Control strategies for two-player games

Abstract

Computer games have been around for almost as long as computers. Most of these games, however, have been designed in a rather ad hoc manner because many of their basic components have never been adequately defined. In this paper some deficiencies in the standard model of computer games, the minimax model, are pointed out and the issues that a general theory must address are outlined. Most of the discussion is done in the context of control strategies, or sets of criteria for move selection. A survey of control strategies brings together results from two fields: implementations of real games and theoretical predictions derived on simplified game-trees. The interplay between these results suggests a series of open problems that have arisen during the course of both analytic experimentation and practical experience as the basis for a formal theory.