A New Approach to Draw Detection by Move Repetition in Computer Chess Programming

TL;DR

This paper introduces a novel method for draw detection in computer chess by utilizing move repetition through variant strings generated during search, aiming to improve efficiency especially in endgame positions.

Contribution

The paper presents a new approach based on variant strings for draw detection, contrasting with traditional matrix and bitboard methods, and demonstrates its implementation in the Axon chess program.

Findings

The new method is more efficient in endgame positions.

Axon chess program achieves master-level strength.

The approach outperforms standard techniques in specific scenarios.

Abstract

We will try to tackle both the theoretical and practical aspects of a very important problem in chess programming as stated in the title of this article - the issue of draw detection by move repetition. The standard approach that has so far been employed in most chess programs is based on utilising positional matrices in original and compressed format as well as on the implementation of the so-called bitboard format. The new approach that we will be trying to introduce is based on using variant strings generated by the search algorithm (searcher) during the tree expansion in decision making. We hope to prove that this approach is more efficient than the standard treatment of the issue, especially in positions with few pieces (endgames). To illustrate what we have in mind a machine language routine that implements our theoretical assumptions is attached. The routine is part of the Axon chess program, developed by the authors. Axon, in its current incarnation, plays chess at master strength (ca. 2400-2450 Elo, based on both Axon vs computer programs and Axon vs human masters in over 3000 games altogether).