Author: Robert Hyatt
Date: 17:03:12 11/08/01
Go up one level in this thread
On November 08, 2001 at 19:21:46, Odd Gunnar Malin wrote: >On November 08, 2001 at 11:58:08, Robert Hyatt wrote: > >>On November 08, 2001 at 02:48:14, Odd Gunnar Malin wrote: >> >>>On November 07, 2001 at 21:53:28, James T. Walker wrote: >>> >>>>Here are 12 games played at 40/40 minutes. Crafty played on an Athlon 1.4G and >>>>CM8K played on an Athlon 1.2G. Final score was 6.5-5.5 for CM8K. After 70 >>>>games at 40/40 or 40/60 it looks like Crafty 18.12(K7) is not as strong as >>>>Crafty 18.10. I wonder if others are having similar results? >>>> >>> >>>Result after 6 rounds are CM8000-Crafty 4-2 >>> >>>I Checked a little on this in SSDF's list. >>>Result after 6 round is in brackets. >>> >>>CM8000-DeepFritz 7.5-32.5 (3-3) >>>CM8000-Junior 5 13.5-26.5 (2.5-3.5) >>>CM8000-Hiarcs7.32 23.5-16.5 (4.5-1.5) >>> >>>Is the SSDF's rating only valid for match play between computers? >>>Or is it a ratinglist on how good the booklearning is? >> >> >>Simple. When a "contest" is held, using a set of known rules and so forth, >>human nature is to take advantage of the contest rules themselves, as much as >>you try to take advantage of normal chess rules. In long matches between two >>programs, book learning is _the_ edge. In tournaments, it has no influence to >>speak of. So SSDF ratings don't say a lot about how the top 10 programs on >>the list would end up in (say) a round-robin tournament. >> > >Then you could ask "What is the purpose of this list?". > >The rules are not clear defined on their page, but they say in the faq: "Our >goal is to always play matches of 20 games between two computers/programs under >test". >But the chairman talk about tournament games in his comment. > >Since the list are used like the FIDE list, maybe it had been better if they >arranged playing too like FIDE. >Instead of playing matches they could perhaps played double round robin or >something like that. Maybe had a 'first saturday' tournament for computers. > >Odd Gunnar > Tournaments would be much better. But then someone comes along and says "XX got stomped in SSDF competition, but it stomped YY when I played then in a 100 game match." :) >>> >>>The exception that confirm the rules is: >>>CM8000-ChessTiger 15.5-24.5 (1-5) >>> >>>Odd Gunnar
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.