Author: ERIQ
Date: 13:25:34 10/05/03
Go up one level in this thread
On October 05, 2003 at 14:40:56, Robert Hyatt wrote: > > >Even after your corrected results, the programs are reasonably close. From >observation, I think they are close enough that it takes a _really_ long >match (>100 games at least) to draw any conclusions at all... > >Just a personal opinion, of course. yes 10000 games *might* prove something but who has the time. I think I can make a good guess at this situation. I think based all of my matches for these two engines It might be fair to say that crafty is about 100 Pt's. higher now. I don't just start the match and then go to sleep I watch a good many of the games and I have noticed that in my first match crafty18.x it would lose badly in all but the opening at fast Tc using gcc2.95 to compile it. At 19.1 I noticed it holding it's own in the middle game and then losing endgame's using same compiler. I then moved to gcc33 and wow crafty ups the heat and is now looking alittle better in the middle game which kept it out of trouble in the endgame even at fast time controls. I then got crafty 19.4 compiled after about pulling out all of my hair,(BTW please fix the damm makefile) now I have no hair left :) so ok now crafty is creating winning positions in middle game and does not have to worry at all in endgame. To me just watching this I feel like crafty went up about 50Pt's. in the middle. But at slow Tc crafty has been the best since 19.1 and I don't even have tablebases yet!. I could play a slower match but I think the beatings would only get worse. It would also seem to me that crafty likes more speed so when I go dual I will have no equal. I really hate to say this, as yace has been my favorite free engine for awhile and still is, but it is nolonger the king on my box :(
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.