Optimal Robot Play in Certain Chess Endgame Situations

Abstract

Robots are an aspiring factor in science, economy and society. We show that the adoption of human strategies and behavior is not always appropriate. Using the example of chess playing robots we demonstrate that the assumption of each move taking the same amount of time is invalid for robots at blitz chess. We present a different approach. Instead of optimizing the remaining number of moves in endgame play the actual time required for a move is considered and used for optimization. It is shown that this brings a significant change in gameplay.