Author: Robert Hyatt
Date: 12:28:59 08/16/99
Go up one level in this thread
On August 13, 1999 at 23:47:35, William Bryant wrote: >On August 13, 1999 at 03:10:30, James Robertson wrote: > >>Today my program encountered JPCrafty in a computer tournament at FICS. In this >>position: >> >>r1b2rk1/4b2p/p1p1p2Q/2q1ppN1/8/8/4B1PP/3R1RK1 w - - id JPCrafty - JRCP, 1999; >> > >My program likes Rf2 but doesn't see a draw > > 3 1031 0:00 -5.25 Rf2 Bxg5 Qxg5+ Kh8 > 4 2882 0:00 -5.20 Rf2 Bxg5 Qxg5+ Kh8 Bd3 > 5 14972 0:00 -5.30 Rf2 Bxg5 Qxg5+ Kh8 Rc1 Qd6 > 6 35002 0:00 -5.25 Rf2 Bxg5 Qxg5+ Kh8 Rc1 Qd6 Bc4 > 7 147262 0:02 -5.05 Rf2 Bxg5 Qxg5+ Kh8 Rd8 e4 Qf6+ Kg8 Rxc8 Rfxc8 >Qxe6+ Kh8 Bxa6 > 8 556161 0:10 -5.05 Rf2 Bxg5 Qxg5+ Kh8 Rd8 e4 Qf6+ Kg8 Rxc8 Rfxc8 >Qxe6+ Kh8 Bxa6 > 9 2971241 0:51 -4.34 Rf2 Bxg5 Qxg5+ Kh8 Rd8 Rxd8 Qxd8+ Kg7 Bxa6 Qc1+ >Bf1 e4 > 10 6658309 1:56 -4.16 Rf2 Bxg5 Qxg5+ Kh8 Bf3 Rb8 Rc1 Qd4 Bxc6 e4 Qh6 > 11 14467904 4:12 -4.34 Rf2 Bxg5 Qxg5+ Kh8 Rd8 Rxd8 Qxd8+ Kg7 Qg5+ Kh8 >Qd8+ > >How are you detecting a draw by perpetual check? > >Are you looking for a three fold repetition created by the checks or by another >method. > >This is a weakness of my program I would like to fix. > >William >wbryant@ix.netcom.com It takes crafty about 45 seconds to find Rf2 with a score of 9.00 on my PII/300 notebook. I do repetition by looking for repeated positions based on the hash signature... however, I count 2 repetitions as a draw, which works fine in a game, but might not work as well if crafty has to 'take over' a human game. IE if you can force it to repeat the position once, it assumes you can force it to repeat it a second time and draw... In the case of JPCrafty, it is a very slow machine so I doubt it could find this in reasonable time...
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.