Checking chess checks with chunks: A model of simple check detection

Abstract

The procedure by which humans identify checks in check positions is not well understood. We report here our experience in modelling this process with CHREST, a general-purpose cognitive model that has previously successfully captured a variety of attention- and perception-related phenomena. We have attempted to reproduce the results of an experiment investigating the ability of humans to determine checks in simple chess positions. We propose a specific model of how humans perform this experiment, and show that, given certain reasonable assumptions, CHREST can follow this model to create a good reproduction of the data.