Author: Dana Turnmire
Date: 21:56:26 04/07/01
If anyone is interested here are a few of the positions where CM5000 did better than CM8000 on the CCR One Hour test by Larry Kaufman. To run the test, set the program on infinite with no book. When you come to the test move simply note the best move after 15" 30" 1min and 2 min. After each time period if the correct move is being evaluated give it one point. If it has the correct move after 15 seconds but the wrong move at the one minute mark it still gets one point for the 15 second period. The most points for any one test is 4 and of course 0 points is also possible. Here are all the positions where CM5000 did better than CM8000 on an AMD k6-2/500/64RAM. Set 4 1.d4 d5 2.c4 e6 3.Nc3 Nf6 4.cxd5 exd5 5.Bg5 c6 6.Qc2 Be7 7.e3 Nbd7 8.Bd3 0-0 9.Nf3 Re8 10.0-0 Nf8 11.Rab1 g6. White to move. Solution 12.b4!, the minority attack and the point of the previous move. CM5000 got 4 points, CM8000 got 2 points. Set 7 1.d4 Nf6 2.c4 c5 3.d5 b5 4.cxb5 a6 5.bxa6 Bxa6 6.Nc3 d6 7.Nf3 g6 8.g3 Bg7 9.Bg2 0-0 10.0-0 Nbd7 11.Re1 Qb6 12h3. Black to move. Solution 12...Rfb8!, so that one rook will be posted on each semi-open file. CM5000 got 4 points, CM8000 got 0 points! Set 16 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.0-0 Nxe4 6.d4 b5 7.Bb3. Black to move. Solution 7...d5!, returning the pawn in the best way, since 7...exd4 will only keep black a pawn up for a few moves. CM5000 got 1 point, CM8000 got 0 points. I thought maybe the test was biased towards tactical moves but these test sets look more like positional evaluations to me.
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.