20...Bxe6 21.Nd4±) 21.Nd6+ Bxd6
22.exd6 0-0 23.Be7 Rf7 24.Ne5 Nxe5 25.Rxe5 c3 26.bxc3 bxc3 27.Rc1 c2 28.h3 Qb2 29.Ree1

This position reminds me of the last game in the World Chess Challenge match vs. Topalov in 2009 where I
tragically went from totally winning to lost in the space of two moves. Had I won that game, I would have continued
to the tiebreaks, with very good chances to win the rapids. As I lost, Topalov went on to challenge Kramnik for the
unified World Champion title. I still sometimes get nightmares about that game and its error.
a2) 16.Nd6+?! Going for forcing lines doesn’t bring White any hope of an advantage. 16...Bxd6

547
17.Bxf5 (17.exd6 Bxc2–+) 17...Be7 18.Bxe7 (18.Be4 N7b6³ Black has completed his development and stands better
thanks to his superior structure.) 18...Nxe7 19.Bc2!

It is important to leave the e4-square for the f3-knight which is going to hop all the way to d6, compensating for the
inferior structure. 19...h6 20.Nd2 0-0 21.Ne4 Nb6 22.Nd6 Rcd8 23.Qe2 Ned5 24.h4 Rxd6

548
Forced, as White was prepared for a battery with Qe4 and h5, penetrating the black king’s defenses. This positional
exchange sacrifice creates a balanced game. 25.exd6 Qxd6 26.Qe4 Nf6 27.Qe5 Rd8 28.Rac1 Nfd5 29.g3= Neither
side can really make progress, as any pawn move on the queenside will open files for the white rooks.
16...Bxe4 17.Bxe4 N7b6 18.Rac1

18...h6! Not actually getting rid of the white bishop on g5, but taking the important g5-square from White’s f3-
knight later on! White must also decide on the final diagonal for the bishop.
[18...Be7 19.Bxd5 Nxd5 20.Bxe7

549
20...Nxe7 (20...Kxe7 21.Ng5 h6 22.Ne4 Rhd8 23.Nd6±) 21.Nd2 c3 22.bxc3 bxc3 23.Ne4± and Black is not in time
to save his pawn.]
19.Be3! Got to be closer to the center. (19.Bh4 a5 20.Nd2 c3 21.Nb1 g5 22.Bg3 Qc4!³) 19...Be7 20.Nd2 0-0

21.Nxc4! Nxc4 22.Bd3 Nc3 23.bxc3 Nxe3 24.fxe3 bxc3 25.Bxa6 Rb8 26.Qd3 Rb2! (26...Bb4 27.a4²) 27.Rxc3 Qa5
28.a3 Rfb8 29.Rec1 Rb1! and, although White is better, the position is very likely to liquidate into a drawish
opposite-colored bishops endgame, which would be the logical result, as neither side has made any mistakes!!
b) Note that 15...f6?? loses on the spot: 16.exf6 gxf6 17.Nxf6+ N5xf6 18.Bxf5+–, while 15...Qb6 would be similar
to the game, plus it makes no sense to deprive the black knights of the b6-square.

16.Nh4!

The beginning of an interesting concept which I felt was very unsettling psychologically for my opponent. It is not
quite clear at first where this knight is going. The first thought is that it’s heading for f4 or e3 via g2, however
White’s plan is more ambitious than that!!

550
16.Nd6+

A computer choice. 16...Bxd6 17.Bxf5 (17.exd6? Bxc2 18.Qxc2 0-0 19.Ne5 Qb5!³) 17...Be7 18.Bxe7 Nxe7 19.Bc2
h6 20.Nd2 Qd5! This must have been my opponent’s idea when he first played ...Qa5. One must give him credit, of
course, for having seen this far! 21.Ne4 Nf5 22.Qd2 0-0=.

16...Bxe4™ 17.Bxe4

17...g6!?

Played after a long think which almost took the last minutes off my opponent’s clock. Black realized that White
simply wants to push his f-pawn, to open the position up for his bishops, with sacrifices if necessary. This would
create very dangerous mating threats to the black king. Or so I thought.
a) 17...Be7 18.Qg4 Bxg5 19.Qxg5 0-0 20.Nf5! is one of the more obvious ideas for White.

551
20...exf5 21.Qxf5 g6 22.Qxd7 Rfd8 23.Qg4 c3 24.bxc3 bxc3 (24...Nxc3 25.Bf3 Rc4 26.e6±) 25.Bc2!± The bishop is
heading for the excellent b3-square, from where it’ll be pointing at everything along the a2–g8 diagonal, leading to a
big advantage for White, since Black’s only trump, the far advanced c3-pawn, will be unable to move because of the
knight being tied to d5.
b) 17...h6 18.Ng6! Rg8 19.Bh4 c3 20.b3 N7b6 21.Qg4² is another idea.
White keeps the initiative.
c) 17...c3 18.bxc3 Nxc3 19.Qf3! and Black is unable to use the blockading d5-square, since the c-file is vulnerable
for infiltration by the white rooks.

19...Nxe4 20.Qxe4 Nb6 21.Rac1! Qd5 22.Rxc8+ Nxc8 23.Qc2 Ne7 24.Rc1! h6 25.Bd2 Qb7 26.g3!

552
The knight goes to e3 or f4 to exchange Black’s d5-knight. 26...Nd5 27.Ng2 Be7 28.Qa4+! Qd7 (28...Kf8 29.Rc6±;
28...Qb5 29.Qxb5+ axb5 30.Rc8+ Bd8 31.Ne3±) 29.Qxa6 0-0 30.Ne3 Bg5 31.Rc2 Bxe3 32.fxe3 Rb8 33.Qc4 Qb7
34.Qc6² with some advantage for White.

18.Qg4!ƒ

White shows that he will have no hesitation in sacrificing pieces for the future f4–f5 break. The computer doesn’t
quite ‘see’ the idea with Nh4 and Qg4, but once it appears on the board, it actually starts to like it.

18...Bg7?!

During the game I was very impressed with my opponent’s ability, even when very short of time, to place his pieces
on protected squares and have them protecting each other, although ...Bg7 obviously serves no function other than to
prepare for short castling.
a) 18...Qb5!? 19.f4 Be7!
553
It would be very hard for Black to make this move, voluntarily giving up control over the dark squares, but it is the
best. 20.Nf3
[20.f5?! doesn’t work: 20...gxf5 21.Nxf5 exf5 22.Bxf5 h5 23.Qh4 Bxg5

24.e6!! Unfortunately this inspired break is only good enough for equality. (24.Qxg5 Ne7–+) 24...Bxh4 25.exd7+
Kf8 26.dxc8=Q+ Kg7 27.Qc5 Bxe1 28.Rxe1 Re8! An important intermediate move. White suffers, but gets a draw.
29.Qxb5 Rxe1+ 30.Kf2 axb5 31.Kxe1 Ne3 32.Be4 Kf6 33.Kd2 Nf5 34.a3! bxa3 35.bxa3 Nxd4

554
36.Bd5! Nf5 37.Bc6 Nd6 38.Kc3 Ke5 39.g3 f5 40.a4 bxa4 41.Bxa4 Kd5 42.Bd1 Ne4+ 43.Kc2 h4 44.gxh4 Ke5
45.h5 Kf6 46.Be2=]
20...h6 21.Bxe7 Nxe7 22.Qh4 Kf8

23.Rac1 (23.a3? b3 24.Rac1 Nb6 and the a3-move only created unnecessary problems for White as the b2 pawn is
weak.) 23...Nb6 24.Qf6 Rh7 25.Qh4 Nbd5 Arguably Black has the easier game.
b) 18...N7b6

555
19.Nf3! White has a simple plan: he wants to invade via the dark squares on the kingside and hit Black’s weak f7-
and h7-pawns, now that Black has created an additional dark-squared weakness on f6.
[19.f4?! was my plan, however it isn’t optimal: 19...Be7! is simple and strong. (19...Rc7?! During the game I
thought this was best, however 20.f5! gxf5 21.Bxf5

21...Kd7! 22.Be4 h6 23.Bc1 Kc8 24.Rf1 Kb8 25.Qe2 Be7 26.Nf3 Qb5 27.Rf2= and we have reached yet another
complicated, but probably balanced, position. Both sides have pluses and minuses which seem to cancel each other
out.) 20.f5 gxf5 21.Nxf5 exf5 22.Qxf5 Bxg5 23.Qxg5 Ne7! Black defends, and, with his material advantage, is the
only one who can play for a win.]
19...Bg7 20.Qh3

556
Black’s position is dangerous, as it is very simple to make a mistake. 20...c3 (20...0-0?! 21.Bh6, with the idea of
Ng5, creates a dangerous attack. 21...Rc7 22.Ng5 Nf4 23.Qh4 Bxh6 24.Qxh6 f5 25.exf6 Rxf6 26.h4‚ ; 20...h6?
21.Bxg6 Kf8 22.Be3 fxg6 23.Qxe6 g5 24.Bxg5 hxg5 25.Nxg5 Rc7

26.Re3! Nxe3 27.fxe3 Re7 28.Rf1+ Ke8 29.Qc6+ Rd7 30.Nf7! Rf8 31.e6 Qd5 32.exd7+ Qxd7 33.Qxb6 Rxf7
34.Qb8+ Ke7 35.Qxb4+ Ke8 36.Rxf7 Qxf7 37.Qb8+ Ke7 38.Qa7+ Kf8 39.Qxf7+ Kxf7 40.Kf2 Ke6 41.Kf3+–

557
and this endgame should be won for White.) 21.bxc3 Nxc3 22.Bc2 Nbd5 23.Bh6 Kf8!

This is the best defense. Black keeps his rook on h8 to protect his kingside. It is not so easy for White to break
through Black’s defense.

19.f4!

558
Going for it!! At this moment I felt great! I had more time on the clock and a dangerous initiative. This is what it
must have been like for Kasparov in his famous Spanish Zaitsev games vs. Karpov!!

19...Qb6!

Karpov has always been considered one of the greatest defenders of his day.
In modern times, the title of Minister of Defense belongs to Karjakin after his World Championship match vs.
Carlsen, where he successfully managed to defend a number of lost positions.
19...0-0? 20.f5! Qb6

21.Kh1! and Black cannot defend all his weaknesses.

[21.fxe6? It’s never too late to make a mistake. 21...Qxd4+ 22.Kh1 Nxe5 23.exf7+ Nxf7 24.Nf3 (24.Rad1 Nf6!µ)
24...Qc5 25.Rad1 Nf6 26.Bxf6 Bxf6 27.Rd5 Qf2 28.Rd2 Qa7 (28...Qb6? 29.Bxg6+–) 29.Rd7 Rc7 30.Red1=]
21...Qxd4™ (21...exf5 22.Nxf5 Qe6 23.Rf1! preparing to take on d5. 23...N7b6 24.Bh6!+– Black suffers decisive
material losses.) 22.fxg6!
559
(22.fxe6 Nxe5 23.exf7+ Nxf7 24.Rad1 Nf6µ) 22...Nxe5
[22...hxg6 23.Nxg6 Nxe5 24.Nxe5 f5 (24...Qxe5 25.Bh7+ Kxh7 26.Rxe5 Bxe5 27.Qe4++–) 25.Bxf5 Qxg4
26.Bxg4+–]
23.gxf7+ Nxf7 (23...Rxf7 24.Bxh7+ Kxh7 25.Qxd4+–) 24.Nf3! (Even stronger than 24.Qxe6) 24...Qxb2 (24...Qb6
25.Rad1+–) 25.Rab1 Qc3 26.Qxe6 Nf6 27.Bd2 Rce8

28.Bxh7+! Kh8 and White wins in many ways, for example 29.Qh3 Ng5 30.Rxe8 Nxe8 31.Bxg5+–.

20.Kh1!

560
Played in the the spirit of the position in accordance with the famous maxim that the threat (in this case of breaking
with f4–f5) is stronger than its execution.

20...Qc6?

And Black makes a mistake.

20...f5!? This was the only move to maintain the balance. Of course, it’s NOT at all easy (in fact probably
unimaginable) to make such a move with seconds on the clock, opening the e-file against your king for the enemy
rook on e1. 21.exf6™ (21.Nxf5? exf5 22.Bxf5 gxf5 23.Qh5+ Qg6–+) 21...N7xf6!? Seems like a losing move, but in
reality is best.
[21...N5xf6 22.Qh3! White forces Black to capture the e4-bishop, thus getting rid of Black’s super-strong d5-knight.
(22.Bxg6+? hxg6! 23.Rxe6+ Kf7 24.Bxf6 Nxf6! 25.Qxg6+ Kxe6∞) 22...Nxe4 And Black in turn needs to get rid of
White’s super-strong lightsquared bishop. (22...0-0? 23.Bc2! c3 24.Bb3! and White is just in time to conclude his
attack. 24...cxb2? 25.Bxe6+ Kh8?! 26.Nxg6#) 23.Rxe4 0-0! 24.Rxe6 Qxd4 25.f5 Nc5! 26.fxg6

26...h5! Black’s only move to keep the game balanced. (26...Nxe6?? loses on the spot: 27.gxh7+ Kxh7 28.Nf3++–;
561
26...Qd3?! 27.gxh7+ Kxh7 28.Nf3+ Kg8 29.Re7‚ ) 27.Nf5 Qg4 28.Nxg7 Qxh3 (28...Kxg7? 29.Re7+ Kxg6 30.Qe3‚ )
29.gxh3 Kxg7 30.Rb6 c3= and after all the excitement we have an equal endgame.]
22.Bxd5! Nxg4 23.Rxe6+ Qxe6 24.Bxe6

24...h6! (24...Nf2+? 25.Kg1 Bxd4 26.Bxc8 Nd3+ 27.Kf1 h6 28.Nxg6 Rg8 29.Ne7 Rf8 30.Bxh6+–) 25.Bxg4 hxg5
[25...Kf7!? 26.f5 hxg5 (26...gxf5? 27.Bh5++–) 27.fxg6+ Ke7 28.Re1+ Kd8

29.Nf3! (29.Bxc8 Rxh4„) 29...Rc6 30.Nxg5 Rxg6 31.Ne6+ Rxe6 (31...Kc8? 32.Bf5+–) 32.Bxe6 Re8! 33.Re2 c3!
(33...Bxd4 34.Bxc4±) 34.bxc3 bxc3 35.d5 Bh6 36.Rc2 Bd2 37.g4 Ke7 38.h4 Kd6 39.Kg2 Rb8 40.g5 Ke5 41.Bf7

562
41...Rb4 (41...Rb2? 42.Rxb2 cxb2 43.Bg6+–) 42.Kh3 Rd4 43.g6 Kf6 44.h5 a5² and Black is very close to the draw.]
26.Nxg6 Rd8

27.Nxh8! A much stronger option.

[27.fxg5?! Rxd4 28.Re1+ Kd8 29.Nxh8 Rxg4! 30.Nf7+ Kd7 31.h3 Rf4 32.Ne5+ Kd6 33.g3! Rd4 34.Ng4!?

563
34...Rxg4!! 35.hxg4 Bxb2! A fantastic position. White has an extra exchange and pawn (!!), but his tripled g-pawns
provide Black with strong counterplay. 36.Rb1 c3 37.Re1 c2 38.g6 a5 39.g7 Bxg7 40.Rc1 a4 41.Rxc2 b3 42.axb3
axb3 43.Rc1 b2 44.Rf1 Bh6 45.Rb1 Bc1=

With an easy draw.]

27...Bxh8 28.fxg5 Rxd4 29.Re1+ Kf8 30.Be6±

564
Despite reduced material and the presence of opposite-colored bishops, White’s material advantage is an important
factor which should be decisive.

21.f5ƒ

Of course.

21...N7b6!

The best move under the circumstances.

21...gxf5 22.Nxf5 exf5 23.Bxf5 h5 24.e6!!

565
Probably the move my opponent missed. 24...N7f6 (24...hxg4 25.exd7++–; 24...Ne5 25.exf7+ Kxf7 26.Qe4+– Rce8
27.dxe5 Rxe5 28.Qf3+–) 25.exf7+ Kf8 26.Qf3 Kxf7 27.Re5!

An important centralizing move. White needs his last piece, the a1-rook, to join the attack. Black is defenseless.
27...Qd6 28.Rae1 Rce8 29.Bxf6 Bxf6 (29...Nxf6 30.Qb7+ Kf8 31.Bg6+–) 30.Rxe8+–

22.Bf6?!

An inaccuracy which, thankfully, doesn’t spoil all the advantages of White’s position.
a) 22.fxe6? Qxe6! (22...fxe6? 23.Rf1±) 23.Nf5 gxf5 24.Bxf5 h5! 25.Qf3 Qc6 26.e6 f6µ
b) 22.fxg6! hxg6 23.Rf1!! with the idea of doubling on the f-file, would have given White a decisive advantage, as
the black king can no longer escape to the queenside.

566
23...Rc7 24.Rf3 c3 The only sensible option left to Black. (24...Rxh4? The exchange sacrifice doesn’t work as the
white queen gets access to Black’s kingside via the h7-square. 25.Qxh4+–) 25.bxc3 bxc3 26.Raf1 c2 (26...0-0 loses
to 27.Bf6 c2 28.Nxg6 fxg6 29.Qxg6+– with a mating attack; 26...Nc4 27.Rxc3+–) 27.Rc1!±

More space means more maneuvering opportunities for White who suddenly targets the c2-pawn which has gone a
little too far. Black cannot defend against all the threats as his pieces are pinned down to various defensive tasks, for
example 27...0-0 28.Bf6 Qe8 29.Nxg6! fxg6 30.Bxg6 Qc8 31.Be4 Qe8 32.Qh4 Nxf6 33.exf6 Rxf6 34.Qh7+ Kf8
35.Rxf6+ Bxf6 36.Qxc7+–.

567
22...Bxf6!

With seconds left, my opponent finds the best defense.

a) 22...exf5 23.Nxf5 Bxf6 24.exf6 gxf5 (24...0-0 25.Ne7++–) 25.Bxd5+ Kd8 26.Qg7+–
b) 22...Rg8 23.Qg5!‚ (23.fxg6 hxg6 24.Bxg6 Bxf6 25.exf6 transposes into the game.)

23.exf6 Kd8!

Again the best move. Black finds the way to move his king to safety on the queenside. In return, he leaves his entire
kingside for White to ravage.

24.fxg6?

But this is bad. Opening the h-file is not good, under any circumstances.
a) 24.fxe6? I didn’t like this move in view of the following line: 24...Qxe6™ (24...fxe6?! 25.f7 …Nf3–e5) 25.Qg3
Qxf6 26.Nf3 Re8 27.Ne5 Rc7„ and Black is going to fight for a win now.
b) 24.Bxd5!!

568
This move didn’t enter my thought process at all. White gives up his great bishop, for what! The point is that White
actually now threatens to get to the black king along the open e-file. 24...exd5 The only move that doesn’t lose right
away. (24...Nxd5? 25.fxe6 Kc7 26.Nf3! Kb7 27.exf7 Rhf8 28.Re6 Nxf6 29.Qh3 Ne4 30.Rf1 Qd5 31.Re5+–)
25.fxg6! hxg6 26.Nxg6! The point. White now gets to use his distant passed f6-pawn as a battering ram.

26...fxg6 (26...Rg8 27.Qe2! Qxf6 28.Qe7+ Qxe7 29.Nxe7± While this looks tricky because Black has a very solid
pawn chain on the queenside, White’s material advantage — a clear exchange — is too great. 29...Rg4 30.Nxc8
Kxc8 31.Rad1 Na4 32.h3 Rh4 33.Re2 c3 34.bxc3 Nxc3 35.Rc2 Kd7 36.Rd3+– and White should be able to convert
his big material advantage.) 27.Re6

569
27...Qd7 (27...Qc7 28.h3 c3 29.Rae1 cxb2 30.f7 Qxf7 31.Qg5+ Kc7 32.Re7++–) 28.f7 White has to be precise here;
one wrong move and he’ll be on the losing side. (28.Rae1? Rc6–+) 28...Qxf7 29.Qg5+ Kd7 30.Rxb6 Rhf8 31.h3
Rc6 (31...Rb8 32.Rxa6+–) 32.Rb7+ Rc7 33.Rxb4±

with an extra pawn and superior attacking chances.

24...hxg6

570
25.Bxg6?!

At this point I understood that I had lost my entire advantage and was hoping for a tactical trick before Black
managed to complete his king run to the queenside. Fortunately the text move doesn’t lose.
25.Nf3! immediately would have been better, as White keeps some tricks up his sleeve. 25...Rc7 26.Bxg6

26...Nxf6 (26...fxg6 27.Rxe6 Qd7 28.Rae1 Rc6 29.Ng5 is much more dangerous, e.g. 29...Nxf6? 30.Qf4 Nfd5
31.Qb8+ Rc8 32.Qe5+–) 27.Qg5 Nbd5 28.Ne5 Qe8=.

25...Rc7

Clearly Black was seriously concerned about making the time control, and had no time to calculate complicated
lines, hence the move in the game. He is still fine though!
25...Rg8! This move, however, would have made my life quite complicated as Black gets to attack this time.
26.Qf3
571
Now Black has two good options. 26...Rc7 (26...Rxg6 27.Nxg6 fxg6 28.Qg3 Nxf6 29.Qxg6 Nbd5 30.Rf1 Nd7
31.Rae1 e5 32.Qg5+ Kc7 33.dxe5 Qe6³ With White’s e-pawn firmly blocked, Black’s king gets to safety on the
queenside and his position has improved as his knights and passed c-pawn can become stronger than White’s rooks.
White might equalize with computer accuracy, but it’s going to be difficult for a human.) 27.Be4 Qd6 28.Rac1 Kc8
29.g3 Kb8 30.Qf2 Qd8 31.Rf1 Rh8³ Black has a much easier game, as the white f6-pawn will fall, and the black
knights are dominating, while White’s h4-knight is far from the battle. White can probably hold this position, but it
will be an uphill struggle, especially after having such a big advantage earlier in the game.

26.Be4 Kc8 27.Nf3!=

During the game I thought I must be better here, however in reality the position is close to equal.

27...Nxf6?!

Black goes for the counter-offensive in a position where he had almost reached comfortable safety.

572
Objectively the move is not bad, however in practice, with only seconds to make thirteen moves in an open position
with lots of checks, it is not to be recommended.
a) 27...Kb8? 28.Ne5 Qe8 29.Qg3! Ka7 30.b3 c3 31.a3± White successfully undermines Black’s queenside pawn
chain, opening files for his rooks and again making Black’s king vulnerable to a potential white attack.
b) 27...Qe8! A very difficult move to find in time trouble. Black prepares to take on f6. 28.Qg5 (28.Ne5 Nxf6=)
28...Qd8 29.Bxd5 Nxd5

Now for example 30.a3 b3 31.Rac1 Qxf6 32.Qxf6 Nxf6 33.Ne5 Rh4 34.Rxc4 Rxc4 35.Nxc4 Rxd4 36.Na5 Rd2
37.Rb1 Rd3∞

and the resulting endgame is not clear. White has an outside passed pawn on the kingside, but his knight is
awkwardly placed, while Black has solid coordination among his pieces. In a way, this position is reminiscent of the
earlier game in this book vs. GM Hulak, where an endgame with similar piece distribution was reached, equally
difficult to handle over the board.

28.Qg7 Nxe4?
573
A terrible choice. Black had to take on h2, of course. I was quite surprised that my opponent left my h-pawn intact.
28...Rxh2+ 29.Kxh2 Nxe4 30.Qf8+ Kb7 31.Qxb4 Ka7 32.Kg1 f5

With the open h-file now in Black’s possession, White has to be really careful not to get mated. 33.Nd2 Probably
forced. The e4-knight cannot be tolerated. 33...Nd5 34.Qa5 Nxd2 35.Qxd2 Rg7 36.Re2 Rg4 37.Rf1 Qb6 (37...f4?
38.Rff2±) 38.Qe1! (38.Rd1? Nf4µ; 38.Ref2) 38...Nf4 39.Rxf4 Rxf4 40.Rxe6 Qxd4+ 41.Qe3=

29.Qxh8+ Kb7 30.Qh4!±

White is just an exchange up.

30...Nd6!

Tougher than 30...f5, when Black creates an additional target for the white rooks, the e6-pawn.
30...f5?! 31.Ng5 Nxg5 32.Qxg5 Nd5 33.Qh6 Re7 34.Re2 c3 35.bxc3 bxc3 36.Rae1+–

574
31.Rac1 a5 32.Qg5 Nd5 33.h4!?

I couldn’t find a way to break the black fortress in the center and on the queenside, so it became logical to use the
outside passed pawn.
33.Ne5 Qb6 34.Qd8 Nf5! 35.Re4 c3 36.bxc3 Rxc3 37.Qxb6+ Kxb6² and White’s advantage is rapidly diminishing
while the strength of Black’s passed pawns on the queenside is increasing proportionally.

33...Rc8?

A logical idea: Black wants to stop the pawn with his rook, however the rook will be badly placed on h8.
a) 33...f6 34.Qg8 (34.Qg6? Nf5=) 34...Nf4 35.Qg4! Nd3 36.Qxe6 White keeps the advantage.
b) 33...Nf5 34.Ne5 Qb5

35.a4! (35.Qg8 Nf6 36.Qd8 isn’t as good: 36...Ne4! 37.a4! Nfg3+ 38.Kh2 Qd5 39.Qxd5+ exd5 40.Ng4 f5 41.Nf6!
Kc6 42.Nxe4 Nxe4 43.Rf1 Rh7 44.g3 Rh5

575
and I’m not sure if this position can be won for White.) 35...Qa6 (35...Qxa4 36.Rxc4+–) 36.Qg8 Nf6 37.Qf8 Qd6
38.Qxd6 Nxd6 39.g4+–

34.h5 Rh8 35.Kg1±

Now Black faces an unpleasant position in horrible time trouble. He has to check for threats everywhere, including
the run of White’s h-pawn.

35...Rh7

35...Nf5!? trying to keep some sort of control over the g7- and h6-squares seems best. 36.Ne5! (36.g4? f6!=;
36.Nh4!? Getting rid of this knight is another good idea: 36...Nxd4? 37.Qg7 Qc5 38.Qxh8 Nf3+ 39.Kh1 Nxh4
40.Qg8+–.) 36...Qb6 37.Rxc4 f6 38.Qg6 fxe5 39.Rxe5 Nxd4 40.Qf7+ Ka6

41.Qf2!! An important trick! 41...Nf5 42.Qxb6+ Nxb6 43.Rc6 Ng7 44.g4 Kb7 45.Rc1! Another important detail: the
black king must be kept cut off from the kingside along the c-file. White’s plan is to bring his king up to support his
connected passed pawns on the kingside, along with the mighty e5-rook. The position is technically won for White.
576
36.h6?

Yet another blunder, as I was running short of time too, yet it was not easy to see that Black’s position is holding
because of his d5-knight, and I had to play 36.Re5 with the idea of being able to sacrifice it for the knight.
36.Re5!

The threat is actually to play Qg5–g8. 36...a4 37.Qg8 Nf6 38.Qf8! and Black is over-extended. 38...Nd5 39.Ng5
Qb6™ ( 39...Rxh5 40.Nxf7+–) 40.Rxd5 exd5 41.Qe7+ Ka6 42.Nxh7 Qxd4+ 43.Kh2! Qf4+ (43...Ne4 44.Rf1+–)
44.g3 Qf2+ 45.Kh3 Ne4 46.Qe5 f6 47.Qc7 Qf5+

48.Kh4!! Ng5 49.g4!! The last two moves were obviously found by the computer. I would never have dared to enter
such a position. However, it does seem to work, as the black position is lost. 49...Qf2+ 50.Qg3 Qxg3+ 51.Kxg3
Nxh7 52.Rd1 Ng5 53.Rxd5+–

36...Qc7?!

577
In this unbalanced position, it is very difficult to make the right decisions, especially with only seconds on the clock.
Black had to transfer his queen to h8!!
36...Qe8 37.Nd2 Qh8! 38.Qe5 Qf8 39.Nxc4 Nxc4 40.Rxc4 Qxh6 41.Qd6 Qh1+ 42.Kf2 Qh4+ 43.Ke2

Now it’s the white king’s turn to run to the queenside!! 43...Qg4+ 44.Kd2 Qg5+ 45.Kc2 Qf5+ 46.Kc1 Qf4+! The
best defense; it will take quite a while to break the resulting fortress with Black’s d5-knight. 47.Qxf4 Nxf4 48.Rf1
Nd5 49.Kd2 Rg7 50.g3 Kb6 51.Rf3±

37.Nd2!

37...Qb6?

37...c3 38.bxc3 bxc3 39.Ne4 Nf5 40.Nxc3 Nxc3 41.Qd2 Rxh6

578
42.Qxc3! The only move that wins. I must admit I didn’t see this far. (42.Rxc3 Qh2+=) 42...Qh2+ 43.Kf2 Qf4+
44.Ke2! and White should win.

38.Qg4?

And fortunately for me Black lost on time in this very complicated position!! Flags were hanging, and I didn’t see
that 38.Nxc4 was winning. The game might have continued in a number of ways from here, and it’s unclear whether
White can win, so I’ll give some analysis here and leave the greater part for the more dedicated fans to discover on
their own!! The lesson here is don’t get into time trouble!!!
38.Nxc4! Qxd4+ 39.Ne3 Nf5

40.Qd8!! Probably the move we both missed. White wins instantly.

38...Rxh6

a) 38...c3!? is very interesting as well: 39.bxc3 Nxc3

579
[39...Rxh6 40.cxb4 Rg6 41.bxa5 Qb4 (41...Qa7 42.Rb1+ Ka8 43.Qh4+–) 42.Qe2 Qxd4+ 43.Qf2 Qxf2+ 44.Kxf2 Nf4
45.Rb1+ Ka7 46.Red1!± and the position should be winning, with major practical problems of course!]
40.Rxc3!?
[40.a3! Rxh6 41.Qf4!

41...Rh8 (41...Nf5?? 42.Rxc3 bxc3 43.Rb1+–) 42.axb4 axb4 43.Rxc3 bxc3 44.Rb1 Nb5 45.Nc4 c2 46.Qe4+
(46.Rxb5 Qxb5 47.Nd6+ Kb6 48.Nxb5 Rc8!!=

and, amazing as it seems, the game is equal, since the queen must keep covering Black’s c2-pawn.) 46...Kc7
47.Nxb6 cxb1=Q+ 48.Qxb1 Kxb6 49.Qf1! Nd6 50.d5 Re8 51.Qf2+ Kb5! 52.Qd4!± and this fantastic position is
probably closer to a draw than a win for White.]
40...bxc3 41.Rb1 cxd2

580
Objectively this is equal. 42.Qf3+ Kc7 43.Rxb6 Kxb6 44.Qb3+ Ka7 45.d5! The last chance for White. 45...Nc4
46.d6 Rxh6 47.d7 Rh8

and Black holds the draw.

b) 38...Qc7!

581
would have made my life complicated as I would been tempted to repeat moves and then go for the correct plan.
39.Nxc4!? This move also leads to White’s advantage, but it was, of course, impossible for us to see at the board.
39...Nxc4 40.b3 Rxh6

41.Re5!! Key move! 41...f5 42.Qg5 Rh1+ (42...Qh7 43.Rxc4+–) 43.Kxh1 Qh7+ 44.Kg1 Nxe5 45.dxe5 Nc3 46.Qe3
Qd7

582
47.a3!± and White has great winning chances. (47.Rxc3?! bxc3 48.Qxc3 Qd5=)

1–0

583
 GAME 19
Gata Kamsky
Curt Hansen
Groningen 1995
In December 1995 I was invited to take part in the annual Groningen chess festival. The tournament was quite
strong; Anatoly Karpov won and I shared 2nd-3rd places with Ivan Sokolov. Peter Svidler, Michael Adams and Joel
Lautier were other young stars taking part. My game vs. Karpov was a huge rollercoaster in yet another Caro-Kann
Advance Variation which ended in a dramatic draw, but the game I present below is the one that remains forever in
my memory. It was played in my classical style, with a long maneuvering period climaxing in a sudden tactical
skirmish, which, just as suddenly, transposed into a technical endgame. The French Defense has been known to be
quite an obstacle for some players, including Bobby Fischer. In my pre-retirement phase, I avoided the topical
forcing lines, preferring to steer the game towards a strategic handling of the resulting structures. It would be after
my return from retirement that I was to get the idea of employing it as a surprise weapon in my special Candidates
match vs. Topalov in 2009. However it was a disaster, and after some further tries with it in important games
(including my game vs. Karjakin in the 2009 Grand Prix which I managed to misplay from winning to losing in a
matter of two moves) it became evident that I should stick to the white side of the opening!

1.e4 e6 2.d4 d5 3.Nc3

This move was taught to me in my childhood by my first official teacher, the unforgettable Vladimir Grigorievich
Zak, who also taught former World Champion Boris Spassky and many times former World Championship
challenger Viktor Korchnoi. I first came to the Pioneers’ Palace after my father decided that chess was to be my
future profession. Given my father’s demanding approach to teaching, none of the coaches would take me. However
I will be eternally grateful for Mr. Zak’s gentle and diplomatic touch. He told my father that he would play a match
of 100 games against me before deciding whether to accept me as a student, although we probably played no more
than ten games before he did so. His great knowledge of how to approach a child and engage with him during a
game, both on chess and non-chess-related subjects, was a doorway to something I had never imagined before. Mr.
Zak gifted me some children’s chess books which I spent countless hours reading and re-reading, slowly getting into
the game with increasing interest, as it provided me with an escape from my normal life, though not for long. To
return to the topic at hand, I still play 3.Nc3 as my main move in the French in memory of my first teacher.

3...Bb4

584
The older main line, which made a brief comeback after I used it in the 2009 Candidates. However the lines are of
such a forced nature that one must have a prodigious memory and be prepared to face new, computer-assisted
variations, a rather dreadful prospect.

4.e5 c5

Black sometimes delays the ...c7–c5 push in these positions, and goes for development with ...b7–b6 and ...Bc8–a6,
however White’s spatial advantage is significant and the absence of Black’s dark-squared bishop means that his
kingside will be under constant pressure.

5.a3 Bxc3+ 6.bxc3 Qc7

The main line continues after 6...Ne7. This older move targets the c3-square and white e5-pawn without declaring
the position of the g8-knight.

7.Qg4

7...f5

The point. Black protects his g7-pawn and at the same time blocks the d3-h7 diagonal for White’s f1-bishop if it
comes to d3. 7...f6 is a more modern approach, as the position becomes dynamic with lots of possibilities for
improvement.

8.Qh5+

8.Qg3 is considered the main move. The theoretical dispute around whether provoking Black’s ...g7–g6 is good for
White or a tempo gain for Black has been going on for a long time. There is no clear verdict, however 8.Qg3
remains the far more common move. My personal view is that the black g6-pawn will require constant attention, as
will the one on h6, as Black will have to play ...h7−h6 to prevent White’s c1-bishop from infiltrating on the dark
squares. The way I’m going to annotate this game is not by providing detailed analysis, but rather by focusing on
what I felt was important, such as the general ideas and motifs.

8...g6

8...Qf7 is possible, but not played, because White simply returns with 9.Qd1 and the black queen has nothing better
585
than to go back to c7, thus giving White a tempo.

9.Qd1

9...Bd7

Black has a choice of where to develop his light-squared bishop. He can play 9...b6 with the idea of playing ... Bc8–
a6, trying to exchange the white f1-bishop and divesting White of the advantage of the bishop pair. The second plan
is the move in the game. Black intends to lodge his bishop on a4, generally an ideal location. From there it exerts
pressure on the white c2-pawn and blocks the advance of White’s a3-pawn which would vacate the a3-square for the
c1-bishop. (From a3 it would keep an eye on a lot of dark squares in the center.) Partly for those reasons, White
sometimes sacrifices the pawn on a5, so that after ...bxa5, there is less support for the c5-pawn.
9...cxd4 10.cxd4 Qc3+?!

This pawn grab is considered dubious since it gives White too much advantage in development. The further capture
586
of the d4-pawn potentially risks the center being opened with the follow-up c2–c4 break. It is quite possible that
with modern day computer analysis, Black can hold this position, but let’s leave that to the theoreticians. 11.Bd2
Qxd4 12.Nf3 Qe4+ 13.Be3 Qa4 14.Rb1° is one possible line.

10.Nf3

10.a4, preventing the d7-bishop from landing on the a4-square, is also possible and has been played before. 10...Nc6
11.Rb1! and surprisingly there are very few games in the database here. Obviously the game is just starting to get
exciting and ready for further analysis. (11.Nf3?! cxd4 12.cxd4 Nb4! The point; White cannot comfortably protect
his c2-pawn without giving something up in return.)

10...Ba4

11.Bd2!?N

A new move at the time. Black’s idea is fully revealed if White plays 11.Bd3, so I intended to cover the c3-square
before continuing development with that move. The computer, however, thinks very highly of 11.Rb1, which has
been played before, and it might be correct, but my move had the advantage that it took my opponent out of any
possible preparation.
a) 11.Bd3?! cxd4 12.cxd4 Qc3+ 13.Bd2 Qxd3 14.cxd3 Bxd1 15.Kxd1

587
15...h6!= An important little move that covers Black’s weak dark-squares on the kingside. Black lacks space, but has
a superior pawn structure thanks to White’s doubled d-pawns. While White has good dynamic chances here to pose
concrete problems, if Black manages to get the ...g6–g5 break in, he will have long-term winning chances due to the
defect of White’s doubled pawns.
b) 11.Rb1!? Nd7 12.Be2!² White prepares the c3–c4 break, after short castling, and if Black plays ...c5–c4, then
White saves a tempo compared to if he had played Bd3.

11...Nd7

The standard development of the b8-knight in such positions. Black also retains the option of moving his king to the
queenside, where it would be safe thanks to the closed nature of the position. This is actually the recommended way
of doing things. White usually needs to keep his rook on the h-file to prevent Black from opening the kingside by
sacrificing the g-pawn via ...g6–g5 breaks.
11...cxd4?!

In general, if Black does not get immediate advantages from such an exchange, he lands up in a worse position, as
588
undoubling White’s c-pawns also opens the diagonal and access to the b4-square for White’s dark-squared bishop.
12.cxd4 Bxc2 This pawn is untouchable. (12...Qxc2?? loses at once 13.Bb5+!+–) 13.Rc1 Bxd1 14.Rxc7 Bxf3
15.gxf3 Nc6 16.Bb5 Rb8 17.Rxb7 Rxb7 18.Bxc6+ Rd7 19.Ke2 Ne7

20.Bb5! Nc8 21.Rc1 Nb6 22.Rc6+–.

12.Bd3

12...c4?!

This move helps White because it fixes the pawn structure, and the battle plans for both sides become clear.
Theoretically, Black needs to close the position when facing the enemy bishop pair, however, in this case, the c4-
square might be well utilized by the d7-knight, which would severely limit White’s space and ability to maneuver.
12...h6!? 13.h4 0-0-0 would be played later with more success for Black.

589
13.Be2 h6

A standard move to cover the dark holes and prepare for ...g6–g5.

14.h4=

And an equally standard response by White.

This position needs certain explanation. At the moment, some of White’s pieces, including the f3-knight and e2-
bishop, are not well placed at all. They need to be repositioned, so the question becomes: what is the ideal set-up for
White’s pieces? The answer is that, with almost the entire board blocked, White has to place his chances in a
successful g2–g4 pawn break which will open the position on the kingside. At the same time, he has to watch for
Black’s counter-breaks, ...g6–g5 if he castles short and ...b5–b4 if he castles long. Therefore, the optimal place for
the white king is in the center! The knight needs easy access to the f4-square, and that means that h3 is the right spot
for it. That leaves the question of how to place White’s heavy pieces. If the a1-rook leaves, then the a3-pawn will be
without protection, so the solution comes naturally. The white queen belongs on c1, protecting the a3-pawn and
creating a battery along the c1-h6 diagonal, forcing Black to keep an eye on the defense of his h6-pawn. One of the
white rooks should be on the h-file to meet its counterpart in case of Black’s ...g6–g5 break, and the other belongs on
g1 to help with the g2–g4 break. Now the game revolves around White attempting to do all of the above, while
preventing possible black pawn breaks. But having said and done all of this, we must point out that the whole white
plan is an illusion and the position, objectively, is equal.

14...0-0-0 15.Qc1 Nb6 16.Rh3!?

It seems that the rook takes this square away from the knight, however I wanted to throw in a mix of plans to test my
opponent for defensive alertness and possibly lull him with long maneuvering before getting ready to proceed. It is
also important that on h3 the rook protects the c3-pawn, relieving the d2-bishop of guard duty to regroup to the
queenside.
16.Ng1 is of course a more direct approach. 16...Kb8 17.Nh3 Qg7

590
18.g4 (18.Rg1 Ne7 19.g4 fxg4 20.Bxg4 Bd7 21.Nf4 Qf7 22.a4

22...Na8!= The knight is transferred to an ideal defensive spot on c7, from where it can hop to e8, covering the hole
on d6, and then via g7 to f5. The position is dynamic, but balanced.) 18...fxg4 19.Bxg4 Qf7 20.Nf4 Bd7 21.h5 g5
22.Ng6 Rh7

591
23.f4 gxf4 24.Rf1 Ne7 25.Rxf4 Qg7 26.Rf6 Ng8 27.Rf2 Ne7= and it becomes obvious that White is not ready for
such expansion, as the a1-rook serves no role.

16...Be8

16...Qg7! was more accurate as the whole Rh3–g3 maneuver now loses its appeal.

17.Rb1?!

A loss of time, but of no risk to White, since the position cannot be changed by force in, say, the next forty moves.
17.Nh2 Qg7 18.g4 Ne7 19.f3 g5³ Black successfully blocks the position.

17...Na4

Black has improved his b6-knight, or so he thinks. The knight serves the important function of keeping the white
forces tied to the protection of the c3-pawn, however it would not be easy for Black to bring this knight rapidly back
592
into the thick of things on the kingside, should it become open.
17...Qg7 Again, a more standard approach for Black, attempting to complete his development by ...Ng8–e7.

18.Kf1!

White starts his whole sequence. At first it seems that perhaps he wants to move his king to h1, bring the a1-rook to
g1 and then simply push g2–g4.

18...Qg7 19.Ng1

19.Kg1?! If White follows through with his plan of putting the king on h1, the following might happen: 19...Ne7

20.g3
[20.Kh1?! White has to pay attention. A careless move like this can quickly turn the game around. 20...f4! 21.Qe1
(21.Bxf4 Nxc3µ) 21...Rf8! (21...g5? 22.hxg5 hxg5 23.Nxg5²) 22.Nh2 (22.g3 g5µ) 22...g5!³ and Black takes over the
initiative.]
20...Kb8 21.Kh1 f4! 22.gxf4 Nf5³ Black successfully locks in White’s dark-squared bishop, which is now nothing
more than a big pawn. Potential breaks on the kingside are impossible and the game is equal.

19...a5

593
Black shows his ambitions. He usually goes for the ...b5–b4 break when he has castled kingside. The idea is to play
...b4–b3, creating an advanced passed pawn. If White captures on b4, then Black takes on d4, dismantling White’s
center. The French Defense has many adherents and is known for its fighting qualities, as Black is actually on the
look-out for any mistakes by White to immediately take over the initiative.

20.Bf3!?

The knight goes to e2!

20...Ne7 21.Ne2 Bc6

At present all Black can do is watch, so he shuffles his pieces around, trying to make useful moves.

22.Rg3

22.Kg1 Kb8 23.Be3 Rdg8 24.Qd2 Bd7

594
25.Kh1? The plan to put the white king on the h1-square to prepare the g2–g4 break simply never works. 25...g5!–+

22...Rdg8 23.Kg1 b5

This ambitious move is quite possible, but a bit too aggressive in my opinion. Personally I would look for ways to
reposition the e7-knight to a more useful location, like the f7-square, though Black would be vulnerable to an attack
on the g6-pawn. Besides, a move like ...b5 potentially opens up the black king too early and a bit too much. An
additional disadvantage of this move is that Black can no longer transfer his knight to the f7-square without leaving
the b5-pawn unprotected as long as the white rook is on b1.
23...Bd7 24.Nf4 Qf7 25.Nh3 Qg7 26.Qe1

26...Qf7 (26...Nc6 27.Nf4 Qf7? Black must also be careful. 28.Bxd5 exd5 29.e6 Bxe6 30.Qxe6+ Qxe6 31.Nxe6±)
27.Be2 (27.Be3?! Nc6 28.Qd2 Qf8 29.Ra1 Nd8 30.Be2 Nf7µ) 27...Nc6 28.Qc1 Qf8 29.Nf4

29...g5!? An interesting attempt to take over the initiative, however all White’s forces are well placed to meet the
aggression. (29...Qf7= and the position is equal, since White must go for the perpetual to prevent the black c6-knight
from completing its journey to f7, when Black’s position will be preferable.) 30.Nh5 f4!
595
31.Rh3 (31.Rf3?! Qf7 32.Nf6 Rg6 33.Rh3 Rxf6! A thematic exchange sacrifice by Black. In such a closed position,
rooks are almost useless and the white bishops even more so. Black can prepare the central ...e6–e5 break and now it
is White who has to sit and watch. 34.exf6 e5) 31...g4 (31...Be8 32.Nf6 Rg7 33.hxg5 hxg5 34.Bg4±) 32.Rh2 f3
33.Bf1 Rg6 34.Qe1∞

This position is difficult to evaluate, as both sides have trumps. The only question is whether Black’s f- and g-pawns
are a strength or a weakness. 34...Rhg8 35.Kh1 Qf5 36.Nf6 Rxf6 37.exf6 Qxf6 38.Qe3 fxg2+ 39.Bxg2 h5 40.f3 g3
41.Rh3∞.

24.Nf4 Qf7 25.Nh3

596
The knight has completed its journey; now the b1-rook must enter the game.

25...Qf8 26.Be2 Kd7!?

Black intends to keep his queen on the f5-square, so the king goes to protect the e6-pawn ahead of any potential
white Nh3–f4, so he can meet this move with the ...g6–g5 break.

27.Ra1

The rook protects the a3-pawn before the queen can move from c1.

27...Rh7 28.Bf4 Rhh8

Black starts to shuff le back and forth. Around here he offered me a draw, which I declined, as I felt that my strategic
plan had a chance of success.

29.Qd2

597
29...Rh7?!

Although it doesn’t change the evaluation of the position, this is a small psychological error which allows me to
complete my plan. But what if Black had played 29...Nc8, preventing my a1-rook from leaving? I do not have the
answer to that question as I would not seriously have considered putting my queen on a2 to proceed with my plan!
Or does it prevent White...!?
29...Nc8! 30.Kh2 Kc7 31.Rh1! Qxa3? 32.Bxh6± The exchange of pawns clearly favors White. Now there are real
targets for the white pieces, the black e6-and g6-pawns, so the conclusion is that Black could not quite prevent the
transfer of the a1-rook to the kingside.

30.Re1 Kc8 31.Qc1 Bd7 32.Bd2²

Phase one of White’s mega-plan is complete. The e1-rook will join the battle shortly. Now White must see to the g4-
break and somehow exploit the misplaced black knight on a4.

598
32...Bc6

Black is a bit careless, giving White too much of a free hand. The plan to transfer the knight to the f7-square was
still good and had to be played.
32...Nc6! was a safer choice. It is difficult for White to improve; for example if he plays 33.Kh2 then Black has the
typical 33...g5!, with the idea 34.hxg5 hxg5 35.Bxg5 f4! 36.Bxf4 Rxg3 37.Bxg3 Nxc3 and he’s doing great.

33.Kh2 Kd7 34.Rh1 Rhh8 35.Kg1 Rh7 36.Nf4

With the idea of transferring the knight back to the e2-square to avoid getting entangled after the potential g2–g4
break. On e2 the knight will be quite passive, but it will support the d4-pawn in case of Black’s ...b5–b4 break.

36...Rf7 37.Bf3?!

White could have attempted to go for the g4-break earlier with the knight on f4.
37.Kf1 Rh7 (37...g5? is premature. 38.hxg5 hxg5 39.Nh3!± This is the correct retreat. White must prevent Black
from playing ...f5–f4.) 38.Rgh3 Rf7
[38...Qf7 39.g4 fxg4 40.Bxg4 Nf5 41.Rg1²

599
Finally the position opens, and the white pieces, in particular the bishops, get concrete targets: advantage to White.
41...Kc8 42.Rf3 Kb7 43.Bxf5!? Not the only way, but the most expedient, as Black cannot recapture with his queen
without losing material. (43.Ne2 h5 44.Bh3 White still has to capture on f5 in order to open the g-file. 44...Nb6
45.Bxf5 gxf5 46.Rxg8 Qxg8 47.Rg3 Rg7 48.Rxg7+ Qxg7 49.Bg5

White is much better. Here is an example of how the game could develop: 49...Na4 50.Nf4 Qf7 51.Bd8 Ka6
52.Ne2± White’s queen is finally able to infiltrate the enemy camp and, as you can see, the difference between the
white and black bishops is quite stark. 52...Qh7 53.Qf4± White plans to improve his position slowly and get to
black’s weak h5-pawn. The important thing in such positions is to limit any possible counterplay by your opponent.
In this case, the only sources of counterplay for Black are the ...f5–f4 or ...b5–b4 breaks. 53...b4 54.cxb4 axb4
55.axb4 c3 56.Kg2 Bb5 57.Ng1!

600
The knight must have access to the f4-square. 57...Qg6+ 58.Qg5 Qh7 59.Qf6 Qg8+ 60.Kh2 Bf1 61.Qg5 Qh7 62.Qc1
Bb5 63.Nh3 f4 64.Nxf4 Qf5 65.f3 Nb2 66.Bg5 Na4 67.Kg3 Qf7 68.Ng2 Qg6 69.Ne1!

Slowly, agonizingly slowly, White neutralizes Black’s counterplay and inches his way to victory. 69...Qf7
70.Qf4+–) 43...Qxf5 44.Rgg3 and White is much better, e.g. 44...Rf7 45.Nxg6 Qxf3 46.Rxf3 Rxf3 47.Qd1 Rf7
48.Qg4 Bd7 49.h5+– and the passed h-pawn will bring White victory.]
39.g4!?

601
39...fxg4 ( 39...g5 40.hxg5 hxg5 41.Nxe6!± Just like in the game, this thematic piece sacrifice works well for White.
Black should have protected the e6-pawn with his bishop instead. 41...Kxe6 42.gxf5+ Nxf5 43.Bg4+– and White
regains the material with interest.) 40.Bxg4 Rf5 (40...Nf5?! 41.Nxe6 Kxe6 42.Bxh6 Qe8 43.Rf3!

This is why protecting pawns with your king, when your opponent has the bishop pair, is NOT recommended! Black
is lethally pinned and loses material.) 41.Rg3 Qf7 42.Rhg1 White is prepared to take on f5, since the g6-pawn will
fall if Black recaptures with the knight. 42...Rxf4 43.Bxf4 h5 44.Bh3 Rf8 45.Rf3+– and White’s material advantage
has become decisive. 45...Nxc3 46.Bg2 Ne4 47.Bd2!

602
47...Qxf3!? A positional queen sacrifice. It looks beautiful, but there isn’t enough compensation for the queen.
(47...Qe8 48.Bxa5+–) 48.Bxf3 Rxf3 49.Be1 b4 50.axb4 axb4 51.Kg2 Rf8 52.Bxb4 Rxf2+ 53.Kh1 Nf5 54.Qa3+–

White’s king is sufficiently defended and he goes for the decisive counter-attack.

37...Rh7 38.Ne2 Rgh8 39.Kf1 Rf7 40.Rgh3 Rhh7 41.Nf4

With the time control past, I decided to go for the g2–g4 break, having spotted the knight sacrifice on e6. Here I was
pretty confident.

603
41...Rhg7?!

Unexpectedly giving White a sudden chance.

42.Be2

a) 42.g4² immediately was strong too.

b) 42.Nxe6 Kxe6 43.Bxh6!? Now, how to evaluate this position? During the game I felt that Black would get too
much counterplay on the queenside after 43...Nxc3 44.Bxg7 Rxg7 45.h5

but it seems that the black king on e6 is way too vulnerable for Black to successfully push ...b5–b4. 45...g5! Black
keeps the position, especially the h-file, closed at all costs to prevent the white rooks joining the game. For this
reason, he even jettisons the g5-pawn. (45...b4 46.axb4 axb4 47.hxg6 Nxg6 48.Bh5!

604
Very strong. The g6-knight, a great defender, is eliminated, and with its departure the white rooks become forces of
destruction. 48...Ne4 49.Bxg6 Rxg6 50.Rh8 Qe7 51.R1h6 Be8 52.Qa1+– and the white pieces get to the black
monarch via the open lines.) 46.h6 Rh7 47.Qxg5 Ne4

48.Qf4 (48.Qc1?! b4 49.axb4 axb4 50.g4! The only way for White to attempt to claim the advantage is to open the
position, otherwise he faces annihilation on the queenside.) 48...Ng6 49.Qc1 b4 50.axb4 axb4

605
51.Rh5 (51.g4?! c3∞ and Black has achieved his dream counterplay with good chances to win.) 51...c3 52.g3! Nd2+
53.Kg2

Now White threatens to transfer the bishop to the d3-square and use the open a-file to transfer the queen from c1 to a
more active location, however... 53...Nxe5!! and the game is on!! (53...Ne7 54.Be2; 53...Bb5 54.Qa1 Qd8 55.Bd1!±
with a completely unbalanced position, although White’s chances seem higher, thanks to his attacking possibilities
against the black king. 53...Nxf3 54.Kxf3± and probably the computer is right that White is much better, thanks to
the far-advanced passed h6-pawn which keeps the h7-rook tied down and out of play. Black has big problems
defending his weak center and the queenside.) 54.Be2 Be8 55.dxe5 Bxh5 56.Rxh5 Rxh6 57.Rxh6+ Qxh6 58.Qa1
Kxe5²

606
A draw is the most likely result.

42...Rg8 43.g4!

Finally!!! Even now, I’m pretty amazed at the patience I had displayed so far in this game. The position now
becomes open and the white pieces finally get targets. The long maneuvering phase also lowered Black’s sense of
danger and as a result he falls for the tactics which I had to calculate exactly.

43...g5?

The losing move, however, if Black just waits, he will be stuck with a strategically inferior position which would be
very unpleasant to defend, e.g.
43...Bb7 44.gxf5 The most logical move of course.
[44.g5?! An interesting but dubious computer resource that I hadn’t even looked at. 44...hxg5 (44...h5 seems
obvious, but the white pawns become too strong after the sacrifice 45.Bxh5 gxh5 46.Nxh5 Kc8 47.Nf6+–) 45.hxg5
607
Rfg7

and it’s not quite clear to me how the computer intends to make progress for White, as the invasion on the only open
file doesn’t bring results against sturdy black defense. 46.Kg2 Qf7 47.Rh8 Rxh8 48.Rxh8 Rh7=]
44...Nxf5 (44...Rxf5 45.Bg4± leads to a position similar to what we have looked at before.) 45.Rg1 Rfg7 46.Bg4 Qf7
47.Rf3± and Black is in for a lot of suffering.

44.hxg5 hxg5

45.Nxe6!!

The only way! I recall this move was a shock for my opponent. He is finally forced to calculate a lot of lines, but it is
too late. White wins, despite being a piece for a mere pawn down; his bishops become monsters.
45.Nh5?! Perhaps Black was hoping for this move, which would have led to a completely blocked and drawish
game. 45...f4! 46.Nf6+ Rxf6 47.exf6 Qxf6
608
and now it is Black who will be fighting for an advantage: 48.Rh7 Rg7 49.Rxg7 Qxg7 50.Kg2 e5 51.Rh5! White’s
only counterplay lies in maximizing the effectiveness of his queen and rook by taking the h-file. 51...e4 52.Qh1 b4!„
and Black expands on all fronts!! White might hold the balance, but he will have to be very careful.

45...Kxe6 46.Rh6+?!

Doesn’t spoil anything, but it was more accurate to take on f5 first:

46.gxf5+ Rxf5 47.Bg4 Kd7 48.Bxg5 Kc7 49.Bxf5

49...Qxf5 The most aggressive defense. (49...Nxf5 50.Qf4 Kb6 51.e6+– and Black’s offside a4-knight is the reason
why his position is lost.) 50.Bxe7 Bd7 51.e6! Perhaps the easiest. White needs to open up the position further so that
the black king will not feel safe even on the queenside. 51...Bxe6 52.Rh7 Bd7 (52...Kc6 53.Qe3+–) 53.R1h6 Nxc3
54.Bd6+ Kc8 55.Rh8!

609
After this rook exchange, the white king will be completely safe, while its black counterpart, on the contrary, will
start to feel the heat.
55...Re8 (55...Rxh8 56.Rxh8+ Kb7 57.Qh6 Ne4 58.Bf4 Nf6 59.Qf8+– with a decisive attack) 56.Qe3 Ne4 57.Bf4
Kb7 58.Rxe8 Bxe8 59.f3+–.

46...Ng6

46...Rg6 47.gxf5+ Rxf5 48.Rxg6+ Nxg6 49.Bg4 Kd7 50.Bxg5 Kc7

51.Qe3!! The only move to keep the game won. White cannot give up his light-squared bishop or he’ll actually be
worse!! 51...Kb7 (51...Nxc3 52.Rh7+ Kb6 53.Bxf5 Qxf5 54.Bd8+ Ka6 55.Qxc3+– Suddenly with mating threats on
the queenside.) 52.e6! White’s main trump, his central passed e-pawn, supported by both bishops, decides the game.
52...Nxc3 53.Rh7+ Ka6 54.e7 Rxf2+ 55.Qxf2 Qxf2+ 56.Kxf2 Ne4+ 57.Ke3 Nxg5 58.Rg7 Nxe7

610
59.Rxg5! b4 60.axb4 a4 61.Kd2 a3 62.Kc3+– and Black’s position is hopeless.

47.gxf5+ Rxf5 48.Bg4 Kd7

49.R1h5!+–

A ver y important intermediate move, and the only one to win the game.
49.Bxg5 Kc7 50.Qe3 Kb6∞ Compared to the 46...Rg6 lines, the presence of the rooks, especially the black g8 rook,
makes all the difference. 51.Bf6 (51.e6?? Rxg5 52.Qxg5 Ne7–+) 51...Bd7 52.Qh3 Qxa3! 53.Bxf5 Qc1+ 54.Ke2
Nf4+ 55.Kf3 Nxh3 56.Rxc1 Bxf5

611
57.Bd8+ Kb7 58.Bxa5 Ra8!! 59.Bb4 Nxc3 60.Bxc3 Ra3 61.e6 Rxc3+ 62.Ke2 Nf4+ (62...Kc7? The position is still
very sharp, and this natural move would be a grave error. 63.e7 Kd7 64.Ra1+–) 63.Kd2 Rf3 64.e7 c3+ 65.Ke1 Bd7
66.Rb1 Be8 67.Rh8

67...Rh3! 68.Rxh3 Nxh3 69.Rb3 Ng5 70.Rxc3 Ne4 71.Ra3 Kc6 72.Ra8 Nd6= and finally it’s a fortress.

49...Kc7 50.Bxf5 Qxf5 51.Rxg5

612
51...Qe4!

The sharpest try. Now White has to take care in a tactical skirmish.
51...Qf7 52.Qd1!+– This re-entry of the white queen into battle via the d1- and g4-squares is decisive.

52.Qd1! b4!

Again the sharpest and best defense by Black.

52...Nb2? 53.Qg4

53...Qxc2 (53...Qxg4 54.Rxg4 Be8 55.f4 Kd8 56.f5 Rf8 57.f6 Ne7 58.Rh7 Nc6 59.Bh6+– is a simple win for
White.) 54.Rgxg6 Qb1+ 55.Kg2 Rxg6 56.Qxg6 Qxg6+ 57.Rxg6 b4 58.cxb4 axb4 59.Bxb4+– with a technical win.

53.axb4 axb4

613
54.cxb4?!

During the game I thought this was the most accurate, and that 54.f3 was an error, however the reverse is true!! In
my defense, I had calculated this far before making my 43.g4 break and evaluated this position as much better, if not
winning. The evaluation is correct, but the calculation needs correction!!
54.f3! is much better. 54...Rf8 (54...Nxc3 55.fxe4 Nxd1 56.Ke1! b3 57.cxb3 Nb2 58.Rgxg6 Rxg6 59.Rxg6 cxb3
60.Rg7+ Kc8 61.Rg3 Nc4 62.Rxb3 dxe4 63.Rb4+– with a technical win) 55.Rgxg6 (55.Kg2? Nh4+–+) 55...Rxf3+
56.Kg1

56...Bd7!
[56...Be8 57.Rg7+ Kc8 (57...Kd8 58.Rh8 Rf7 59.Rxf7 Qg6+ 60.Bg5+! Qxg5+ 61.Kf1+–) 58.Rh8+– The threat of a
linear mate is too strong.]
57.e6! (57.cxb4?? Only this move spoils everything. 57...Qxd4+ 58.Kh1 Qe4= 59.Kg1!= keeps the balance.)
57...Bxe6
614
[After 57...bxc3 58.Be1! Black’s counterattack should fail. Some nice lines: 58...Nb2 59.Qa1! White bravely moves
his queen further away from his exposed king!! Despite the ominous-looking threats, the white king is safe.
(59.Qxf3! is another incredible win: 59...Qxf3 60.Bg3+ Kd8 61.Bh4+ Kc7 62.exd7 Qd1+ 63.Kh2 Qxc2+ 64.Rg2+–)
59...Ba4 (59...Qe3+ 60.Kh2 Qf4+ 61.Bg3+–) 60.Rg7+ Preparing to roll the pawn to its destiny. 60...Kc6 61.e7+ Kd7
62.Rh8!!+– and we have reached an unbeliveable position.

The white king is wide open, the Qa1 is far from the action, Black has numerous checks with possibilities to capture
the white rooks — but White wins!! 62...Qxd4+ 63.Kh2 Qe5+ 64.Bg3 Qe2+ 65.Kg1 Qe3+ 66.Kh1+– and the white
king escapes the checks.]
58.Qe1!!

The move that I missed. White wins here. 58...bxc3 59.Qxe4 dxe4 60.Bf4+! The decisive intermediate check.
60...Rxf4 61.Rxe6 Rf8 (61...Nb2 62.Re7++–) 62.Rc6+ Kb7 63.Rxc4+–

54...Qxd4?

The losing mistake. Black has fought bravely and could have posed significant problems with the natural 54...c3:
615
This move was the last chance, however, with precise play, White will come out on top. 55.Be3 Bd7! Black
successfully gets his sleeping bishop into the game with great chances to hold his inferior position.
a) 55...Bb5+ 56.Kg1 Ne7 57.Qh5 Be8 ( 57...Qxc2 58.Rxg8 Nxg8 59.Qf7++–) 58.Qh3!! The only move. White needs
to keep his attack going. ( 58.Qh4?! Qxh4 59.Rxh4 Rxg5+ 60.Bxg5 Nf5!

Setting up a blockade on the light squares. Despite White’s material advantage, Black’s defense holds.
61.Rh7+ Kc6 62.Rh8 Bd7 63.Be3 Nb2 64.Rf8 Be6 65.Rf6 Kd7 66.Kf1 Nc4 67.Ke2 Na3 68.Kd3 Nb5!=

616
The fortress is complete. Draw!!)
58...Rxg5+ ( 58...Qxc2? 59.Qe6 Rxg5+ 60.Bxg5 Nc6 61.Qxe8 wins material.) 59.Bxg5 Nf5 60.Rh8!

60...Bg6 61.Rg8 Qe1+ 62.Kg2 Qe4+ 63.Kh2 Qxc2 64.Rxg6 Qxf2+ 65.Qg2+– The blockade is broken and White
wins.
b) 55...Nh4!? A curious attempt to box White in!! And it almost works!! 56.Rxc6+™ ( 56.Rxh4? Qxh4 57.Rxg8
Qh1+ 58.Rg1 Bb5+–+) 56...Kxc6 57.Rxg8 Qh1+ 58.Rg1 Qh3+ 59.Ke2

617
59...Nb2 (59...Qf3+ 60.Ke1 Ng2+ 61.Rxg2 Qxg2 62.Qd3 Kb6 63.Ke2+– and White crawls out of the box with a
winning material advantage.) 60.Rg3! Easiest. (60.Qa1!? Qf3+ 61.Ke1 Ng2+ 62.Rxg2 Qxg2 63.Qa6+ Kd7 64.Ke2!
The winning move, as Black loses by just one tempo. 64...Qh1 ...Qd1 65.e6+ Ke7 66.Bg5++–) 60...Qf5 61.Qc1

61...Kb5! (61...Qh5+ 62.f3 Nf5 63.Rg2 Nh4 64.Rg4+–) 62.f3 Nc4 63.Kf2+– White’s material advantage is too great.
56.Kg1 Ne7 57.Qh5 Rxg5+ 58.Qxg5 Nf5 59.e6! Again, lines of attack and opening up the white bishop are more
important than pawns.

618
59...Bxe6 60.Rh7+ Bd7 (60...Kc6? 61.Qd8+–) 61.Bf4+ Kc6 62.b5+ Kxb5 63.Rxd7 Qe1+
[63...Nb6 64.Rb7 Qe1+ (64...Qxc2 65.Qg6+–; 64...Nxd4!? The position is so irrational that, for a human player, it
looks like Black has very dangerous counterplay, but the computer sees it as an absolutely clear win for White.
65.Qf6! Kc4 66.Kh2!+– and I stopped trying to make sense of the mess.) 65.Kg2 Qe4+ 66.Kh2 Qxc2 67.Qg2 Kc6

68.Rc7+ Kb5 69.Qf1+ Nc4 70.Qe1! and under the computer’s strong guiding hand, White wins.]
64.Kg2 Qe4+ 65.Kh2 Nxd4 66.Rxd5+ Kc4

619
67.Ra5!! Just look at this position and White’s last move — absolutely unbelievable and beautiful!! White finally
got to Black’s wayward a4-knight!!! 67...Nf3+ 68.Kg3 Nxg5 69.Rxa4+ Kd5 70.Rxe4 Nxe4+ 71.Kf3 and the ending
should be winning.

55.Rg4!+–

Black’s fate is sealed. He loses material.

55...Nc3

55...Qxe5 56.Rhxg6+– … Bf4

56.Rxd4 Nxd1

57.b5??

620
Trying to play beautiful chess doesn’t always work. Sometimes it’s just better to collect the point safely. This move
could have cost me the game, which at this point had already lasted for many hours.
57.Ke2! Nb2 58.Rg4 Be8 59.f4 Ne7 60.Rxg8 Nxg8 61.Rh7+ Kc6 62.Ke3! Na4 63.Kd4

Completely dominating the position, especially the black knights, which are a sad sight to behold. 63...Bd7?!
64.b5++–

57...Bxb5?

A simple blunder. Black was probably completely demoralized when he played this move, which allows a
picturesque finish. However if my opponent had thought more, he would have realized that there was a much better
defensive resource.
57...c3! is more persistent: 58.bxc6 cxd2 59.Rxd2 Nb2! (59...Nc3 60.Rd3 Ne4 61.Rxd5 Kxc6 62.Rd4 Ng5
63.Rd6++–) 60.Rxd5 Nc4 61.e6 Kxc6 62.Rd4 Kc5±

and there is still a lot of struggle ahead as the black knights have finally achieved coordination and support each
621
other beautifully to prevent the white pawns from advancing.

58.Ba5+ Kb7 59.Rxd5! Ba6 60.Rd7+ Kb8 61.Rd6

A picturesque position — all Black’s minor pieces are under attack by the d6-rook. Black resigned. 1–0

Quite an error-laden game, especially the last part, but that is to be expected in a position where normal standards are
erased and the material balance is replaced by a dynamic balance involving lots of nuances and tactics where, even
with the computer’s help, it is not easy to get through the maze of complications. Hopefully the reader will set aside
some time and attempt to re-examine or train himself to seek tactical discoveries in the position that arose after
Black’s 52nd move. Perhaps then he’ll feel humbled and glad at the same time that chess can be so difficult, even for
the top players!

622
 GAME 20
Gata Kamsky
Viswanathan Anand
Dos Hermanas 1996
I decided to take part in the Dos Hermanas event because I needed some practice before the upcoming FIDE World
Championship match vs. Anatoly Karpov, which was to take place less than a week after the Spanish event. All the
top players, with the exception of Mr. Karpov, were present, including Mr. Kasparov, whom I drew against in the
first round. I lost two games vs. Topalov and Gelfand, and had no wins going into the last round. However, given the
previous history between Mr. Anand and me, I was looking forward to a fight, as I strongly felt that my opponent’s
complicity in obtaining an unfair win vs. me in the PCA match was less than honorable behavior on his part. As a
result, I would continue to play against Mr. Anand with additional motivation, even after my subsequent return to
chess in 2004.

1.e4

I declare my intentions with this move.

1...c6

This somewhat surprised me. Clearly I wanted to keep all of my preparation for Mr. Karpov hidden and felt that my
opponent had decided to take advantage of that. After thinking for quite a bit, I decided to go for one of the older
main lines, having analyzed some of them on several occasions, but not as deeply as I would have wished.

2.d4 d5 3.Nc3 dxe4 4.Nxe4 Nd7

In modern days 4...Bf5 is played almost exclusively. 4...Nf6 is also tried, but mostly by players who feel the need to
play for a win against lower-rated opponents. We are not going to discuss the theory as that is not the purpose of this
book, although I personally would adopt this system with Black as it would be quite a fighting opening, compared to
the ‘Spanish torture’.

5.Ng5

The main move in this line, popular since the famous Kasparov-Karpov match in Seville. White attempts to provoke
Black into weakening some squares on the kingside, which could become useful after White’s standard queenside
623
castling.

5...Ngf6

If Black plays 5...h6 then White demonstrates the main point of his last move with 6.Ne6!, as the knight is
untouchable due to Bg6 mate.

6.Bd3

6.Bc4 is another main alternative.

6...e6

Considered to be the main line. Other moves like 6...g6 have also been tried, but require further testing.

7.N1f3 Bd6

Now that Black has developed his f8-bishop he is finally ready to chase the white g5-knight with ...h6. To do so
without developing the bishop is countered by the knight sacrifice on e6 which is considered to give White a
dangerous initiative as compensation. Modern analysis in this line is very heavy and, as in most openings with
forced lines, requires a good memory.

8.Qe2

The point of this move is to recapture on e4 with the queen, which can then move further to the kingside in
preparation for the attack.

8...h6 9.Ne4 Nxe4

10.Bxe4

This probably came as a slight surprise to my opponent as 10.Qxe4 is considered to be the main move by far.
However I did not want to enter the heavily-analyzed lines after 10...Qc7. Black is now forced to play his own game,
unaided by theory, and, given my opponent’s distaste for playing positions with limited space, I felt I had made the
right choice.
624
10...Nf6

Probably a bit too early. Usually Black reserves this tempo until White declares which side he is going to castle.
Both 10...0-0 and 10...Qc7, followed by ...b6, ... Bc8–b7 and ...0-0-0, were serious alternatives.

11.Bd3 Qc7

Or 11...b6, as the moves are interchangeable. Black too remains flexible with the option of castling, as he has no
desire to fall into a direct kingside mating attack.

12.Bd2

Played after some thought. I wanted to keep the tension re castling, while continuing with development. The bishop
is modestly placed on the d2-square, however White will have the option of pushing c2–c4, then transferring it to the
c3-square on the long a1–h8 diagonal, which would provide him with a small advantage in space.

12...b6

Both sides continue with the development of their pieces, however now I decided to embark on a rather ambitious
plan.

13.Ne5!?

The idea behind this move is to try to prevent the ...c6–c5 pawn push, which not only opens up Black’s light-squared
bishop, but also undermines White’s central d4-pawn. I also started to see the outline of my general plan of castling
short and an all-out direct attack on the kingside, as Black’s set-up with the queen on c7 and bishop on d6 is not
ideal for fighting against White’s isolated d4-pawn.
13.0-0-0 is of course the most logical continuation. 13...Bb7 14.Kb1 0-0-0 15.Rhe1 c5 16.dxc5

625
16...Bxc5 (16...bxc5?! This pawn recapture is a bit dubious as it exposes the black king, limits the d6-bishop and
creates a possible weakness on c5 if White manages to fully block the c4-square. 17.Ba6²) 17.Bc3 (17.Ba6 Bxa6
18.Qxa6+ Kb8 19.Qe2² is also possible. White’s pawn structure is preferable, as he has more control in the center
and a great potential outpost on e5 for the f3-knight. In fact this is a standard position where Black is slightly worse,
but has solid, albeit slightly passive, development, with good chances of a draw.) 17...Kb8 18.g3²

Getting the kingside pawns out of the X-rays of the black bishops. White has a small advantage as he has greater
flexiblity in maneuvering his pieces. Nevertheless I feel it is better for White to play this sort of position by
exchanging the lightsquared bishops.

13...c5N

An ambitious and new reply at the time. The usual reply in this position is the calm 13...Bb7, which results in
positions similar to those arising after 13.0-0-0. The main difference is that White’s f3-knight is already on the
important e5-square and supported by the f4-pawn. Whether it is better for him is still under discussion.

626
14.Bb5+

Of course. Given the chance, White has to give this check, as it forces Black to give up castling rights. Other options
lead to a passive game for Black. The only question is whether the black king should go to the e7- or f8-square.
Intuitively it should go to the e7-square, connecting the rooks and possibly followed by ...Rh8–d8 and ...Ke7–f8,
when Black has avoided a direct kingside attack and managed to deploy all his forces to the center. However the fact
that after this game very few people ventured into this line suggests that not too many people are happy to play
Black in this position.

14...Ke7

a) 14...Bd7? 15.Nxd7 Nxd7 16.d5± would be a strategic disaster for Black, as White’s strong bishops and the very
weak light squares in Black’s camp make the position very difficult to defend.
b) 14...Nd7? 15.Nxf7! Kxf7 16.Qf3+ Nf6 17.Qxa8 cxd4 18.0-0-0 Rd8 19.Qc6+– and White is just in time to return
with his queen before it gets trapped. His material advantage would be decisive.

15.c3!

White agrees to be left with an isolated d-pawn, however he gets much in return in the shape of Black’s misplaced
king and the potential f2– f4–f5 standard break, which would make the black king very vulnerable in the center.
15.0-0-0 is another possibility, but I didn’t feel much confidence in castling into a potential queenside attack. This
move also sets up a tactical trap, which, however, is easily avoided. 15...Bb7 16.dxc5!?

627
16...Qxc5 (White dreams of 16...Bxe5? 17.cxb6 with the lethal threat of Bb4+, and if 17...Qc5 then 18.Qxe5!!+–

and White reclaims his material with interest as the queen is untouchable in view of Bd2–b4+.) 17.Bc3 Rhd8 18.Bd4
Qc7 19.f4 Bc5= Black starts to exchange pieces, leading to an equal game.

628
20.Bxc5+ Qxc5 21.Bd3 Rd4 22.Rhf1 Be4! Destroying White’s last hope for the attack, the d3-bishop, without which
White has no attack at all. Following massive exchanges on the d-file leading to the endgame, Black’s king will be
well placed in the center.

15...Bb7 16.0-0

I was quite happy with my position here, as it is not easy to play for Black. White has the clear-cut plan of
centralizing his rooks and getting in the f2–f4–f5 push.

16...Rhd8 17.f4

I went for the pawn push immediately, fearing that my center would not hold if I played slowly. 17.Rad1?! Kf8
18.Rfe1 cxd4 19.cxd4 Rac8 20.Bd3 Qe7 21.a3 Kg8 and Black has the better chances thanks to White’s worse pawn
structure and no possibility of attack.

17...Ne4
629
17...Kf8

18.Bd3 Probably White’s most principled reply, returning the bishop, which was no longer effective on the b5-
square, back to the important d3–h7 diagonal, helping to cover both the c2- and e4-squares.
a) 18.Rae1 cxd4 19.cxd4 Qc2!µ

An excellent defensive move that pins White’s forces along the second rank. White cannot allow the exchange of
queens due to his inferior pawn structure. 20.f5 The only chance for White to fight for an advantage. 20...Qxb2! Not
fearing ghosts.
[20...exf5 21.Nxf7

630
21...Re8! (21...Kxf7? 22.Bc4+ Kg6 23.Bd3 Ba6 24.Bxc2 Bxe2 25.Rxe2 Re8 26.Bxf5+ Kf7 27.Bd3± White keeps an
extra pawn and strong pair of bishops.) 22.Ne5 Qxb2 23.Bxe8 Qxd4+ 24.Kh1 Rxe8 25.Rxf5 Rxe5 26.Rxe5 Bxe5
27.Qxe5 Bxg2+ 28.Kxg2 Qxd2+ 29.Re2 Qd7 30.h3²

Despite reduced material, this position still has a lot of play left in it.]
21.Nxf7 Kxf7 22.fxe6+ Kg8

631
23.Rxf6 (23.e7 Bxe7 24.Qxe7 Qxd4+! 25.Be3 Qd5 26.Rf3 Qxb5 27.Qe6+! Kh8 28.Rxf6 Qd7³!) 23...Qxd4+!
(23...gxf6 24.Qh5= leads to a draw.) 24.Rf2 Qh4 25.h3 Rf8!³ and it is Black who is attacking.
b) 18.f5 Looks strong, but the tactics work out fine for Black. 18...cxd4 19.cxd4 Bxe5! Black’s most logical and best
choice, as he opens the central lines for his heavy pieces and gets rid of White’s powerful e5-knight, which is more
important than keeping the pair of bishops. 20.dxe5 Ne4

21.Bxh6! The only move for White to keep the balance. (21.Rac1 Qxe5 22.Bf4 Qf6³) 21...exf5!? (21...gxh6 22.fxe6
Qxe5 23.Rxf7+ Kg8 24.Qg4+ Ng5 25.Rxb7 Qxb5 26.h4 Qc6 27.e7 Rd5!= and Black keeps equality; 21...Qxe5
22.Bf4 Qc5+ 23.Be3 Qe5 24.fxe6 Qxe6 25.Rad1 Kg8 26.Bd4 Rac8 27.b3 Qg6= and the position is approximately
equal, since White’s strong bishop pair is compensated by Black’s great coordination and his perfectly placed e4-
knight.) 22.Rxf5 gxh6 23.Qh5! Kg8! 24.Raf1 Qc5+ 25.Kh1

632
25...Qd5 (25...Qxb5 26.Qxf7+ Kh8 27.Rf6!=) 26.Qg4+ Kh8 27.Qh4 Ng3+ 28.Qxg3 Rg8 29.Qf3 Rac8 30.Rf6 Rxg2

31.Bc6™ 31...Bxc6 32.Rxh6+ Kg7 33.Rxc6= The massive tactical fireworks fizzle out into an equal rook endgame.
18...Kg8 (18...Be7 19.f5 exf5 20.Rxf5 With a strong attack.) 19.Kh1∞
A complex middlegame is in progress.

18.Be3 Kf8

633
Black has successfully prevented White from breaking the position open with f4–f5, however the e4-square is not an
ideal location for the black knight, as it can be easily evicted, giving White additional tempi. Still, the position
remains balanced.

19.Rae1

White methodically brings all his pieces into active duty. The only one missing is the b5-bishop.

19...Kg8∞

Consistent, but perhaps it was time to declare the situation in the center.
19...cxd4!? 20.Bxd4 (I would most likely have played 20.cxd4. I wanted to avoid exchanges as much as possible as I
was afraid of reducing my attacking potential. The position is complex with many possibilities for each side.) Let’s
have a look at a beautiful line: 20...Bc5 21.Bxc5+ Qxc5+ 22.Kh1 Qc7 23.Bd3 Nf6 24.f5 exf5 25.Rxf5 Rd6! 26.Bc4

Setting the last trap. 26...Bd5! (26...Re8?? 27.Ng6+!! Kg8 28.Qxe8+ Nxe8 29.Rxe8+ Kh7 30.Nf8+ Kg8 31.Ne6+
634
Kh7 32.Nxc7 Rd1+ 33.Rf1+–) 27.h3 Rad8 28.Bxd5 Rxd5=.

20.Qg4?!

An interesting attempt for White to improve the position of his queen before going for the f4–f5 push.
20.Bd3 would be a normal alternative. 20...Nf6

21.Bf2 With the idea of transferring this bishop to the h4-square. (21.a3 cxd4 22.cxd4 Rac8 23.h3 preparing g2–g4
would make the game quite unclear, though Black would be much better prepared for White’s coming offensive on
the kingside.) 21...cxd4 22.cxd4 Rac8 23.Bh4 Be7=

20...Nf6!

Black immediately evicts the white queen from its best location and at the same time opens the possibility for
repositioning his b7-bishop to the e4-square, taking control of the vital b1–h7 diagonal.

21.Qg3?!
635
21.Qh3 would be more accurate to keep an eye on the f5-square. 21...Be4!
[21...cxd4 22.cxd4 Qc2!? Black attempts to transfer his own queen to the kingside where it will help defend the
black king and create additional threats. 23.Re2 Qh7 24.Bf2!

This is probably the correct way for White to play this position. The black queen is awkwardly placed on the h7-
square, and if it goes to f5, then Black’s structure could potentially be hazardous in the endgame, although in the
current sitution the f5-pawn allows Black to get a very strong grip on the e4-square. Both sides have chances to play
for a win here. 24...Qf5 25.Qxf5 exf5 26.Bh4∞]
22.Bf2! cxd4 (22...Bf5?! 23.g4 Bh7 24.Qf3! and White takes control of the a8–h1 diagonal, leading to a slightly
better game.) 23.cxd4 a6 24.Ba4 Rac8

25.Bb3! We have reached probably the critical position in this line. Both sides have their trumps and weaknesses.
The final verdict on the battle will have to be played out another time.
[25.Bh4?! Bf5! Using the fact that g2–g4 is now impossible. 26.Qf3 (26.g4? Bxe5 27.gxf5 Bxd4+ 28.Kh1 exf5–+)
26...Bxe5 27.dxe5 (27.fxe5? Rxd4–+) 27...Rd3 28.Re3 Rxe3 29.Qxe3 Nd5 30.Qd4 Qc4 31.Qxc4 Rxc4 32.Bb3
636
Rxf4µ with serious winning chances for Black.]
25...Bd5 (25...Bf5? 26.g4±; 25...Bxe5 26.fxe5 Bf5 27.Qf3 Nd5∞ with yet another unclear position.) 26.Bh4 Bxb3
27.Bxf6 gxf6 28.Ng4 Bxf4 29.Nxf6+ Kf8 30.Qxb3 Rxd4 31.Kh1

31...Bg5 (31...Be5? 32.Qe3±) 32.Ne4 Kg8 33.Qf3° This position is very unclear, with Black’s extra pawn being
compensated by White’s superior pieces and open black king.

21...cxd4 22.cxd4

22...Nd5?!

While this move is not bad, Black willingly blocks his own strong b7-bishop without getting anything in return. He
prepares to play ...f7– f5 to prevent White’s f4–f5 push, however, by doing so, he creates another weakness, the e6-
pawn.
22...Be4! An extremely strong centralizing move that exposes the weakness in White’s previous strategy. 23.Be2!
637
Probably best. White takes control of the critical g4- and h5-squares before he goes for his f4–f5 break. The game
remains roughly balanced.
a) 23.f5

This dangerous-looking move is probably the reason why Black didn’t play 22...Be4, however Black has sufficient
defensive resources.
23...exf5!
[23...Bxf5 Even this move is possible. 24.Rc1 (24.Bxh6? Nh5 25.Qf3 Bxe5–+) 24...Qb8 25.Bc6! (25.Bxh6? Nh5
26.Qg5 Bxe5 27.dxe5 Qxe5 28.Qxh5 gxh6 29.Qxh6 Qxb2!µ and Black has defended successfully.) 25...Ne4
26.Bxe4 Bxe5 27.dxe5 Bxe4 28.Bxh6 Bg6 29.Rxf7 Kxf7 30.Rf1+ Kg8 31.Qxg6 Qxe5=]
24.Bxh6

24...Nh5! (24...Ng4?? 25.Bxg7 Kxg7 26.Rxe4+–) 25.Qh4 Bxe5 26.dxe5 Qc5+ 27.Rf2 Qxb5 28.Qxh5 Qxe5!µ Black
gets a very strong bishop on e4, which, along with his material advantage, gives him a serious advantage.
(28...gxh6? 29.Rxe4=)
638
b) 23.Rc1!?

An interesting move to harass Black’s queen which was ideally placed on the c7-square, from where it could enter
the game with great effect via the c5-square. 23...Qe7 24.f5 exf5 (24...Nh5? 25.f6 Nxf6 26.Bxh6 Nh5 27.Qg4±)
25.Bxh6 Nh5 26.Qg5 Qxg5 27.Bxg5 f6 28.Be2 fxg5 29.Bxh5 Rf8 30.h3 Rad8 31.Bf3³

23.Bd2

After this quiet retreat, the black position becomes dangerous.

23...a6?!

Black decides to push the bishop away, since it was hitting a lot of important squares in the black camp, but this
really helps White as the bishop is even better placed on the d3-square than on b5.
a) 23...Ne7?! This move was probably my opponent’s original intention, but White wins an exchange here. 24.Rc1
Qb8 25.Nd7 Rxd7 26.Bxd7 Qd8 27.Ba4 Nf5 28.Qh3!
639
Preventing the black queen from reaching the perfect h4-square. Black has a certain amount of compensation in view
of his active pieces and extra pawn, but whether it will be enough is not clear. 28...Nxd4 29.Kh1²;
b) 23...Rac8? A normal developing move like this doesn’t quite work here. White is ready for a direct attack on the
black king, the full potential and danger of which is well illustrated by the following line:
24.f5 exf5 25.Bxh6 f6 26.Bd3!

The white bishop joins in with deadly effect. 26...fxe5 27.Bxf5 Rf8 (27...Re8 28.Qg6+–) 28.Qg6 Rxf5 29.Rxf5+–
c) 23...f5 24.Qg6 Bc8

640
25.Be8!! The key move, taking control of the f7-square. Black has to give up material to maintain equality.
25...Bxe5 26.fxe5 Qe7 27.Bc6 Rb8 (27...Bb7 28.Bxb7 Qxb7 29.Qxe6++–) 28.Bxh6 Qf7 29.Qg3 Bb7 30.Bb5! a6
31.Be2 Rd7 32.Bg5± with a serious advantage in the shape of an attack and extra pawn.
d) 23...Qc2 a bit late now, though Black keeps certain defensive chances. 24.Rf2!

24...Nf6 (24...Qxb2? 25.f5! Qxd4 26.Bxh6 Bf8 27.fxe6 fxe6 28.Qg6+–; 24...Bxe5 Again White’s e5-knight is a
monster, so it makes sense for Black to exchange it, even though this exchange opens up the f-file and White’s d2-
bishop. 25.fxe5 Qg6! 26.Qh3 Ne7 27.Bd3 Nf5 28.Rf4! with a very strong attack. I do not envy Black in this
position.) 25.Bd3 Qxb2 26.Bc3 Qa3 27.Nc4 Qa4

641
28.d5!! All White’s pieces join in the attack. Black cannot avoid losing material. 28...Nh5 29.Qh3 Bc5 30.Qxh5
Bxf2+ 31.Kxf2 Rxd5 32.Qg4 g5 33.Qh3! Rxd3 34.Qxd3 Qc6! 35.Qd4!

The king will be safe in the center. White forces Black to transpose into an inferior endgame. 35...Qxg2+ 36.Ke3
Qf3+ 37.Kd2 Qd5 38.Rg1! Rd8 39.Qxd5 Rxd5+ ( 39...Bxd5 40.Bf6 Rc8 41.Ne5+– and despite the limited pieces on
the board, White creates a mating net around the black king.) 40.Kc2!

642
40...f6! (40...Rc5 41.Ne5+–) 41.fxg5 (41.Bxf6? Rc5=) 41...Rxg5 (41...hxg5 42.Bxf6+–; 41...fxg5 42.Ne5+–)
42.Rxg5+ fxg5 43.Nd6 Bd5 44.a3± Despite Black having three pawns for the missing piece, he has to defend very
attentively and accurately to hold the position. Very possibly White is winning as all the black pawns are quite weak.

24.Bd3±

Suddenly, within the space of two slightly inaccurate moves, Black’s position has become critical.

24...f5

Probably the best, given the situation. Black cannot allow f4–f5 at all.
a) 24...Nf6 This doesn’t help much, but only if White finds the next move. 25.Rf2!! Best, and a very important move
which forces favorable complications. White safequards himself against any possible checks along the a7–g1
diagonal, and, more importantly, protects the d2-bishop. White’s f4–f5 push next will open the position when Black
is not ready. (25.Kh1 is less good: 25...b5 26.f5 Bxe5 27.dxe5 exf5! 28.Rxf5 Ne4! The point, as White’s d2-bishop
is not protected. 29.Bxe4 Rxd2³) 25...Bf8 (25...Rac8 26.f5 Bxe5 27.dxe5 Rxd3 28.Qxd3 Ng4 29.Rf4 Nxe5
643
30.Qg3+– White gets a material advantage in addition to his attack.) 26.Bc3!

White is clearly better, e.g. 26...b5 27.f5 exf5 28.Rxf5 b4 29.Rxf6 bxc3 30.Rxh6!+– wins on the spot. 30...Rd6
31.Bh7+ Kh8 32.Rh5 Rh6 33.Rxh6 gxh6 34.Qg8#
b) 24...Bb4 looks sensible, but is too late. White gets through to the black king by brute force: 25.f5! Bxd2 26.fxe6

26...f5 (26...f6 27.Qg6 Kf8 28.Qh7+– with a mating attack. 26...Bxe1 27.exf7+ Qxf7 28.Bh7+ Kxh7 29.Qd3+ Kg8
30.Rxf7 Rab8 31.Qg6+–) 27.Rxf5 Be3+ 28.Kh1! Bxd4 29.Rf7 Bxe5 30.Rxe5 Qc1+ 31.Bf1 g5 32.Qd3+–

25.Qg6!

644
Direct and good. White simply breaks through with g2–g4, opening up his monster d3-bishop. It is also noteworthy
that White’s entire army is mobilized for action!

25...Re8?

Amounts to capitulation. I recall my opponent not being happy with his position at all.
25...Bc8! would have put up stronger resistance. 26.Rc1! (26.g4? is not good now: 26...Bxe5 27.fxe5 Ne7 28.Qh5
Rxd4–+ and Black is winning.) 26...Bxe5 Most likely the move we both underestimated.
[26...Qb7!? 27.Rxc8!? A picturesque idea to break through Black’s defenses by tactical means, although it is
possible for White to play the position by slower, positional means. 27...Qxc8 28.Rc1 Nc7 (28...Bc7 29.Qf7+ Kh7
30.g4± and White breaks through on the light squares.) 29.Nc4!

29...Rb8 (29...Bf8 30.Nxb6 Qb7 31.Nxa8 Rxa8 32.Rxc7 Qxc7 33.Qxe6+ Qf7 34.Bxf5+– and White’s material
advantage is decisive.) 30.Nxd6 Rxd6 31.Bxf5+–]
27.Rxc7 Bxd4+ 28.Kh1 Nxc7 29.Rc1 Ra7
645
30.Bc4! Kh8 (30...Bxb2?! 31.Rc2 Bd4 32.Qxf5± Removing the key defender, the black f5-pawn, gives White a clear
advantage and the attack.) 31.Bc3 Bxc3 32.bxc3± And despite close to material equality, Black’s pieces are very
passive, which leads to serious winning chances for White.

26.g4!+–

White’s advantage is decisive. The attack cannot be stopped unless Black gives up a lot of material.

26...Nf6

Black attempts to utilize the justopened a8-h1 diagonal to create threats against the white king, but it is an illusion.
26...Bxe5 27.fxe5 Qf7 28.Qxf7+ Kxf7 29.gxf5 exf5 30.Rxf5+ Kg8 31.Bc4!

646
White’s pair of bishops, very strong connected passed pawns in the center and active rooks make this a winning
position, despite some technical hurdles.

27.gxf5 Bxe5

27...exf5 28.Bxf5 Re7 29.Re3!

The rook swings over to the g3-square, from where it will both attack and defend. 29...Qd8 30.Bd3! Rc8 (30...Bxe5
31.fxe5 Qd5 32.exf6+–) 31.Rfe1! With an extra pawn and superior position, White is in no rush to score the direct
hit. Instead it’s better to slowly improve the position. 31...Bd5 (31...Bxe5 32.fxe5+–) 32.Rg3 Kh8

647
And finally White is ready. 33.f5!? Qf8 34.Rh3 Rcc7 35.Bxh6 gxh6 36.Rxh6+ Rh7 37.Qxf6+ Qxf6 38.Rxf6+–.

28.fxe5 Qc6 29.Re3!?+–

The most aesthethically pleasing way to victory.

a) 29.exf6?? loses for White! 29...Qh1+ 30.Kf2 Qxh2+ 31.Ke3 exf5+ 32.Be4 Rxe4+ 33.Kd3 Rxd4+! 34.Kxd4 Rd8+
35.Kc4 Bd5+ 36.Kc3 Qc7+ 37.Kd3 Be4+ 38.Ke2 Qc4+ 39.Kf2 Rxd2+ 40.Kg3 Rg2+ 41.Kh4 Bc6+–+
b) 29.Qg3 This also wins, albeit prosaically. 29...Nh5 30.Qh3 exf5 31.Rxf5 Rad8 32.Be4 Qg6+ 33.Rg5!+–.

29...Qh1+ 30.Kf2 Qxh2+ 31.Ke1

The white king escapes the checks and Black loses material.

31...Ne4

31...Qh4+ 32.Qg3 (32.Ke2 Qg4+ 33.Qxg4 Nxg4 34.Rg3 h5 35.f6 is crushing as well.) 32...Qxg3+ 33.Rxg3 exf5
648
34.Bxh6+–

32.Bxe4 Qh4+ 33.Kd1!

The final touch. Black loses a piece and with it any hope.

33...Bxe4 34.Be1 Qh2 35.Rxe4 Qh3 36.Ke2 1–0

After this game, I went to Elista for my match vs. Karpov with at least some hope of success.

649
650
 GAME 21
Gata Kamsky
Anatoly Karpov
Elista 1996
After the final Candidates matches in India in early 1995 there was a lot of intrigue about my match vs. Karpov.
Despite my father representing me, there was, as always, very little information coming our way from the FIDE
President, Mr. Ilyumzhinov, who had just been elected to the post the previous year. It was pretty obvious that
Karpov was trying to get the best possible conditions for himself so he could get the title and possibly go on to
challenge Mr. Kasparov for the unified title. However I was still in the way and he had to beat me in the match.
Sometime late in 1995, Mr. Karpov asked me and my father for a meeting to discuss the chess aspects of the match.
Several provisions were agreed upon, including getting rid of adjournments, which were a relic of the past. My
delegation, or would it be more accurate to say my father’s delegation, since I only had a say in the matter of
choosing my chess assistants, consisted of a number of relatives who would assist us in the matter of catering. The
event took place in Elista, the capital of Kalmykia, the only region in Russia that claims Buddhism as the most
popular religion. Both Karpov and I were given houses for the duration of the match. One important factor was that
our house had no access to the internet, which would play a certain role later in the match. My chess assistants
included GMs Loek van Wely and Predrag Nikolic. The original assistants who had helped me before the start of the
whole PCA and FIDE Candidates cycles had fallen out of favor with my father and were deemed untrustworthy.
Upon arrival, I was handed the rules of the match, without prior consultation, and was told to accept them or leave
the match. The rules included adjournments!! It had become quite obvious to me that Mr. Karpov had had a certain
say in the matter and despite the welcoming facade on the face of the new FIDE President, the organization of the
match was doing everything possible for Mr. Karpov to win it. I lost the first game without putting up much of a
fight. However in the second game I was able to utilize a new idea that I felt was interesting. My assistants were not
convinced, but were unable to find holes in the lines, so off I went to play the game.

651
1.e4 c6 2.d4 d5 3.exd5

The first surprise. In almost all the games I had played vs. Karpov up until that moment, I had played 3.e5, leading
to the Advance Variation.

3...cxd5 4.c4

The Panov Attack. I glanced at my opponent, and it was obvious that he was not prepared for this option. Of course
he had played the Caro-Kann all his life and had extensive experience in all lines, including the Panov, so the
question in his mind was where the surprise was going to be.

4...Nf6 5.Nc3 e6 6.Nf3 Bb4

Black chooses the most solid setup, as this position can also be reached via the Nimzo-Indian Defense, my
opponent’s main defense to 1.d4.

7.cxd5

This is less common. White usually proceeds with 7.Bd3 and after 7...0-0 8.0-0 dxc4 9.Bxc4 b6 there arises the main
tabiya, which has been extensively analyzed and not played much by White in modern times.

7...Nxd5 8.Bd2 Nc6 9.Bd3

652
9...Be7

One of the three main moves in this position, the other two being 9...0-0 and 9...Nf6. It looks a bit paradoxical for
Black to first develop his f8-bishop to b4 and then voluntarily return with it to e7. The idea is that White’s d2-bishop
is not well placed and hinders the defense of the isolated d4-pawn. Black also has the possibility of placing the
bishop on the long diagonal after ...Be7–f6, followed by ...Bf6–g7, exerting additional pressure on White’s isolated
pawn and helping to protect the black king against any direct attack.

10.0-0 0-0 11.Qe2

The position after 10....0-0 is still being heavily analyzed and played in modern times. White’s other two moves,
11.Re1 and 11.a3, are preferred though. Both moves lead to a sharp game with a lot of possibilities for both sides —
the truth is still out there. The reason I chose 11.Qe2 is that I strongly believe in piece coordination, especially
amongst the heavy pieces, and also that White’s rooks are best placed on the d1- and e1-squares, both defending the
d4-pawn, which gives White some extra space, and preparing for the standard breakthrough in these positions, d4–
d5. The white queen could also potentially find a more active role on the kingside via the e4-square.

11...Nf6

This move was considered best at the time, and still counts amongst the most popular moves today. Alternatives
were 11...Bf6 and 11...Ncb4, with complex play.

12.Ne4!?N

653
And here was my new idea. Black’s f6-knight is a very strong defensive piece that needs to be exchanged.
Several games had been played before with this move, but this was the first game at the highest level, and it gained
in popularity, becoming the main line for White after 11.Qe2.

12...Bd7?!

A bit timid. In the fourth game of our match, my opponent’s team had prepared a big improvement, 12...Qb6,
exposing the drawback of White’s Nc3–e4 maneuver. White no longer has the option of Nc3–a4 and the black queen
is ideally placed on the b6-square, exerting pressure on both the b2- and d4-pawns. Having not seen such a move in
my preparation, I quickly fell into an inferior position and went on to lose the game, which did not help my
psychological state. Overall, it has to be said that in the openings that my opponent had expected from me, his
preparation was vastly superior.
12...Nxd4?! is of course the first option that I had to consider for Black. 13.Nxd4 Qxd4 14.Bc3

14...Qb6 (14...Qd5 15.Rfd1! Nxe4 16.Bxe4 Qg5 17.Rd3! g6 18.Rad1ƒ I felt that in this position White’s activity
more than compensated for the sacrificed pawn, since Black’s queenside development had effectively been stopped
654
and his queenside pawns might become easy targets.) 15.Nxf6+ Bxf6 16.Qe4 Rd8 17.Qxh7+ Kf8 18.Rae1 Bd7
19.Re3!

19...Ke7 (19...Bxc3?! 20.bxc3 Qc5 21.Qh8+ Ke7 22.Qxg7 Rg8 23.Qh6±) 20.Bxf6+ gxf6 21.Bg6 Be8 22.Rfe1 Rd6
23.h4!² White obtains the superior structure and the initiative.

13.Rad1

White is developing his rooks according to the outline described above.

13...Rc8

Here too 13...Qb6 would be preferable. However, since we are all taught as beginners that it is not a good idea to
develop your queen before the rest of the pieces, this rule becomes sort of ingrained and difficult to ignore in
concrete positions. Also, White has already developed his a1-rook to the center and Black’s 13...Qb6 can be well
met.
13...Qb6 14.b4!

655
Very strong, and not the kind of move that I would have considered at the time. It would only be around 2013, in my
game vs. Nakamura, that I would see an idea very similar to this one. White’s pawns are untouchable due to multiple
pins along the b-file. Furthermore, the b4-pawn restricts the mobility of Black’s c6-knight and gives White some
space on the queenside.
14...Nxe4 (14...Rad8?! 15.b5 Nb8 16.Nxf6+ Bxf6 17.a4±; 14...Nxb4 15.Nxf6+ gxf6 16.Rb1+– Black loses a piece.)
15.Qxe4 f5 16.Qe2 Bf6 17.Bc4 Kh8 18.Bg5!² with complex play where White has slightly preferable chances.

14.Rfe1

14...Nd5!?

All the white pieces are developed and Black has to be careful. The idea behind the f6-knight returning to the d5-
square is to prevent White’s queen from probing for weaknesses on the kingside. White’s e4-knight is also blocking
its own forces, so White’s next move is actually forced.

656
14...Qb6 is quite dangerous. 15.Nxf6+ Bxf6

16.Ne5!? White attempts to prevent Black’s f8-rook from joining the game.
[16.Qe4 is a more logical alternative. White creates threats on the kingside to provoke Black into weakening some
squares for the white pieces. 16...g6 17.Bh6 Bg7 (17...Rfe8 18.Qf4ƒ) 18.Qh4!‚

Black now has to be careful as White is ready for the thematic d4–d5 breakthrough in many lines. 18...Rcd8 (18...f6
19.d5! exd5 20.Bc4!!±) 19.Be4 f6 20.d5 Ne7 21.Be3! Qa5 22.Bd2 Qb5

657
23.Ng5!!ƒ and Black faces some very serious pressure from White.]
16...Rcd8 (16...Nxe5 17.dxe5 Be7 18.Be3 Qc7 19.Bxh7+! Kxh7 20.Qd3+ Kg8 21.Qxd7±; 16...Be8 17.Be3 The
game becomes quite messy. Personally I would prefer not to place White’s dark-squared bishop on the e3-square, as
it blocks White’s heavy pieces along the e-file, but in this case the bishop also hits Black’s queen on b6.) 17.Nc4

17...Qa6 (17...Qc7? 18.Qe4 g6 19.Bf4 Qc8 20.Nd6 Qa8 21.d5!±) 18.Qe4 g6 19.Bh6 Bg7 20.Bg5!² Forcing Black to
make some structural concessions. White’s chances should be preferable in the ensuing battle, for example 20...b5
21.Nd6 Rb8 22.d5 exd5 23.Qxd5 Nb4 24.Qc5 Nxd3 25.Rxd3 Qxa2

658
26.Rd2! (26.Be7 Qxb2 27.Bxf8 Rxf8 28.Qxa7 Be6„) 26...Qa5 27.Qe3 Ra8 28.h4! h5 29.Be7 Be6 30.Qg5².

15.Nc3!

White’s knight returns to the original position, but the changes are in White’s favor, since Black’s d7-bishop is not
ideally placed and will become a potential tactical target.
15.Neg5 h6 16.Ne4

White has managed to provoke Black into the weakening ...h7–h6, but at the cost of two tempi which Black now
uses to equalize fully. 16...Ncb4 17.Bb1 Nc2! 18.Rf1 a6 The threat of ...Bd7–b5 forces White to undertake massive
exchanges which diminish his attacking potential. 19.Nc3 Nxc3 20.Bxc3 Na3 21.Bd3 Nb5„

15...Nf6?!

15...Bf6 was a logical continuation.

659
16.Ne5!
[16.Qe4?! g6 17.Bh6 Re8 18.Nxd5 (18.Ne5 Nce7! 19.Qf3 Bc6 20.Ne4 Nf5= Black has managed to untangle his
minor pieces and both sides can play for an advantage.) 18...exd5 19.Qf4 (19.Qxd5 Rxe1+ 20.Rxe1 Be6 21.Qxd8+
Rxd8=) 19...Rxe1+ 20.Rxe1 Bh8=]
16...g6 (16...Nxd4? is impossible due to tactics: 17.Qe4 Nf5 18.Nxd7 Qxd7 19.Nxd5+–; 16...Nce7 17.Ne4 Bc6
18.Qg4 Nf5 19.Nxf6+ Nxf6 20.Qh3

and White achieves one of his goals; he gets a strong pair of bishops, which in turn allows him to create strong
pressure on the black king. White’s centralized e5-knight and the absence of Black’s dark-squared bishop allow for
many potential tactics. 20...h6 21.Bb4 Ne7 22.Bb1 a5 23.Bd2 Nf5 24.Bxf5 exf5 25.Nxc6 Rxc6 26.Qxf5 b6²) 17.Qg4
Nce7 18.Ne4 Bg7 19.Qh3 All the white pieces are strongly coordinated.

660
19...Nf5 A thematic move for Black in similar structures. The f5-knight blocks White’s d3-bishop and creates
pressure on the d4-pawn. (19...Bc6? 20.Ng5 h6 21.Nxe6!+–) 20.Ng5 h6 21.Ngf3 The threat of g2–g4 forces Black to
push his h6-pawn further. 21...h5 22.g4!? Since Black has set up a strong defense around his king, piece
maneuvering alone is no longer enough. (22.Ng5?! gives Black too much time: 22...Bf6 23.Ngf3 Kg7=) 22...hxg4
23.Nxg4∞

The position becomes extremely unbalanced, but I believe that White has superior chances due to the black king
being more vulnerable.

16.a3!

661
Around this point in the game I was running low on time as well, since my preparation essentially ended after
Black’s 12...Bd7, and I had only general plans of development. 16.a3 is quite thematic in these structures as it takes
away the b4-square from Black’s c6-knight, which is important, as that knight usually goes to d5 via b4. Now Black
is stuck with a bad c6-knight, as it hinders the deployment of the d7-bishop to c6. Having said all that, however,
16.Bb1 was a viable alternative, starting to prepare the usual Bb1 + Qd3 battery against Black’s kingside.
16.Bb1 Nb4 (16...Qb6? 17.Bg5 Suddenly Black is in big trouble. 17...h6 18.Bh4 Rfd8 19.d5+– and the pawn cannot
be recaptured.) 17.Ne5 Bc6 18.a3 Nbd5 19.Qd3² is a standard way for White to play in such positions. In general, if
White manages to achieve such a set-up, his position is considered superior because of his attacking chances on the
kingside.

16...Qc7?!

16...Qb6 This would have been a more logical alternative. Black needs to place his f8-rook on the d8-square to
prepare the e8-square as a retreat for the d7-bishop, where it would help defend the kingside and not interfere with
the coordination of the pieces. 17.Bg5 h6 The white g5-bishop is very powerful, so it is logical for Black to take the
g5-square under control. However this gives White the possibility of sacrificing on h6 in the near future. 18.Be3!
[18.Bh4!? leads to a very interesting line, but is objectively not as strong. 18...Nh5

662
Thematic. If Black manages to exchange the dark-squared bishops, then White’s attack on the kingside will not be as
dangerous. However Black has to watch out for tactics, for which, it seems, White is better prepared. I would
recommend the reader to have a look at this and calculate some lines on his own to acquaint himself with the
possibilities. 19.Bg3!ƒ (19.d5 Nf4 20.Qe4 Nxd3 21.Rxd3 Bxh4 22.Qxh4 exd5 23.Nxd5 Qxb2 24.Rd2 Qxa3 25.Nf6+
gxf6 26.Rxd7 Rfe8 27.Rf1 Qf8 28.Rxb7 Qg7=) 19...Nxg3 (19...Bf6 is less accurate: 20.Ne4²) 20.hxg3 Bf6 21.Ne4!
Bxd4 22.Bb1

22...Rfd8 (22...e5 23.Nd6 Bg4 24.Nxc8 Rxc8 25.b4 g6 26.Qd2 Rd8 27.Rc1± and despite Black’s strong-looking d4-
bishop, White retains both the initiative and a material advantage.) 23.Nd6! Rc7 24.Qc2 g6

663
25.Nc4 (25.Nxf7 Ne5! leaves Black with good counterplay.) 25...Qc5 26.b4 Bxf2+ 27.Kf1 Nxb4 28.axb4 Qxb4
29.Qxf2 Qxc4+ 30.Kg1 Qc5 31.Rd4 Rdc8 32.Kh2! Be8 33.Qd2 Kg7 34.Rh4 g5 35.Rd4‚]
18...Qc7 (18...Nd5 19.Nxd5 exd5 20.Bb1 Rfe8 21.Ne5!± The white bishops are far superior to their black
counterparts, and with the upcoming f2–f4 and Qe2–h5 White has a very strong attack.) 19.Qd2!

Threatening to take on h6, which forces Black to give up control of the g5-square and weaken his kingside even
further. White’s initiative starts to look more and more dangerous. 19...h5 20.Bg5 Qd8 21.Qf4±

17.Bg5±

664
White has achieved a dream set-up. Black’s pieces lack coordination and any counterplay, while White prepares for
infiltration on the kingside and a possible d4–d5 break.

17...Qa5?!

In time trouble Black plays this move in order to provoke White into forced lines.
17...Rfe8 This logical move was preferable as White still needs to demonstrate what his spatial advantage and
superior piece coordination will bring. 18.Bb1 Nd5 (18...g6?! 19.Ba2 Rcd8 20.h3 Bc8 21.d5!

White is finally ready for this breakthrough. Black’s position is lost as he cannot capture on d5. 21...exd5 22.Bxf6
Bxf6 23.Qxe8+ Rxe8 24.Rxe8+ Kg7 25.Nxd5 Qd7 26.Rg8+ Kh6 27.Nh2+–) 19.Qe4 g6 20.Qh4 Qd8 21.Ba2!±
White starts to exert greater pressure on the black position and Black is forced to make serious structural
concessions. White’s light-squared bishop positions itself on the a2-square to prepare for the eventual d4–d5 break,
which will further open the a2–g8 diagonal, allowing this bishop to start hitting Black’s weak point, the f7-pawn,
creating numerous threats. 21...Nxc3 22.bxc3 h5 23.h3! An important move, creating a luft for the king and also
preparing for the potential g2–g4 break that will open the h-file for the white queen. 23...Kg7 24.Qf4 Qc7 25.Bh6+
665
Kh7 26.Qd2

26...Bf6 (26...Na5? Black’s c6-knight is still tied to the control of the e5-square. 27.Ne5‚) 27.Bf4 Qa5 28.Be5!± Qd8
(28...Bxe5 29.dxe5 Re7 30.Ng5+ Kg8 31.Qf4 Qxa3 32.Ne4 Kg7 33.Qf6+ Kg8 34.Bb1 Qb2 35.Qg5 Kg7 36.Rd2
Qb6 37.Nf6+– with a decisive infiltration on the dark-squares.) 29.Bxf6 Qxf6 30.Ng5+ Kg7 31.c4!

Finally, after all the maneuvering, White is ready for active pawn moves and his last move opens the a1–h8 diagonal
for the use of his queen, since its black counterpart will be forced off the same diagonal, leaving Black’s monarch
particularly vulnerable. In addition, White’s c-pawn is going to the c5-square, creating a great outpost on the d6-
square for the white knight. The game is strategically won for White. 31...e5 ( 31...Rcd8 32.Ne4 Qe7 33.c5 Bc8
34.Qc3 f6 35.d5 exd5 36.Nd6+–) 32.Ne4 Qe7 33.dxe5 Be6 34.Nd6+–

18.d5?!

An obvious and quite flashy move, but possibly not the strongest. White should have followed the adage ‘the threat
is stronger than its execution’.
18.Bb1!
666
Preparing the d4–d5 break might have been even stronger. Now White’s d1-rook is participating in the attack.
18...Rfe8 (18...Rfd8 19.d5! exd5 20.Bxf6 Bxf6 21.Nxd5 Bd4 22.Nxd4 Qxd5 23.Nf5 Qe6 24.Qh5 Qf6 25.Rxd7 Rxd7
26.Ne7+ Kf8 27.Nxc8 h6 28.Qc5++–) 19.Ne5 Nxe5 (19...Rcd8 20.Bxf6 Bxf6 21.Bxh7+ Kf8 22.Rd3!+–) 20.dxe5
Nd5 21.Bxe7

21...Rxe7 (21...Nxc3 22.Qd3 Nxb1 23.Qxd7+–) 22.Qd3 f5 (22...g6 23.Ne4! Kg7 24.Nd6 Rc7 25.Qf3± and Black’s
weak dark-squares give White a very strong attack.) 23.Nxd5 exd5 24.Qxd5+ Qxd5 25.Rxd5 Be6 26.Rd4±

667
White has a healthy extra pawn, guaranteeing a big advantage which should be sufficient for a win.

18...exd5 19.Bxf6 Bxf6 20.Bxh7+ Kxh7 21.Rxd5

Black’s d7-bishop cannot be protected by 21...Qc7 in view of 22.Qd3 with a double attack.

21...Bxc3!?

This is the position that Black was going for with 17...Qa5. The following moves are forced. 21...Qc7 22.Qd3+ Kg8
23.Rxd7 Rcd8 24.Rxc7 Rxd3 25.Rxb7 Bxc3 26.bxc3 Rxc3 27.Rc7±

22.Rxa5 Bxa5 23.b4 Kg8 24.bxa5

668
24...Bg4!?

Black wants to set up a fortress and eliminates White’s last minor piece, however with such limited time his defense
is not easy.
a) 24...a6 Black has no time to prevent White from undermining the c6-knight. 25.Qd3 Rcd8 26.Ng5+– is an easy
win for White.
b) 24...Rfe8 25.Qd3 Rxe1+ 26.Nxe1 Be6 27.a6!± Black has insufficient material compensation for the queen and his
position should be lost.

25.a6!

I thought this move was very important, to undermine Black’s c6-knight and therefore diminish any possibility of a
fortress on the queenside for Black.

669
25...bxa6 26.Qe4!

An important finesse which allows White to keep his kingside pawn structure intact. 26.Qxa6 Bxf3 27.gxf3 Nd4
28.Kg2 Ne6 29.Qxa7 Ra8 30.Qe3 Ra4 31.Qc3 Rfa8 32.Ra1 g6² is the kind of fortress Black wanted to set up.

26...Bxf3 27.Qxf3 Rfe8 28.Ra1!!

I need to keep the rooks on to create threats to the black king. It’s possible that without rooks the position is a win
for White, but the conversion would be more difficult.

28...Re6 29.h3+–

Hereabouts I realized that the position is very close to a technical win. I need to combine slowly improving the
position of my pieces with creating threats to the black king.

29...Rd8 30.Qc3 Rdd6 31.Rb1

Black can only watch as he has no targets for his pieces. He mustn’t allow the white rook to infiltrate his camp or his
king will become vulnerable to a mating attack.

31...Rd7 32.Qc4!

670
This move forces Black to give up control of the b5-square, allowing my rook to swing to the kingside.

32...a5

32...Nd8!? is an interesting computer suggestion, trying to close the eighth rank while keeping the a6-pawn intact.
33.Rb8 g6 34.Kh2 Red6 White’s plan involves the eventual breakthrough with f4–f5 to get to the black king. 35.Ra8
Kg7 36.Qe4 Kf8 37.f4 Rd4 38.Qe5 R4d5 39.Qe3 a5 40.Rxa7 Rxa7 41.Qxa7

41...Rf5 (41...Kg7 42.Qa8+–) 42.Qb8 Ke8 43.g4 Rc5 44.a4 Kd7 45.Qa7+ Kd6 46.Kg3 Rd5 47.h4+– Black’s d8-
knight is tied to the defense of the f7-pawn, thus leaving him with only rook moves. White eventually forces
zugzwang: 47...Rc5 48.Kf2 Rc2+ 49.Ke3 Rc3+ 50.Kd2 Rc5 51.Kd3 Rd5+ 52.Kc4

671
Zugzwang is reached. Black loses one of his pawns, and with it, the game. 52...Nc6 53.Qxf7 Rd4+ 54.Kc3+–

33.Rb5 Rd1+ 34.Kh2 Rd2

35.Rf5

35.Rc5 would have been more accurate.

35...Rd4 36.Qc3 Rdd6 37.Rc5 Rf6 38.Rc4

An immediate 38.f4 is also possible, but I wanted to avoid having to play such drastic pawn moves before I was
absolutely sure that I had to.
38.f4 g6 39.g4 Kh7 40.f5 gxf5 41.g5 Rfe6 42.Rxf5 Kg6 43.Rf2 Ne5 44.Qxa5 Rd3 45.Kg2±

38...Rfe6 39.Rc5 Rf6 40.Qe3 Rfe6 41.Qg3 Rg6 42.Qb3!

672
Finally deciding to transfer the queen to the c7-square, where it will completely tie Black’s c6-knight to the defense
of the queenside pawns and thus not be able to take part in the defense of Black’s king.

42...Rgf6 43.Qb7 Rfe6 44.Qc7 Rf6 45.f4!

Finally. Black will soon be forced to abandon his fortress of doubled rooks on the sixth rank, due to the the well-
established zugzwang method.

45...g6 46.f5 gxf5 47.Rxf5 Rde6 48.Rh5 Rh6 49.Qg3+ Kf8 50.Rd5 Rhg6 51.Qf2 Rgf6 52.Qb2

Some of White’s last moves were intended to prepare to land his rook on the f-file without Black being able to play
an immediate ...Rf6, offering the rook exchange. I felt that once I managed to double my heavy pieces on the f-file to
attack Black’s f7-pawn, his sixth rank defense would collapse.

52...Ke7 53.Rh5 Rh6

673
Preventing White’s rook from reaching the h8/c8/c7-squares.

54.Rb5 Rhf6 55.Qc3 Kf8

a) 55...Rh6 56.Qg7+– The point of the maneuvering on the previous moves. White wants to infiltrate the kingside
with his queen.
b) 55...Rd6 This would have put up stronger resistance, forcing White to utilize his last asset, the h-pawn.

56.h4 Kf8 57.h5 Rh6 58.Rg5 Rhf6 59.Qc5 Ke8 60.Rg8+ Ke7 61.Qb5 Rd4 62.Qg5 Rdd6 63.h6+–

56.Rh5 Rh6

56...Ke7 57.Rh8+–

57.Rf5!

White achieves his desired set-up.

674
57...Rhg6

57...Ke7 58.Qg7 Ref6 ( 58...Rhf6 59.Qg5!+–) 59.Rb5 Rhg6 60.Rb7+ Ke6 61.Qf8 Rf2 62.Qe8+ Kd5 63.Rd7++–

58.Qf3+–

Finally Black’s defense based on control of the f6-square is crushed and the white pieces infiltrate the black camp.

58...Rg7 59.Qf4 Kg8 60.Qc7

60.Rxf7! would have been even more decisive: 60...Rxf7 61.Qg4++–.

60...Kf8 61.Qc8+ Ke7 62.Rd5 Kf6 63.Qh8 Re4 64.Rh5 Ne7 65.Rh7

Black resigned. 1–0

This win gave me a measure of hope, but in the next eight games I suffered four losses, including two devastating
losses with White. This effectively ended my hopes, since there were only sixteen games. Of course at that point I
recalled the famous Karpov vs. Kasparov match in 1984, where Karpov had led 5–0 at one stage. Furthermore, when
Kasparov finally won three games, the match was infamously stopped by the FIDE President at the time,
Campomanes. A repetition of such an experience in my match was impossible. Karpov had learned his lessons.

675
 GAME 22
Gata Kamsky
Anatoly Karpov
Elista 1996
This was the tenth game of the match. At this point I had lost five games and won only one. My opponent’s
preparation was vastly superior and objectively there was no hope, since there were only sixteen games in the match.
Thus the decision was made to attempt to steer away from any possible preparation by my opponent at all costs and
just play chess, hence the choice of opening in this game. Up until this point I had always allowed the Nimzo-Indian
Defense as White, going for isolated pawn positions, but after my loss in the fourth game, I no longer felt confident
in my ability to play the resulting positions.

1.d4 Nf6 2.c4 e6 3.Nf3

The first surprise. The game now transposes into the Queen’s Indian Defense. Obviously my opponent possessed
vastly superior experience of playing this opening, however I felt that wasting my opponent’s time and energy on
preparation that did not happen would give me a certain psychological advantage. I had played this line against
Karpov earlier in 1995, in the Amber Rapid event, however Black achieved an advantage in the opening and I lost
the game. It was difficult to assess whether Karpov’s team would prepare for a repeat of that line, given that my
main weapon for White was 1.e4.

3...b6 4.a3

The Petrosian Variation. The point of this move is to prevent the development of Black’s f8-bishop to the b4-square
once and for all. The move also follows the rapid game mentioned above.

4...Bb7

The other main move in this position is 4...Ba6 which leads to a lot of theory.

5.Nc3 d5

One of the main moves which Karpov has employed for many years. Since my opponent’s main weapon for White
is 1.d4, it is neccessary to mention that he had played this line for White on numerous occasions as well and
continued to do so after this match.
676
6.cxd5

Other alternatives for White are 6.Bg5 Be7 7.Qa4, or 7.e3, however Black appears to score well in both lines.

6...Nxd5

Recapturing with the pawn would not make sense, since it would block the b7-bishop, and statistics show a
favorable score for White after both 7.g3 and 7.Bf4. The structure after the move in the game resembles more a
Tarrasch opening, which, in modern times, has been revived by Kramnik, as recently as the Candidates tournament
in Berlin in spring 2018.

7.Bd2

Less popular than 7.Qc2 or 7.e3, but still following our rapid game from 1995. The idea is for White to recapture on
c3 with his bishop from where it will potentially target Black’s kingside along the long diagonal.

7...Nd7 8.Qc2

Both sides are making logical developing moves.

8...c5

9.Nxd5!?

Here is the suprise. The database shows only two games played with this move prior to this one. Our previous game
followed the main line 9.e4 Nxc3 10.Bxc3 cxd4, which is considered to be equal. Black now has to make serious
strategic decisions with regards to the structure.
Both recaptures on d5 are possible, but result in vastly different consequences.

9...exd5

A very prudent decision after considerable thought on my opponent’s part. After 9...Bxd5 Black has to face the
thematic pawn sacrifice 10.e4 Bb7 11.d5!?, leading to an extremely sharp, double-edged, tactical position with
advantage to the side with the better preparation! It is obvious that Karpov’s team had analyzed this line, as
677
demonstrated by his games in later years. However the line has not been played much since that time and lots of
possibilities remain.

10.dxc5!

White must follow up with this exchange, or risk the black pawns rolling on the queenside, creating a potential
structural advantage with a far-advanced passed pawn. The move in the game forces Black to make another
important strategic decision with regards to his structure. Knowing my opponent’s distaste for positions with
structural deficiencies, it was a safe bet that Black would recapture with the pawn, going for a hanging pawns
structure. However, given the awkward position of the white queen on the c2-square and his underdeveloped
kingside, it would have been quite OK for Black to recapture on c5 with his bishop, resulting in a dynamic position
with equal chances.

10...bxc5

As mentioned above, a predictable choice by my opponent. Now, however, Black has to be very attuned to the
dynamic qualities of the position, as White has standard programmed breaks like e2–e4 or b2–b4 which would both
result in Black giving up central blockading squares and easily slipping into an inferior position. However that was
not my plan before the game. My intention was to create tension, similar to the ideas in the Polugaevsky Variation of
the Grünfeld, where White leaves his king in the center and combines threats to the black king with breakthroughs in
the center. While Black has no problems in the opening, the position as a whole retains a lot of dynamic possibilities,
exactly what I needed given the overall score in the match.
10...Bxc5 11.e3 Rc8 12.Bc3 0-0 13.Qf5 leads to complicated, roughly balanced positions.

11.e3

a) 11.g3 This move is a viable alternative leading to a more positional game. 11...Be7 12.Bg2 0-0 13.0-0 Re8 14.e3
Nf6 (14...Bf6 15.Rac1 Rc8 16.Rfd1 Qb6 17.b4 cxb4 18.Qa4²) 15.Rfd1 Ne4 16.Bc3

678
Black’s e4-knight is a very strong piece, therefore White invites its exchange, which would be favorable to him, but
Black can simply keep his central knight which will both block White’s g2-bishop and keep pressure on the
queenside. 16...Qb6 17.Nd2 f5 18.Rac1 a5! A standard move in such positions. Black prevents White’s potential b2–
b4 break which would weaken his hanging pawns and reduce their control of major central squares. Black has a
comfortable game, given White’s lack of a clear plan for further improving his position. 19.b3 Qe6 20.Bb2 Rec8!=

21.Bh3?! d4! If Black can achieve this thematic breakthrough, he usually gets the advantage.
b) 11.e4 is another interesting idea that deserves serious analysis. It was first tried in a 2008 game between
Shcherbakov and Riazantsev which Black went on to win after 11...Be7 (11...dxe4 12.Ng5!?∞) 12.exd5 0-0 13.0-0-0

679
13...c4?! (13...Bf6 leads to a very unclear game) 14.Bc3?! (14.h4!² would have been a major improvement for
White.) 14...Nb6!³.

11...Be7

Again Black plays very logical and safe developing moves, the recommended way of playing when faced with a
surprise move by your opponent.
11...Bd6 12.Bd3 Qe7 is a possible improvement as the bishop is more actively placed on d6.

12.Bd3

My general preparation ended here with an evaluation of unclear but playable. My opponent and I were now
entering uncharted territory.

12...g6!

680
My opponent was famous for his Petrosian-like quality of sensing danger in advance. It is clear that White has the
possibility of using his h-pawn as a battering ram if Black castles on the kingside. Given the potential danger there,
and that he cannot hide on the queenside, it becomes logical to leave the king in the center. This, however, creates a
certain lack of coordination among his pieces and, given the hanging pawns structure, the game becomes extremely
unbalanced. Furthermore, Black must be ready for either of White’s b2–b4 or h2–h4 breaks, or even e3–e4, all of
which lead to completely different scenarios. At this point, the price of each move becomes very high for both sides.

13.h4

A useful move. White makes sure that Black’s kingside castling is taken out of the equation altogether.
13.b4!?

This move is viable, but reveals White’s hand too early and eliminates the tension, which is more important than
fixing the structure. Black has a choice here. 13...0-0 Black castles and completes his development. (13...c4?! cannot
be recommended as it gives White a great outpost on the d4-square and weakens the dark squares. 14.Be2 Nf6
15.Bc3 0-0 16.Nd2²) 14.bxc5 Nxc5 15.0-0 Rc8 16.Be2 Bf6 17.Rac1 Ne4 18.Qd3 Nxd2 19.Qxd2 Qd6=

681
Black’s strong dark-squared bishop compensates fully for the isolated d5-pawn. The game is likely to simplify
further into an equal endgame, e.g. 20.Qa5 d4!= A strong move that leads to massive exchanges and equalizes
completely. 21.Nxd4 (21.exd4 Bxf3 22.Bxf3 Bxd4³; 21.Rcd1 Bxf3 22.Bxf3 Be5 23.g3 Qf6=; 21.Qxa7? d3!µ)
21...Bxd4 22.exd4 Qxd4 23.Rxc8 Rxc8 24.Rd1 Qc5 25.Qb4 Qxb4 26.axb4 Rc2 27.Bf1 Bc8=.

13...Qb6

A move that is often played in similar structures as Black’s queen is ideally placed on the b6-square, where it
supports a potential ...d5–d4 central break, protects the c5-pawn and keeps an eye on White’s queenside pawns.
Black also attempts to develop his pieces to more active squares, and can even potentially castle queenside. White
doesn’t have much choice but to continue with his idea.

14.h5 Bf6

14...0-0 15.Kf1!

An important move. Prior to the potential opening of the center with ...d5–d4, White moves his king to the safety of
682
the kingside. The rook actually plays an important role on h1 and is better placed there than on f1. 15...c4! This is
probably Black’s best plan. White’s d3-bishop is too strong and, despite giving White access to the d4-square, Black
now has a clear plan for the development of his pieces. The d7-knight has to go to the e4-square and the e7-bishop to
the f6-square to help cover the weak dark squares around Black’s monarch. White’s lack of space for maneuvering
and king on f1 do not allow much hope for achieving an advantage. However such a decision would still make any
elite player pause, as voluntarily fixing your own structure on an inferior location is something that goes against the
very basis of chess knowledge. Clearly, the dynamic of chess development is a balancing act that must be carefully
evaluated.
[15...a5 16.Rd1 Ba6 17.Bc3 Bf6 18.Kg1

White has finished his development and Black’s hanging pawns require careful defense. 18...Bxd3 19.Qxd3 Rfd8
20.hxg6 hxg6 21.Qc2 Bxc3 22.Qxc3

22...Qf6 (22...Nf6 23.Rh4 Rab8 24.Rd2² and given Black’s unsafe king, White’s chances are slightly preferable.)
23.Qc2 Rab8 24.Rxd5 Rxb2 25.Qd3 Rdb8 26.Rh4 with double-edged play.]
16.Be2 Nc5
683
17.Nd4! (17.Bc3?! Ne4 18.Bd4 Qe6= As a result of White’s bishop taking the important blockading d4-square from
his f3-knight, Black’s e4-knight is perfectly placed to negate any white attacking chances on the kingside.) 17...Bf6
18.Bc3 Ne4 19.Bf3 Nxc3 20.bxc3 Rab8 21.g3 Bxd4 22.exd4

22...Bc6= The game seems balanced, with Black’s control of the b-file compensating for White’s pressure on the d5-
pawn.

15.Rb1

684
Fortunately for me, it is much easier to play this position for White than for Black.

15...Rc8

Another logical prophylactic move. Black guards against a possible b2–b4 break, as long as the white queen remains
on c2. Black is now ready to play ...c4 and castle kingside, so I had to take action.

16.Qa4!?

This move attempts to achieve several things. One, White pins the d7-knight, temporarily preventing castling. Two,
he prevents the potential exchange of light-squared bishops on the a6-square. Three, he vacates the c2-square so the
d3-bishop can remain on this important diagonal in case Black pushes ...c5–c4. The fourth, and most important, idea
is to transfer the queen from its rather passive position on the c2-square to a more active location on the kingside to
create additional potential threats there. If one move is able to achieve at least two goals, it is considered to be pretty
good, but in this case it pursues four goals simultaneously. The only reason I give it a question mark is to show that,
despite the reasoning, the evaluation of the position does not change overall.
a) 16.Bc3?! d4³ is the justification for Black’s last move.
b) 16.Kf1! would have been preferable.

685
White intends to play Bd2–c3 and Rb1–d1, exerting pressure on Black’s hanging pawns, which, along with the
exchange of the dark-squared bishops, would give him some chances for the advantage. 16...c4
[16...Qe6 17.Bc3 Ba6 should also be survivable for Black. (17...0-0? 18.hxg6+–; 17...d4? 18.exd4 cxd4 19.Re1+–)]
17.Be2 0-0 18.Bc3 Nc5 19.Kg1 Ne4 20.Bd4 Qd6

The game remains complicated, now that Black has managed to safeguard his kingside.

16...Bc6?!

While this move is not bad, it is inaccurate. Usually in such structures the ideal location for Black’s queen is on the
e6-square, from where it exerts pressure everywhere in the center, but in this particular case it would also deny the
white queen the g4-square, from where it would exert pressure on Black’s kingside and keep the d7-knight under
attack.
16...Qe6! The move that could have turned the game. Black most likely didn’t want to sacrifice his a7-pawn for
nebulous compensation, however, if taken, White’s attacking potential would have become zero and Black would
686
have had sufficient compensation to equalize.

17.Qxa7
a) 17.Kf1 0-0 18.Kg1 (18.Bc3? Inviting Black to play ...d5–d4 is suicidal for White. 18...d4!µ 19.exd4 cxd4
20.Bxd4 Bc6–+; 18.Qf4 Ne5 19.Nxe5 Bxe5 20.Qh4 Ba6 21.Bxa6 Qxa6+ 22.Kg1 Bf6 23.Qg4 Rb8³ and the white
attack is blunted.) 18...d4³
b) 17.hxg6!? fxg6 (17...hxg6 18.Rxh8+ Bxh8 19.Qxa7 Bc6 20.b4 d4

21.bxc5 Bxf3 22.gxf3 Nxc5 23.Bb5+ Kf8 24.Bb4 Kg8 25.Bxc5 dxe3 26.Bxe3 Bc3+ 27.Kf1 Qf5 28.Rd1! — if
28.Rb3, Black takes on f3 and is fine — 28...Qxb5+ 29.Kg2±) 18.b4 (18.Qxa7 Bc6 19.Be2 0-0ƒ is hardly
recommendable for White; 18.Qf4 0-0 19.0-0 Ne5³) 18...c4∞
17...Bc6 (17...Qc6 18.Qa5 Ra8 19.Qb5 0-0!„) 18.b4

687
18...Bd8!! Probably the move we both missed. The queen is trapped on a7. 19.bxc5 Ra8–+
Note that if Black had inserted ...c4 before playing ...Bc6, then White would have been better playing Qb4.

17.Qg4!?

I decided to continue playing in the same risky manner with the intention of confusing my opponent, who was
rapidly approaching serious time trouble. Returning with my queen to the d1-square, resulting in a repetition, was
not an option.

17...Bb5!?

Black attempts to take over the a6–f1 diagonal, which results in the white king being unable to escape to the
kingside.
17...Qb3! hitting White’s strongest piece, the d3-bishop, seems a stronger option for Black. 18.Be2 (18.Ba6 Rb8µ)
18...Rb8!
688
Without the d3-bishop, White has no attacking chances, resulting in a position where Black would threaten to
assume control of the game. Having said that, White can now suddenly change his game plan completely. (18...Bb5
19.Bxb5 Qxb5 20.Rd1 Qxb2 21.e4 Qb7 22.exd5 0-0) 19.0-0! White now has to switch to defense to equalize, or risk
his queenside being overrun. (19.hxg6? hxg6 20.Rxh8+ Bxh8 21.Kf1 Qa2 22.Rd1 Rxb2–+) 19...0-0
[19...Bxb2 This would be too risky for Black. 20.Bc1 Qc3 (20...Qxa3 21.Rxb2 Rxb2

22.Nd4!!+– Turning the tables by taking advantage of Black’s weak dark squares and lack of piece coordination.
22...Ba4 23.Nb5+– White gains a decisive material advantage. 20...Qc2 21.Rxb2 Rxb2 22.Nd4! cxd4 23.Qxd4 Rb6
24.Bd3 Qb3 25.Qxh8+±) 21.Qh4!° and Black has to be careful.]
20.Rfc1!

689
20...Bxb2!? Risky, but possible. (20...Bb5 21.Bxb5 Qxb5 22.e4=; 20...Rfe8 21.Bb4!²) 21.Bc3! Tactics and the
slightly open black king give White a way to equalize. 21...Qxa3 (21...Bxc1? 22.Rxb3 Rxb3 23.hxg6 hxg6 24.Bd3
Rxc3 25.Bxg6±) 22.Rxb2 Rxb2 23.hxg6 hxg6 24.Qh4

24...f6 (24...d4 25.Bxd4 cxd4 26.Rxc6∞) 25.Qh6 with compensation, e.g. 25...Kf7 26.Qh7+ Ke8 27.Nd4!?
(27.Qxg6+ Ke7 28.Bxb2 Qxb2 29.Re1 Ne5=) 27...Rb6 (27...cxd4 28.Bxb2 Qxb2 29.Rxc6± with a dangerous
attack.) 28.Qxg6+ Ke7 29.Nxc6+ Rxc6 30.Qh7+ Rf7 31.Qc2

690
31...Rb6 (31...Nb6? 32.Ra1+–) 32.Bf3 Qb3 33.Qf5= and White has sufficient compensation.

18.Bc2

Of course White keeps his lightsquared bishop. At this point I was getting quite optimistic about my chances, as
Black does not have any forced way to eliminate the tension.

18...a5

A general purpose move.

18...Ba6!? Attempting to set up a Ba6/Qb5 battery against the white king was an alternative. 19.Bc3!! A spectacular
move, especially with the following idea in mind. I have to say I was very pleased when I saw this idea.
[19.Ba4 Bb5 20.Bc2 (20.Bc3 d4!³) 20...Ba6= is a repetition.]
19...Qb5 20.Kd2!! The white king escapes to the queenside, where it would be relatively safe, while the black king
would start to feel uncomfortable, since the white rooks are now connected and Black no longer has any potential
691
...d5–d4 break, since it would open the e-file for White’s artillery.

20...Ke7 21.Rbe1!‚ White has now covered all the weak squares around his king and is ready to take aggressive
measures on the kingside. 21...Qb6! 22.Bxf6+ Nxf6 (22...Qxf6? 23.e4!± and White has successfully opened the
center. 23...d4 24.e5 Qe6 25.Qg5++–) 23.Qg5 Qxb2

24.hxg6! (24.Rb1 h6! 25.Qf4 Ne4+ 26.Qxe4+ dxe4 27.Rxb2 exf3 28.gxf3 g5=) 24...hxg6 (24...fxg6 25.Rb1+–; My
editor, Romain Edouard, pointed out that Black may have an unbelievable saving move: 24...Ke6!! Black threatens
...Ne4+! E.g. 25.Qf5+ Ke7 26.Rb1!? Ne4+ 27.Qxe4+ dxe4 28.Rxb2 exf3 29.gxh7 fxg2 30.Rg1 Rcd8+ 31.Kc3 Bf1=)
25.Rxh8 Rxh8 26.Rb1 Qxa3 27.Qe5++–

19.hxg6?!

Given how complex and unbalanced the position is, I was not sure how to proceed. I decided to relieve some of the
tension in an attempt to improve the position of my pieces.
19.e4?!

692
A very interesting move. While it looks incredible, considering both kings are in the center and the white king is
especially open due to Black’s b5-bishop, White attempts to open the position, hoping to exploit Black’s lack of
coordination. He also hopes to bring his heavy pieces to the central files. 19...Rd8! A very strong waiting move.
a) 19...0-0 20.a4 dxe4 (20...Bc6 21.exd5 Bxd5 22.hxg6 hxg6 23.Qxd7+–) 21.hxg6 hxg6 22.Ng5!±
Leading to wild fireworks that look better for White in view of the vulnerable position of the black king.

E.g. 22...Ne5 23.Qh3 Rfd8 24.axb5 Rxd2 25.Qxc8+ Rd8 26.Qa6 Qxa6 27.bxa6 Bxg5 28.Bxe4+–.
b) 19...dxe4?! 20.Ng5!! The move I missed. White keeps his queen on the g4-square, pinning the d7-knight and
denying Black the chance to castle kingside. Opening the center favors White as he has more control there and
potential access to the central files with his heavy pieces. Black has to be really careful here. 20...Bxg5 21.Qxg5 Qe6
22.Bc3 (22.Qf4!? Ne5 23.Bc3 Nd3+ 24.Bxd3 exd3+ 25.Kf1 Rg8 26.Kg1ƒ; 22.Bxa5 Ne5 23.Bc3² keeps a small edge
for White.) 22...Rg8 23.Rd1! Preparing for the white king to run to the queenside.

693
White’s bishop pair is extremely strong and, given the lack of coordination among Black’s heavy pieces, White’s
chances are superior. 23...Bd3 24.Ba4 Kf8 25.Rh3ƒ
20.Rd1!?

(20.exd5 0-0„; 20.Bg5 h6³) 20...d4!

[20...0-0 21.hxg6 (21.Qh3 Rfe8 22.hxg6 hxg6 23.Qh7+ Kf8 24.Bh6+ Ke7–+) 21...hxg6 22.Bg5! White needs to
eliminate Black’s defensive f6-bishop. The fireworks which now follow are hard to evaluate. It is very possible that
the following moves can be improved upon. 22...dxe4 23.Qh4 Rfe8 24.Nd2 e3 (24...Bg7!? 25.Bxd8 Rxd8 26.a4!
Bd3 27.Bxd3 exd3 28.0-0²) 25.Ne4 exf2+ 26.Kxf2 c4+

694
27.Ke2!! Quite unbelievable. The white king walks bravely alone under the full gaze of Black’s entire army — and
is OK!! (27.Kg3 Be5+µ) 27...c3+ 28.Kf3 Rxe4!! (28...Ne5+ 29.Kg3+–) 29.Bxe4 Ne5+ 30.Kg3 Rxd1 31.Rxd1 Bxg5
32.Qxg5 f6 33.Qh6! Bc4 34.bxc3 Qc7! 35.Qf4

and there is still a lot of play left.]

21.Bf4!

695
(21.Bg5?! h6µ) 21...c4
[21...Bc4 22.hxg6 hxg6 23.Rxh8+ Bxh8 24.Ba4 Bb3 (24...Bb5 25.Bxb5 Qxb5 26.Qh3 Bg7 27.Qh7 Bf6 28.e5 Nf8
29.Qg8 Be7 30.e6 fxe6 31.Bh6!ƒ with a continuing attack.) 25.Bxb3 Qxb3

26.Bd6! (26.e5 Nf8!³) 26...Nf6 (26...Qxb2 27.e5+–) 27.Qf4 Nh5 28.Qh2 d3 (28...Qxb2 29.g4∞) 29.Be5 Qe6 30.Qh3
Qe7!∞ This quiet retreat helps Black to maintain his initiative in the center, as transposing into the endgame would
make the game preferable for White because of Black’s weak d3-pawn.]
22.e5 Both sides are using their trumps and the position becomes very unbalanced. 22...d3 23.Bb1 (23.exf6? dxc2
24.Rc1 c3–+) 23...Be7 24.Ba2!∞

696
and in this very unbalanced position, it’s hard to say who’s better.

19...hxg6 20.Rxh8+ Bxh8

21.Ng5

At this point I decided on a very concrete plan of going for one of the central breakthroughs. There was an
alternative, 21.Bc3, but I didn’t believe in it, since it is clear that after Black’s next move, White will have to go on
to the defensive. 21...d4! 22.exd4

697
22...f5! An important intermediate move. (22...cxd4 23.Bxd4 Bxd4 24.Nxd4 Kf8 25.Rd1!?² It seems that White
manages to keep both his king safe and his extra pawn.) 23.Qh3™ (23.Qf4? cxd4 24.Bxd4 Qe6+µ) 23...Qe6+
24.Kd2 Qe2+ 25.Kc1 Qf1+ 26.Bd1 Bf6!

I don’t believe that Black is worse here, given how bad White’s piece coordination is.

21...Bf6?!

An inaccuracy; Black overlooks a small but important tactical motif.

21...Rd8!

698
This seems an important move in this type of position. Black gets out of the X-ray of the white queen on g4 and
prepares the ...d5–d4 push in the center. The following sequence looks forced. 22.f4! Preparing the flight of the king
to the g1-square via f2. 22...Bg7 (22...d4 23.Kf2 c4 24.Qh3 Bf6 25.exd4 Bxd4+ 26.Kg3 Bf2+ 27.Kh2 Nf6 28.Bc3∞)
23.Kf2 Nf6 24.Qh4 Kf8 25.f5 gxf5 26.Bxf5 Re8 27.Kg1 Re5∞

The game remains unclear, but obviously Black has achieved a comfortable degree of coordination among his pieces
and should be not worse.

22.a4!?

699
Now I had seen the following line and decided to go for it, hoping that my calculation was good enough. White
needs to get his king to safety before he can open up the center.

22...Bc6

Black is forced to take his bishop off the a6–f1 diagonal, allowing the white king to transfer to the kingside, opening
the way for his b1-rook.
22...Ba6? 23.Nxf7! Black not only loses the important f7-pawn, but is unable to exploit White’s massively
uncoordinated pieces. 23...Ke7 24.Qxg6 Rf8 25.Ng5± leads to a serious advantage for White.

23.Kf1

White is now ready for the e3–e4 push. Suddenly, with very limited time, Black is faced with no longer potential,
but very real, threats to his king.

23...Ke7!

Black plays his best defense! He wants to bring his c8-rook into the game.

24.e4

White has to go for this central break, otherwise Black threatens to take over the initiative.

24...Qa6+?

A serious inaccuracy after which the black defenses come perilously close to failing. The idea is to protect the c8-
rook with tempo to free up the d7-knight to move to the center, followed by closing the center with ...d5–d4.
However the black queen is misplaced on the a6-square and this gives White tactical possibilities.
a) 24...Rh8!

700
25.Kg1!?
[25.exd5 Bxd5 (25...Rh1+? 26.Ke2 Rxb1 27.Qe4+! Ne5 28.Bxb1± Bd7 29.f4 Bf5 30.Qe3 Bxb1 31.fxe5 Bxg5
32.Qxg5+ Kd7 33.d6± leads to an almost winning position for White.) 26.Kg1 Qb8!

Not the only defensive resource. Black’s position remains quite solid. 27.Bf4 (27.Re1+ Kf8 28.Bf4 Qb4!³) 27...Be5
28.Re1 Qb4! 29.Re3 Kf8 30.Rxe5 Nxe5 31.Qc8+ Ke7 32.Qc7+ Nd7 33.Bd6+ Kf6 34.Be5+ Kxg5 35.Bc3 Qg4!µ]
25...d4!

701
This move allows the closing of the center, since after 22.a4 White no longer has the option of blowing open the
position on the queenside via the b2–b4 break. The black position is now favorable. 26.e5!? Nxe5 27.Re1 Qxb2
28.f4 The only chance. 28...Qxc2 29.fxe5 Qxd2 30.exf6+ Kxf6 31.Rf1+ Ke7 32.Rxf7+

32...Kd8! and in this remarkable position with both kings under attack, there’s nothing better for White than
perpetual check. 33.Ne6+ Ke8 34.Nc7+ Kxf7 35.Qe6+ Kg7 36.Qe7+ Kh6 37.Qh4+ Kg7 38.Qe7+=
b) 24...d4? immediately is impossible:

702
25.e5! Qa6+ 26.Kg1 Nxe5 27.Re1!+– and Black loses material.

25.Kg1 Ne5

26.Qf4!

26.Qg3?! would be inaccurate. 26...Qe2!

26...d4 27.Nh7!?

Quite logical. White hits the f6-bishop, Black’s key defender, which holds his entire position together.
27.Re1!?
This slow improving move deserved serious attention as well, simply preparing for Qg3 and rolling the central
pawns with f2–f4. Black would be hard-pressed to defend. 27...Qc4 (27...Rh8 28.Qg3 Qb7 29.f4 Nc4 30.e5 Bxg5
31.Qxg5+ Kf8 32.Qf6!‚) 28.Rc1!ƒ I had not seen this move of course. White keeps switching ideas to throw Black
703
off guard. (28.Bb1?! Qb3 29.Nh7 Bg7 30.Qh4+ Kd7 31.Nf6+ Bxf6 32.Qxf6 Qe6∞) Now 28...Qe2? would be bad
because of 29.Bd1! Qd3 30.Rxc5+–.

27...Bh8?

A serious mistake in Black’s calculation, though it is not easy to see the reason immediately.
27...Bg7! is much stronger.

However at this point my opponent had almost run out of time and in such a situation it is practically impossible to
defend such a complex position. 28.Qh4+ (28.Re1 This quiet intermediate move, with the idea of bringing the bad
white b1-rook into the game and taking the e2-square under control, looks good, however it doesn’t work, since
Black has a forced draw after 28...Rh8! 29.Qh4+ Kd7 30.Nf6+ Ke6= 31.Nh7 Kd7=.) 28...Kd7 (28...f6? loses
immediately: 29.f4+–) 29.Nf6+ (29.f4 Qe2! 30.fxe5 Qxd2 31.Rf1 Qe3+ 32.Kh1

32...Rc7!! 33.Rxf7+ Kc8 34.Rxc7+ Kxc7 35.Qe7+ Kb6 36.Qxg7 Qe1+ 37.Kh2 Qh4+= is only a draw.) 29...Bxf6
(29...Kc7?? 30.Bf4+–; 29...Ke6 30.Nd5 Bxd5 31.exd5+ Kd7 32.Re1 Rh8 33.Qg3 Qd6 34.Bxa5 Qxd5 35.Bd2±
White retains the advantage thanks to his strong pair of bishops and outside passed pawn, along with the presence of
704
queens on the board.) 30.Qxf6 Re8 31.Re1 Qc4! 32.Bd1 Qe6 33.Qg5² White keeps a certain advantage.

28.Qh4+

28...f6?

The losing move. Now Black is defenseless.

28...Kd7!! 29.Ng5! Possibly White’s best attempt to prevent the black king from leaving the center.
a) 29.f4 Qe2 30.fxe5 Qxd2 31.Rf1 Qe3+ 32.Kh1 Re8!

705
This defense is very difficult to find in time trouble, but it gives Black the equality he desired. 33.Rxf7+ Kc8 34.Nf6
Bxf6 35.Qxf6 Qc1+ 36.Kh2 Qh6+ 37.Kg3 Qe3+ 38.Kg4 Qe2+ 39.Kg3 (White has to be satisfied with a draw, since
he loses after 39.Kg5? Qxg2+ 40.Kh6 Qh2+–+.)
b) 29.Qh3+!? Kc7

30.Ng5! The only move to pose problems for Black. White keeps creating threats to Black’s weakened kingside.
(30.f4 d3 31.Bb3 Bd7 32.Qg3 Nc4!∞) 30...Bd7!? (30...Kb8 31.Bf4 Ka7 32.Bxe5 Bxe5 33.Bd3 Qb7 34.Nxf7 Bc7
35.e5 Rf8 36.Bc4²; 30...Kb7 31.f4 Qe2 32.fxe5 Qxd2 33.Nxf7 Kb8 34.Bd3 Bg7 35.e6² with advantage to White.
30...Qe2? 31.Bxa5+ Kb8 32.Bd1!² White gets a material advantage while Black gets serious counterplay, thanks to
his dangerous central pawns.) 31.Qg3 Kb7

706
32.f4 ( 32.Nxf7 Nxf7 33.Qb3+ Ka7 34.Qxf7 Qe6!= Black’s central pawns are so strong that he’s willing to give up
his a5-pawn as well to transpose into an endgame where his strong bishops and central pawns will guarantee him
strong compensation at the very minimum.) 32...Qe2 33.Qe1 (33.Rd1? Nc4µ) 33...Qxe1+ 34.Bxe1 f6! 35.Nf3 Nxf3+
36.gxf3 Kb6∞
c) 29.Re1!?

Anot her impor tant candidate move to consider, as White brings his last reserve into the battle. The former b1-rook
deprives Black’s queen of the e2-square and is now ready to support central expansion with e4–e5. 29...Kc7 Now
White has a choice of moves of about equal strength, however, in almost every case, Black gets sufficient
counterplay as all his pieces are well placed. 30.b3!? To take the c4-square under control, which is key in a lot of
lines. (30.Qe7+ Kb6 31.Nf6 Bxf6 32.Qxf6 White has achieved one of his goals by eliminating Black’s dark-squared
bishop, however Black’s king has found safety on the queenside and his other pieces and pawns are really well
placed, giving him good counterplay. 32...Qc4 33.Qxe5 Qxc2 34.Bg5 Qxb2 35.Rc1 Bxa4∞; 30.f4 Nd3 31.Bxd3
Qxd3 32.Bxa5+ Kb8 33.Ng5 Bf6 with good counterplay for Black.) 30...Kb7

707
31.Bb1 (31.Qe7+ Ka8 32.Bf4 d3 33.Bd1 c4 34.Bxe5 Re8 35.Qxf7 Bxe5 36.bxc4 Qb7!

Despite being two pawns down, it is Black who now hopes to play for an advantage. 31.Ng5!? With the idea of
hitting Black’s weak spot, the f7-pawn. 31...Qb6 With a complex game. 32.f4 d3 33.Bd1

708
33...Nc4!!µ The only move, and the one that turns the tables from worse to better for Black! 34.bxc4 Bd4+
(…...Rh8) 35.Be3 d2 36.Re2 Qb1µ) 31...c4!

Black must act quickly before White has time to consolidate and push his central pawns. (31...Ka8 32.Qg3! Rb8
33.f4 Nd7 34.e5±) 32.bxc4 Qxc4∞ (32...Nxc4? 33.Bd3+–) 29...Qe2 (29...Kc7 30.Nxf7 Nxf7 31.Qe7+ Kb8 32.Qxf7
Bb7 33.Qf4+! An important check with the double function of taking away the e5-square from Black’s h8-bishop
and protecting White’s d2-bishop.

709
33...Ka7 34.Re1 d3 35.Bd1! White’s safer king is the number one factor in this difficult-to-evaluate position.
Black’s c5-pawn guards the black king, but is needed to support Black’s attempt to push his central d3-pawn, which
means that at some point Black will have to risk opening up the lines and diagonals leading to his king in order to
have a chance of counterplay. White is not hindered by similar considerations, as his e-pawn is rolling with the
support of the rook, queen and coming Bd1–g4. The advantage is with White, although the game remains extremely
complicated with chances for both sides. 35...Bd4 36.e5±) 30.Qf4! f6 31.Nf3

31...Ke7!! Suddenly Black’s king changes course to stay in the center and the game becomes really hard to evaluate.

29.f4±

710
Simple and best. White’s attack is rolling by itself.

29...Nd7 30.Re1!

The last preparatory move before the e4–e5 break. White’s attack is decisive!
30.e5? immediately wouldn’t work: 30...Qe2 31.exf6+ Kd6! and Black starts to play for a win.

30...Qc4 31.Bb1?!

31.Bd1! would have been more accurate: 31...Kd6 (31...Kd8 32.Bxa5++–) 32.Nxf6! Kc7 33.e5 Nxf6 34.exf6 Kb8
35.f5 gxf5 36.Re7+–.

31...Kd6

32.e5+??
711
Absolutely terrible. I was trying to play fast to use my opponent’s time trouble to my advantage, however, in doing
so, I almost spoiled the hard work before this moment.
32.Nxf6!+– wins an important pawn and with it the game. 32...Bxf6 33.e5+ Kc7 34.exf6 Rf8

I was not sure about this position, since it looks like Black is getting a certain degree of play thanks to my open king,
however it’s an illusion. White’s strong pair of bishops has everything under control and the black king is more
vulnerable than mine. 35.f5! gxf5 36.Bxf5 Kb6 37.Bxd7 Bxd7 38.Re7 Qd5 39.Qf4 Rg8 40.Rg7!+–.

32...fxe5

33.fxe5+

Playing for tricks.

The alternative, 33.Qg3!, has more potential, but Black can neutralize it with precise play. 33...d3!

712
[33...Kc7 34.b3! Qd5 35.Be4 Qd6

36.Ng5!? (36.Bxc6 Kxc6 37.Qd3! Kb7 38.Ng5 Ka7 39.Qb5 Qc6!=) 36...Bxe4 37.Nxe4 Qc6 38.Ng5! Re8 39.Bxa5+
Kc8 40.Qd3‚]
34.Qxg6+ (34.Bxd3 Qd4+ 35.Kh1 exf4 36.Bxf4+ Be5!„) 34...Kc7 35.Qxd3 Qxa4!∞ ( 35...Qxd3 36.Bxd3 exf4
37.Bxf4+ Kb6 38.b3 Bd5 39.Bc4²)

33...Nxe5??

My opponent’s final error.

33...Bxe5! and suddenly Black defends everything while holding on to an extra pawn. 34.Bf4 (34.Rxe5 Nxe5 35.Bf4
Qe2µ) 34...Re8 35.Nf6 Looks decisive, but isn’t. (35.Bxg6 Qd5 36.Qg5 Bxf4 37.Qxf4+ Re5=) 35...Re6

White has no advantage any more.

36.Nxd7 (36.Qg5? Bxf4 37.Qxf4+ Ke7µ) 36...Bxf4 37.Qxf4+ Kxd7 38.Qf7+ Kd6 39.Qf4+ Ke7=.

713
34.Bf4

Finally I could relax as this bishop creates a deadly pin from which Black has no escape. White wins material.

34...Qd5

34...Re8 35.Nf6+–

35.Be4 Qe6 36.Ng5

The knight joins the game with a decisive tempo.

36...Qe7 37.Bxc6 Kxc6 38.Bxe5 Bxe5 39.Qe4+ Kd6 40.Qxg6+

Leading to the win of the piece.

40...Kc7 41.Ne6+

714
Following time trouble, Black keeps making moves on autopilot, but the result of the game is clear.

41...Kd6 42.Nf4+ Qf6 43.Qxf6+ Bxf6 44.Re6+ Kd7 45.Rxf6 Rb8 46.Nd3 c4 47.Ne5+ Ke7 48.Rf4 Rxb2 49.Nxc4
Rb4 50.Rxd4 Rxa4 51.Kf2 Ra2+ 52.Kf3 a4 53.Ne3 Ke6 54.Ke4 Ra1 55.g4 Ke7 56.Rd5 Kf6 57.Ra5 Kg6 58.Kf4
Kf7 59.Ra6

Black resigned. 1–0

I cannot say that this game gave me renewed hope, but at least it gave me the chance to lose the match with an
honorable result. I was intent on making life very difficult for my opponent in the following games.

715