Perception of the statistical structure of a random series of binary symbols.

Abstract

The apparent ability of human observers to perceive and reproduce sequential statistical structuring in a random series of binary events was explained in terms of specific responses which the observers make to specific sequences of events occurring prior to each trial. These prior events included: the events which had been occurring in the symbol series, the predictions which the observers had already made, and the correctness of the prior predictions of the observers. The relevance of these results to the formulation of information transfer was discussed.