Author: Ryan B.
Date: 21:44:19 12/08/05
Go up one level in this thread
On December 08, 2005 at 23:55:32, John Rose wrote: >50 game match @ 3min + 0 - Rybka x64 vs Gambit Fruit 1.0 Beta 4bx > >Results - Rykba x64 +25 =19 -6 69% > Gambit Fruit 1.0 Beta 4bx +6 =19 -25 31% > >Previous match results - > > Rykba x64 +33 =9 -8 75% > Deep Junior9 (2 threads) +8 =9 -33 25% > > Rykba x64 +20 =14 -9 68% > Deep Shredder9.12 x64 (2 threads) +9 =14 -20 32% > > Rykba x64 +24 =18 -8 66% > Fritz 9 +8 =18 -24 34% > >Time control - 3min + 0sec >System - Windows x64,4gb RAM,2 dual core AMD Opteron 275 (2.2ghz)=4 total cores >Interface - Chessbase GUI Fritz 9 Oct 17 >Book - Perfect 9.0, option=normal >Hash - 256mb each engine >Ponder - On >Threads - Rybka=1, Gambit Fruit=1 >EGTB- all common 3-4-5, some 6, total about 90gb >Engine Options - default > >Cross Table > >Rybka ½½01½1½01½1½11011½11111½1111½1001½1110½1½½½1½½½1½½ 34.5/50 >G-Fruit ½½10½0½10½0½00100½00000½0000½0110½0001½0½½½0½½½0½½ 15.5/50 > > >Comments - Time control 3min +0 is the most popular setup in >playchess.com's engine room and that is my motivation for running >4 engine vs engine matches at that time control. It's clear that >Rybka dominates this kind of short blitz match. I observed many games >where Rybka made moves very quickly and was able to build up a >significant time advantage. Here is one game where Rybka appeared >lost but due in part to time advantage was able to get a draw. > > [D] 2r1q1k1/3n1p2/r1bB1np1/6Np/p1RPP2P/5PQB/6P1/1R4K1 b - - 0 42 > >Gambit Fruit (White) has just played 42. Rxc4 with an eval of +3.16. >White has 35 seconds left on the clock. Rybka (Black) now plays >42 ...a3 eval +1.66 and 1:39 left on the clock > > [D] 8/8/4Pnk1/3P4/1B2P1p1/8/p1r3P1/R5K1 w - - 0 59 > >With 25 seconds left Gambit Fruit (White) plays 59. e7 eval +2.33 >Perhaps instead 59. Ba3 would have won here ? Later analysis shows >that Gambit Fruit will find 59. Ba3 at depth 13 which requires just >a few seconds, so time pressure was a factor ! Rybka (Black) responded >with 59...Kf7 eval +0.65 with 1:05 clock time > > [D] 4kB2/4P3/8/8/8/6P1/4r3/6K1 w - - 0 80 > >Final position. Game drawn. >Gambit Fruit (White) had 18 seconds left and Rybka (Black) had 49. > >My next few matches will be much longer time controls. Stay tuned ! :-) Rybka is clearly the stronger chess program overall at this time but I think Gambit Fruit stands a better chance at longer time controls. I hope to see more tests at various time controls. Thanks, Ryan
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.