Author: Mike Hood
Date: 15:33:12 04/03/02
Go up one level in this thread
On April 03, 2002 at 08:52:42, Sune Fischer wrote: >On April 03, 2002 at 08:14:02, Daniel Clausen wrote: > >>On April 03, 2002 at 07:07:03, Sune Fischer wrote: >> >>>Perhaps there is a shortcut, maybe we don't need to search the entire tree, all >>>we need is to invert a 64x64 matrix :) >>>I can't get it out of my head that it might be possible to solve the problem >>>mathematicly if we find the right model. >> >>Well, mathematically, chess is solved. The compression algorithm is also very >>efficient - it's the rules of the game of chess. It's just that usually the >>decompression takes quite some time. >> >>:) >> >> >>Sargon > >Ah, but is like saying we understand everything about turbulence just because we >know some stochastics of how the molecules interact :) > >I guess here the positional play is macro canonial behavior and tactics is micro >canonial :) > >Perhaps it is provably unsolvable, a three body problem of game theory! >Even so, pertubation theory should tell us something, LOL. > >-S. Okay, now let me tell you what was behind my original message. I've had 32-piece tablebases on my mind for a few months, even before Chessbase's excellent April 1st posting at http://www.chessbase.de/nachrichten.asp?newsid=714 (unfortunately in German only). A 32-piece tablebase would "solve chess", as it's so quaintly called. But would the existence of a 32-piece tablebase ruin chess or eliminate chess as a sport? I think not. The technical problems of creating a 32-piece tablebase are immense, but it is nowhere near as difficult as the task of a human memorizing this tablebase well enough to use it in his daily play. Just to take one example: the KQB-KRR already exists, but is there one single person who has studied it well enough to play this endgame faultlessly? No. The biggest impact of the existence of such a tablebase would not be on chess, it would be on computer chess. It would eliminate the need for chess algorithms. Instead of there being competing chess engines, there would be only one chess program necessary which simply performs table look-ups. "Mate in n" problems would be solved faster than the positions can be set up, and created almost as fast. Time controls and pondering for computer games would be unnecessary, because all moves would be made immediately. Those are my thoughts. There were relatively few replies posted to my question, but I hope it has made a lot of you think. But to answer my own question: I'm a person who is fascinated by computers, by chess, and by computer chess in particular. I couldn't stand someone destroying my hobby overnight. Normally I'm not a violent person, but in this case I would definitely shoot the programmer.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.