We compared the traditional intuitive criterion (fixed at 5 dB) with the variable Statpac 2 criterion of the Humphrey Field Analyzer for identifying visual deterioration at individual points within a visual field. The separation of depressed from stable points in 123 follow-up fields of 17 patients, when all points within a field were considered, was reasonably equivalent between the two methods (kappa = 0.55 +/- 0.027 [+/- SE]). Centrally located points within the field demonstrated a stronger agreement (kappa = 0.68 +/- 0.028), because the fixed criterion more often labeled the edge points as having progressed than did the Statpac 2 algorithm. The glaucoma change probability printout is more simply obtained than the manual comparison required for intuitive interpretation with a fixed criterion, and it is potentially more accurate because of its reliance on a database that permits a variable criterion to be applied.