Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.2 9/18/84; site rti-sel.UUCP Path: utzoo!linus!philabs!cmcl2!seismo!harvard!talcott!panda!genrad!decvax!mcnc!rti-sel!trt From: trt@rti-sel.UUCP (Tom Truscott) Newsgroups: net.chess Subject: Re: Why can't a machine be World's Checkers Champ? Message-ID: <338@rti-sel.UUCP> Date: Fri, 9-Aug-85 12:05:34 EDT Article-I.D.: rti-sel.338 Posted: Fri Aug 9 12:05:34 1985 Date-Received: Mon, 12-Aug-85 08:20:23 EDT References: <474@oakhill.UUCP> <10913@rochester.UUCP> Organization: Research Triangle Institute, NC Lines: 31 > I'm almost sure that a few years ago I read that man could no longer beat > a checker program. And the only way even the world champ could win a game > was if the computer always moved second. ... I too have read such reports. They are sincere but wrong. In particular Marion Tinsley (math prof., world checkers champion) has expressed non-trivial contempt for computer checkers programs. Here is a snip from a paper I wrote on computer checkers: ``I should say that at present, there are several thousand just average Class B players who could beat either [the Duke or Stanford] computer without difficulty.'' -- W.B. Grandjean, \fIACF bulletin\fP, August 1977 .QP ``Although computers had long since been unbeatable at such basic games as checkers, ...'' -- Clark Whelton, \fIHorizon\fP, February 1978 .PP In computer checkers, as in many areas of artificial intelligence, misconceptions abound as to the present capabilities of machines. ... > ... I'm talking main frame well done checker > programs here, not a radio shack toy. The Duke checkers program was written entirely in S/360 assembly language and run on an IBM 370/165, so I guess that qualifies. But I am sure someone could write a TRS-80 program that could beat it. Look at Kathe and Dan Spracklin's computer chess successes on a micro. Tom Truscott