eric@aplvax.UUCP (08/15/83)
In an episode of Startrek, Spock proves that someone has been tampering with the computer by arguing that it had started to lose to him in 3D chess. Since he had programmed it, he argued that it should have been impossible for him to achieve anything better than a draw against it. What does this have to do with net.chess? Simply that this got me curious about something. We all know that computers can not yet beat grand masters, but what about their implementors? Can anyone tell me who wins most the matches between the best of the chess programs and the person who wrote the program? Does Ken Thompson regularly beat Belle? What about the others? If you have ever programmed a strategy game, such as chess, you know where where the weak spots in the algorithms are for that program, and this "inside" knowledge is often sufficient to beat the program. Is this still the case with chess? eric ...!seismo!umcp-cs!aplvax!eric
eric@aplvax.UUCP (08/16/83)
It sounds like Belle does consistently beat Ken Thompson, thanks to all of you who sent me mail about this. What prompted this continuation was one letter I got. Ken Thompson can't beat Belle. So what? He asks. How fast can Seymour Cray add? This is a lousy argument. Is the author (name withheld to protect the innocent) saying that chess program algorithms are as well developed as addition algorithms? What I was trying to get across in the original posting is that chess algorithms are far from perfect, they miss valid strategies, and they can sometimes fall into stupid traps. Now, with the annual computer chess tournament coming up, I can understand authors not wanting to give away the chinks in their programs, but can they use them to consistently beat them. We have heard about Thompson and Belle, does anyone know about NUCHESS, CRAY BLITZ, CHAOS and the others? eric ...!seismo!umcp-cs!aplvax!eric