Effects of Chess Software on Competitive Chess: Accuracy, Preparation, and Decision-Making

Abstract

This paper provides a comprehensive empirical and analytical study of the impact of chess software on elite competitive chess from the 1970s to 2020. Using Average Centipawn Loss (ACPL) as the primary accuracy metric, the study integrates statistical analysis with detailed move-by-move annotated case studies of World Championship games across five decades. The findings demonstrate a clear decline in error magnitude, substantial deepening of opening preparation, and a structural shift in decision-making under the influence of chess engines and databases. The research contributes to literature on decision sciences, human–machine interaction, and performance optimization in complex strategic environments.