Computer Science Department
School of Computer Science, Carnegie Mellon University


A Game-Playing Program that Learns by Analyzing Examples


Steven Minton

May 1985

This paper describes a game-playing program that learns tactical combinations. The program, after losing a game, examines the opponent's moves in order to identify how the opponent forced the win. By analyzing why this sequence of moves won the game, a generalized description of the winning combination can be produced. The combination can then be used by the program in later games to force a win or to block an opponent's threat. This technique is applicable for a wide class of games including tic-tac-toe, go-moku and chess.

28 pages

Return to: SCS Technical Report Collection
School of Computer Science homepage

This page maintained by