Author: Terry McCracken
Date: 09:17:28 05/04/02
Go up one level in this thread
On May 04, 2002 at 01:59:00, Uri Blass wrote: >On May 04, 2002 at 01:33:49, Terry McCracken wrote: > >>On May 04, 2002 at 00:58:59, Russell Reagan wrote: >> >>>I'm curious if anyone knows anything about the computer checkers scene. I know >>>about computer chess, how chess programs work, how good chess programs are >>>compared to the best human chess players, etc. and would like to find out about >>>checkers programs. I know that (for example) go programs aren't anywhere close >>>to master strength. So are checkers programs as good as the best master checkers >>>players? I would guess that checkers would probably use some of the same basic >>>methods as a chess program, such as search and evaluation, but are there any >>>special methods that a checkers program would use that chess programs don't? Or >>>any methods a chess program uses that a checkers program wouldn't use >>>successfully? >>> >>>I've only recently started playing checkers to get a feel for the game, but it >>>seems like there are a great number of positions where zugzwang occurs and so >>>null move probably wouldn't work in checkers, or maybe it would need >>>modifications. I'm mainly interested in checkers programs and that whole >>>"scene", but also if anyone has any info to offer on other game playing programs >>>I'd like to hear that as well. I'm more a fan of game playing programs and AI >>>methods in general than only chess. >>> >>>One last thing I thought about was endgame tablebases. Are these available for >>>checkers or other games? It seems like it might be both faster and smaller to >>>generate the tablebases for checkers, but I'm not totally sure on that. There >>>are fewer pieces on the board and fewer squares of the board that are in use, so >>>it seems like endgame tablebases might play a more significant role if they were >>>created for checkers. >>> >>>Thanks for your info. >>> >>>Russell >> >>I really can't say how checker programmes in general, or even the best >>commercial programmes play, relative to chess programmes. However, I _do_ know >>Chinook is unbeatable, it plays perfect checkers since a bug was removed >>after plaing Tinsley in 1994. And yes Chinook has endgame tablebases:o) > > >How do you know that it plays perfect checkers? > >A player can be unbeatable not because of playing perfect but because of the >facts that the opponents do not know to take advantage of it's mistakes. > >Uri Uri to the best of my knowledge Chinook plays checkers perfectly. Chinook doesn't make mistakes for _any_ player to take advantage of! How do I know this? I know a very good friend of the creator Chinook. He has detailed information from Jonathan Schaeffer whom has been his personal friend since 1971! When Chinook faced Dr. Tinsley in 1994 Chinook hadn't been perfected, since that time it has. Without doubt it plays checkers at least 99.9% of perfection! No one can claim that in chess at present. Then again, chess is far more difficult to solve. So Uri I wouldn't make such a claim without some _inside_ information. Terry P.S. The Online Version of Chinook is a "Neutered Version" of this beast!
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.