Author: Sune Fischer
Date: 12:50:51 03/20/02
Go up one level in this thread
On March 20, 2002 at 15:40:18, Dann Corbit wrote: >On March 20, 2002 at 15:29:29, Sune Fischer wrote: >[snip] >>Sure, but still finite mind you :) >>I get the feeling that some people do not fully understand the meaning of >>_infinite_ ;) > >True enough. Even a single draw in 1e1000000000 games will drub a perfect score >ELO. The point I was making was that the size of ELO figures is irrelevant. In >fact, it would make a lot of sense to have the average player get an ELO of >zero. > >Unfortunately, this sort of list won't make anyone happy: > > Program Elo + - Games Score Av.Op. Draws > 1 Yace 0.99.50 : 307 31 106 189 86.5 % -15 12.2 % > 2 AnMon-509 : 268 41 102 124 81.9 % 6 15.3 % > 3 Amy-07 : 224 36 109 159 83.0 % -52 8.8 % > 4 ZChess-222 : 216 31 77 229 79.5 % -19 12.2 % > 5 ZChess-120 : 176 35 72 194 75.8 % -22 14.4 % > 6 WildCat-257 : 147 46 83 120 74.6 % -40 17.5 % > 7 FranWB-090 : 142 55 82 96 69.3 % 1 17.7 % > 8 Yace-020 : 133 37 64 192 70.6 % -19 14.1 % > 9 FRANWB-081 : 104 37 60 210 67.6 % -24 11.4 % > 10 Amyan-1423 : 69 28 40 391 64.3 % -33 13.0 % > 11 Fortress-162 : -8 48 40 205 51.5 % -18 22.0 % > 12 Faile-14 : -58 46 45 202 44.8 % -21 15.3 % > 13 Amyan-111 : -60 36 33 352 42.8 % -9 13.9 % > 14 Aldebaran-070 : -190 51 26 358 27.4 % -21 12.8 % > 15 31337-30 : -250 93 32 199 19.3 % -2 7.5 % Hehe, guess it won't, but now that we have programs that can perform at a constant strength, we could "define" crafty version X on Y hardware to e.g. 2500 elo and then scale everything by that. But then ratings aren't very precise to begin with... -S.
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.