Jeffrey Tarrant is a Wall Street guy. He spent the last thirty years investing in new hedge funds. As the founder and CEO of a firm called Protege Partners, he compares himself to Sam Altman, the president of Y Combinator. What Altman does for Silicon Valley tech startups, Tarrant does for hedge funds. "I help seed them and incubate them," he says.
The analogy has never been more true than right now. Tarrant is beginning to explore hedge funds built on ideas that have sparked so many companies and created so much wealth in Silicon Valley---ideas that span artificial intelligence, crowdsourcing, and digital currencies. He believes the hedge fund world is on the verge of a new revolution, a revolution he calls the Third Wave.
The 1970s saw the rise of discretionary funds, where iconic investors like George Soros used their very human judgments to find new opportunities in the market. Then came the "quants" at funds like Renaissance Technologies, who found even greater opportunities through statistics and computer algorithms. Now, Tarrant says hedge funds are moving beyond the quants.
As a prime example, he cites Numerai, a San Francisco hedge fund that makes trades using machine learning models built by thousands of anonymous data scientists paid in bitcoin. Funds such as Quantopian and Quantiacs are tapping the wisdom of the masses in other ways. And then there's Polychain, a fund that invests exclusively in bitcoin and other digital "tokens" housed on a blockchain, the distributed online ledger that makes cryptocurrencies possible. As its name suggests, Polychain isn't just investing in digital coins—it's investing in a radically new breed of businesses owned, funded, and operated entirely by decentralized networks of anonymous online investors.
Such funds aren't always easy to wrap your head around. But as Wall Street tries to make sense of them, these new tech-driven approaches to investing are proliferating. In the late `90s, Tarrant helped build an online directory of hedge funds called AltVest. Now, he's building a directory for this new wave of funds. It includes roughly fifty players, many of whom have yet to publicly announce themselves---though Tarrant admits that only about half have demonstrated real promise so far.
Not surprisingly, some financial vets question how effective Tarrant's new wave will be. In a recent Bloomberg story, several fund managers said that recent enthusiasm for machine learning is overblown. In some cases, even the founders of these Third Wave funds urge caution.
"Regardless of what method you use in quantitative finance---be it machine learning or traditional quant methods---there are an infinite number of ways to fail," says Martin Froehler, a former quant with Superfund Asset Management GmbH in Switzerland who went on to found Quantiacs. Machine learning models are no "superweapon," he says. In his experience, ninety percent of live machine learning tests fail.
But Froehler's fund benefits from machine learning, too. Based in Silicon Valley, Quantiacs attempts to crowdsource the quant model, and many of the quants feeding the fund are using machine learning technologies. Among other things, they're making use of deep neural networks, complex mathematical systems for recognizing patterns in vast amounts of data. In other words, the Third Wave is not just about using one new technique. It's about combining techniques, from machine learning to crowdsourcing to the blockchain.
Nor is this just a battle of the old guard and the new. The founder of Renaissance has invested in Numerai, and Point72 Asset Management, the fund founded by billionaire Stephen Cohen, has put money into Quantopian. "These people who I considered old school really understood what I was getting at," says Numerai founder Richard Craib. "And I thought I was going to be ahead of my time."
Even the apparent skeptics are embracing the trend. "I'm concerned that people may have unrealistic expectations of what is possible with the current state of the art," David Siegel, co-founder of storied quant fund Two Sigma Investments, said last fall. But more recently, his fund ran an online contest through Silicon Valley data scientist marketplace Kaggle, offering a $100,000 prize for the best machine learning model. One company director indicated the contest was more of a recruitment tool than a full embrace of crowdsourcing or machine learning. But whatever the intention, it was yet another example of Silicon Valley and Wall Street drawing closer than ever before.