RankCut - A domain independent forward pruning method for games

Abstract

Forward pruning, also known as selective search, is now employed in many strong game-playing programs. In this paper, we introduce RankCut - a domain independent forward pruning technique which makes use of move ordering, and prunes once no better move is likely to be available. Since game-playing programs already perform move ordering to improve the performance of αβ search, this information is available at no extra cost. As RankCut uses additional information untapped by current forward pruning techniques, RankCut is a complementary forward pruning method that can be used with existing methods, and is able to achieve improvements even when conventional pruning techniques are simultaneously employed. We implemented RankCut in a modern open-source chess program, CRAFTY. RankCut reduces the game-tree size by approximately 10%-40% for search depths 8-12 while retaining tactical reliability, when implemented alongside CRAFTY'S existing forward pruning techniques. Copyright © 2006, American Association for Artificial Intelligence (www.aaai.org). All rights reserved.