Toward a Model of Human Game Playing

This paper presents an overview of a game playing model which is based on human perceptual and problem solving abilities. The representation of games, acquisition of rules, learning of strategies, and selection of moves are outlined. Details of move selection, includin scanning of the board, use of familiar patterns to suggest move candidates, evaluation of moves, and lookahead are described. Finally, there is a discussion of the means by which the model learns to play a better game.