Author: martin fierz
Date: 06:07:39 08/12/03
Go up one level in this thread
On August 12, 2003 at 04:58:45, Uri Blass wrote: >On August 12, 2003 at 04:39:36, martin fierz wrote: > >>On August 11, 2003 at 20:51:54, ludicrous wrote: >> >>>A little bit off topic, but I'd like to know if there is a SSDF-like list for >>>Checkers Programs? I've heard of KINGSROW and it seems to be the Ruffian >>>equivalent of Checkers. >>> >>>Which is the strongest? >>> >>>What is the rating systems used in Checkers? >> >>there is no SSDF-like list for checkers programs. there are a couple of strong >>checkers programs out there, all are by far good enough to beat any human. there >>are two strong free programs, you should check out these first; kingsrow and >>cake "sans souci" for the checkerboard interface. >> >>the strongest commercial engine is nemesis, which won the computer world >>championship in front of kingsrow and cake in august 2002, but with a very >>narrow margin. i don't think it is any better than the improved 8pc versions of >>kingsrow and cake which have been published after the championship - nemesis was >>never updated since then. >> >>the chinook team has completed the 10pc endgame database early this year, and >>with that, chinook will be the strongest program there is. > >How do you know it? > >the fact that a program has more tablebases does not mean automatically that it >is stronger. > >Without testing it is impossible to know. > >Uri in principle you are right. but there is good evidence to go on: if i use my program with either the 6-pc or with the 8pc endgame database, the difference is HUGE. you can let them play against each other, and the 8-pc program wins handily. of course, the assumption here is that you can probe the database efficiently, which is not quite easy with the 10pc db: the 6pc db is 40MB, the 8pc db is 4GB and the 10pc db will probably be a few 100 GB.... but if you have the kind of computer you need to run this efficiently (e.g. 64GB ram or so), then it will be much more powerful than any 8pc program. the only weird point is that even though it's much more powerful, it will hardly ever beat the 8pc program, since that makes very few errors already. the big difference is that the 10pc program can announce certain wins/losses/draws much earlier (based upon what i see between 6 and 8pc programs). in checkers, captures are forced, so very many sequences lead to fewer pieces. already now, the 8pc db is hit very often right out of the opening. cheers martin
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.