Counting backward during chess move choice

Level of playing skill was correlated with move quality and evaluation accuracy, and interacted with both counting and moving such that stronger players were more, rather than less, affected by counting interference.

Abstract

Backward counting was used to interfere with working memory during tree search from chess game positions. Skilled players at different levels of expertise analyzed games while counting or not, and also while moving the pieces or not. Counting had the effect of reducing the maximum depth of search, the number of branches searched, the total number of moves considered, and the quality of the moves. Moving the pieces increased the depth of search at the expense of alternative branches slightly, but with no reliable effect on move quality. Level of playing skill was correlated with move quality and evaluation accuracy, and interacted with both counting and moving such that stronger players were more, rather than less, affected by counting interference. Such interference seems more likely to affect the central executive than the verbal or visuospatial components of working memory.