Author: Shaun Brewer
Date: 04:09:46 10/09/02
I wanted to remove the effect of opening books and opening book selection (specifically relevant when testing 17.xx versions) from my experiment so I created 36 positions to test. These are hopefully fairly varied. I chose not to use the Nunn positions as there is a risk that if these are used by an author to test program improvements that the engine will become tuned to these positions. PC: Tbird 850 512mb ram approx 128mb only 4 piece TB files present. So each engine pays each position as white 72 games from 36 positions. Game in 1m Individual statistics: (1) Crafty-18.15 : 72 (+ 29,= 26,- 17), 58.3 % (2) wcrafty-17.14 : 72 (+ 17,= 26,- 29), 41.7 % Individual statistics: (1) wcrafty-17.10 : 72 (+ 26,= 23,- 23), 52.1 % (2) Crafty-18.15 : 72 (+ 23,= 23,- 26), 47.9 % Game in 5m Individual statistics: (1) wcrafty-17.14 : 72 (+ 22,= 32,- 18), 52.8 % (2) Crafty-18.15 : 72 (+ 18,= 32,- 22), 47.2 % Individual statistics: (1) wcrafty-17.10 : 72 (+ 27,= 24,- 21), 54.2 % (2) Crafty-18.15 : 72 (+ 21,= 24,- 27), 45.8 % Individual statistics: (1) Crafty-18.15 : 72 (+ 25,= 30,- 17), 55.6 % (2) Crafty-19.0 : 72 (+ 17,= 30,- 25), 44.4 % Game in 15m Individual statistics: (1) Crafty-18.15 : 72 (+ 23,= 27,- 22), 50.7 % (2) wcrafty-17.14 : 72 (+ 22,= 27,- 23), 49.3 % Individual statistics: (1) wcrafty-17.10 : 71* (+ 25,= 26,- 20), 53.5 % (2) Crafty-18.15 : 71* (+ 20,= 26,- 25), 46.5 % * 1 game aborted crafty or winboard problem? Individual statistics: (match in progress) (1) Crafty-18.15 : 40 (+ 14,= 15,- 11), 53.8 % (2) Crafty-19.0 : 40 (+ 11,= 15,- 14), 46.2 % Longer game 40 moves in 30 minutes XP2000+ approximatley 2x speed of my 850. Individual statistics: (match in progress) (1) wcrafty-17.10 : 60 (+ 16,= 30,- 14), 51.7 % (2) Crafty-18.15 : 60 (+ 14,= 30,- 16), 48.3 % Shaun
This page took 1.87 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.