Page 1 of 1

Ethereal 11.25 64-bit Gauntlet for CCRL 40/40

Posted: Thu Mar 14, 2019 11:17 pm
by Graham Banks

Re: Ethereal 11.25 64-bit Gauntlet for CCRL 40/40

Posted: Thu Mar 14, 2019 11:18 pm
by Graham Banks

Code: Select all

CCRL 40/40 Rating List - Custom engine selection
1000012 games played by 2377 programs, run by 22 testers
Ponder off, General books (up to 12 moves), 3-4-5 piece EGTB
Time control: Equivalent to 40 moves in 40 minutes on Athlon 64 X2 4600+ (2.4 GHz), about 15 minutes on a modern Intel CPU.
Computed on March 10, 2019 with Bayeselo based on 1'000'012 games
Tested by CCRL team, 2005-2019, http://computerchess.org.uk/ccrl/4040/

Rank               Engine                Elo   +    -   Score  AvOp  Games
1 Ethereal 11.25 64-bit              3242  +18  -18  47.4%  +15.9   898
  Ethereal 11.00 64-bit              3225  +14  -14  47.0%  +18.6  1552
  Ethereal 10.55 64-bit              3191  +18  -18  49.2%   +4.9   951
  Ethereal 10.00 64-bit              3092  +15  -15  50.7%   -1.2  1286
  Ethereal 9.65 64-bit               3042  +18  -18  55.0%  -32.2   959
  Ethereal 9.30 64-bit               3004  +26  -26  50.5%   -3.3   460
  Ethereal 9.00 64-bit               2968  +18  -18  48.6%   +8.9  1006
  Ethereal 8.62 64-bit               2884  +24  -24  47.5%  +15.8   563
  Ethereal 8.37 64-bit               2814  +27  -27  52.1%  -14.7   438
  Ethereal 8.28 64-bit               2745  +29  -29  48.9%   +4.9   370
  Ethereal 8.19 64-bit               2685  +24  -23  52.8%  -18.7   582
  Ethereal 8.02 64-bit               2620  +31  -31  52.1%  -15.0   359
  Ethereal 8.12 64-bit               2611  +26  -26  51.9%  -11.3   509
  Ethereal 8.05 64-bit               2599  +28  -28  48.1%  +12.4   430
  Ethereal 7.78 64-bit               2586  +25  -25  52.5%  -16.3   549
  Ethereal 7.70 64-bit               2512  +31  -31  47.1%  +15.2   365

Re: Ethereal 11.25 64-bit Gauntlet for CCRL 40/40

Posted: Fri Mar 15, 2019 7:48 am
by AndrewGrant
Thanks, Graham. I'm very happy to that the gains are seen in this test. I was worried after seeing minimal progress in on 40/4 4-CPU list. I suspect a small or unlucky sample size is to blame, and this data for 40/40 1-CPU makes me feel a bit better :)