Author: Vincent Diepeveen
Date: 12:53:49 01/01/05
Go up one level in this thread
On January 01, 2005 at 12:21:50, Peter Berger wrote: You are correct in one aspect, because of booklearning and the huge number of games and other probably human errors, ratings are closer than in reality they are. I personally feel 7.04 is better than shredder8; however because it's newer s8 is better in killing engines it had already trained against at home so should on paper get a higher rating than 7.04. Note that you can get major inaccuracies in testing already when in between matches the learning doesn't get cleaned of books. I'm sure no one is doing that, as fritz simply doesn't have that functionality inside its GUI. So if engine A with its book first plays S704 and then plays against a fresh installed S8, that's a major advantage for engine A. Additional Ed Schroeder has shown a 5% difference in score when an engine doesn't get loaded with the interface. Trivial all ratings here are within 5% score difference from each other. Yet in sports it doesn't matter whether you are 1000 points better or 0.00001 point better. What matters is who is at #1 spot. That happens to be S8 simply, so the rest of the discussion here is complete theoretic. >On December 31, 2004 at 07:51:37, Thoralf Karlsson wrote: > >> THE SSDF RATING LIST 2004-12-31 100049 games played by 267 computers >> Rating + - Games Won Oppo >> ------ --- --- ----- --- ---- >> 1 Shredder 8.0 CB 256MB Athlon 1200 MHz 2805 23 -22 1075 71% 2645 >> 2 Shredder 7.04 UCI 256MB Athlon 1200 MHz 2804 23 -22 1041 70% 2653 >> 3 Deep Fritz 8.0 256MB Athlon 1200 MHz 2791 25 -24 896 72% 2628 > >Two thoughts: > >a.) I was surprised how close Shredder and Fritz actually are - in fact too >close to call. > >THE SSDF RATING LIST 2004-04-22 97872 games played by 264 computers > Rating + - Games Won Oppo > ------ --- --- ----- --- ---- > 1 Shredder 8.0 CB 256MB Athlon 1200 MHz 2818 34 -32 481 70% 2673 > 2 Shredder 7.04 UCI 256MB Athlon 1200 MHz 2809 24 -23 967 71% 2648 > 3 Deep Fritz 8.0 256MB Athlon 1200 MHz 2790 26 -25 855 72% 2625 > > >b.) The major change to the previous list is that Shredder 8 had to play some >lower-rated opponents which hurt its performance. It's interesting to compair >the average rating of the opponents anyway. There is a slight tendency for a >higher average rating of the opponents to result in a higher final rating for an >entry. Maybe there is some rating incest going one here? Maybe choice of >opponents has a too strong influence. > > >Just some thoughts - no criticism intended. > >Peter
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.