Asri-unix.143 net.chess utzoo!decvax!ucbvax!menlo70!sri-unix!mclure Wed Nov 25 14:57:17 1981 John McCarthy and computer chess As many of you know, the first major computer chess games were played between a Soviet program and a Stanford program supervised by John McCarthy, and written by a student (I forget his name). I had a chance to have an extensive discussion over a Chinese dinner with Professor McCarthy a few days ago about his interests in computer chess. He had a few amusing comments. He became dissatisfied with the current direction when less and less "AI" was put into programs, and instead the programs leaned toward brute force. This is quite true; however it appears that the brute force method may have recently hit a brick wall (according to Thompson) in playing strength. CRAY BLITZ and BELLE are just barely masters, and even that may be questionable when masters begin to study their past games. It seems that the current methods produce extremely good tactical play, with full-width lookahead 7-8 ply in the middle game, rapidly increasing in depth as pieces are exchanged. My own view is that there will be a world computer champion in 25-40 years (Thompson thinks 25-30), which will incorporate this tactical play with an extremely large database of tens/hundreds of thousands of chess patterns, which will be used to trim the search tree and increase the depth enormously. McCarthy offered a $10 bet to the person who came up with the shortest win against the Chess Challenger 7 (a rathermediocre player); one person gave him a 6 move win. ----------------------------------------------------------------- gopher://quux.org/ conversion by John Goerzen of http://communication.ucsd.edu/A-News/ This Usenet Oldnews Archive article may be copied and distributed freely, provided: 1. There is no money collected for the text(s) of the articles. 2. The following notice remains appended to each copy: The Usenet Oldnews Archive: Compilation Copyright (C) 1981, 1996 Bruce Jones, Henry Spencer, David Wiseman.