The performance evaluarion tests for environmental research (PETER) program is a series of studies involving extended practice on a variety of task. One of these tasks is a compensatory tracking task which requlres the subject to keep a moving circle centered in a horizontal track by making approptlate left-right manipulations of a control stick ( 1 ) . Eighteen subjeas practiced this task every working day for three consecutive weeks, 10 1-min. trials a day. Approximately 1 yr. later 10 of the same 18 subjects practiced a commercially available video game, Atari's Air Combat Maneuvering. The video task requires the subject to align a black approximately triangular attack jet with a same-sized white target or drone jet so that a fired missile will intercept ( 2 ) . These requirements traditionally define a pursuit tracking task. This task was also practiced 10 trials (games) a day every working day for three consecutive weeks; each game lasted 2 1/3 min. Differential stability refers to that point in practice after which individual learning curves are parallel and at most slowly increasing, except for random error; after this point all intertrial (interday) correlations are the same except for sampling variations ( 2 ) . Such constancy indicates that neither the subjects' strategies nor the nature of what is being measured are changing with practice, that is, the task has stabilized. Analysis is confined to stabilized trials (days) only. For the game the average correlation among stabilized trials was .93. The. average correlation between stabilized trials on the game and compensatory tracking was .78. The average correlation among stabilized trials in compensatory tracking was also .78. Hence, the correlation between game scores and compensatory tracking was as large as the reliability of the latter would allow. The implication is that the video game, a pursuit tracking task, measures the same thing at stabilized levels as compensatory tracking does, only more reliabiy. It appears, therefore, that this particular video game may have a future as a portable, low-cost substitute for traditional computer-driven laboratory tracking tasks. This future, moreover, may include predictive testing and training as well as evaluation of performance.