Author: Uri Blass
Date: 08:49:34 06/21/04
Go up one level in this thread
On June 21, 2004 at 10:57:21, Jaime Benito de Valle Ruiz wrote: >In this position, engines such as Crafty or Fruit need to reach depths of about >13 (15-30 secs on an Athlon 1800+) before they find the winning, while many >other (even amateur) engines show very high scores from depth 1: Crafty shows high scores from depth 1 New game, 7k/6p1/1p6/p2Q4/2P1qPKP/8/8/8 b - - 0 1 Analysis by Crafty 19.13: 1...Qxd5 2.cxd5 -+ (-6.52) Depth: 1/5 00:00:00 1...Qxd5 2.cxd5 Kg8 -+ (-6.62) Depth: 2/5 00:00:00 1...Qxd5 2.cxd5 Kg8 3.d6 -+ (-5.89) Depth: 3/7 00:00:00 1...Qxd5 2.cxd5 Kg8 3.d6 Kf7 -+ (-6.19) Depth: 4/10 00:00:00 1...Qxd5 2.cxd5 Kg8 3.Kf5 a4 4.d6 -+ (-5.74) Depth: 5/13 00:00:00 1...Qxd5 2.cxd5 Kg8 3.d6 Kf7 4.d7 Ke7 5.Kf5 -+ (-5.46) Depth: 6/15 00:00:00 4kN 1...Qxd5 2.cxd5 Kg8 3.d6 Kf7 4.d7 Ke7 5.Kf5 a4 -+ (-5.71) Depth: 7/17 00:00:00 10kN 1...Qxd5 2.cxd5 Kg8 3.d6 Kf7 4.d7 Ke7 5.Kf5 a4 6.Ke5 -+ (-5.61) Depth: 8/19 00:00:00 20kN 1...Qxd5 2.cxd5 Kg8 3.Kf5 a4 4.Ke6 Kf8 5.d6 Ke8 6.h5 -+ (-5.21) Depth: 9/21 00:00:00 44kN 1...Qxd5 2.cxd5 Kg8 3.Kf5 a4 4.Ke6 Kf8 5.d6 Ke8 6.h5 -+ (-5.61) Depth: 10/23 00:00:00 83kN 1...Qxd5 2.cxd5 Kg8 3.Kf5 a4 4.Ke6 Kf8 5.d6 Ke8 6.d7+ Kd8 7.h5 -+ (-5.98) Depth: 10/23 00:00:00 88kN 1...Qxd5 2.cxd5 Kg8 3.d6 Kf7 4.f5 b5 5.d7 Ke7 6.Kf4 a4 7.f6+ gxf6 8.h5 -+ (-5.58) Depth: 11/25 00:00:00 160kN 1...Qxd5 2.cxd5 Kg8 3.d6 Kf7 4.f5 Ke8 5.Kg5 a4 6.f6 gxf6+ 7.Kxf6 a3 8.Ke6 a2 -+ (-5.05) Depth: 12/27 00:00:01 267kN, tb=1 1...Qxd5 2.cxd5 Kg8 3.d6 Kf7 4.f5 a4 5.d7 Ke7 6.f6+ gxf6 7.h5 a3 8.h6 a2 9.d8Q+ Kxd8 -+ (-3.20) Depth: 13/29 00:00:02 499kN, tb=3 1...Qxd5 2.cxd5 Kg8 3.d6 Kf7 4.f5 a4 5.d7 Ke7 6.f6+ gxf6 7.h5 a3 8.h6 a2 9.d8Q+ Kxd8 -+ (-3.60) Depth: 14/35 00:00:05 1712kN, tb=13 1...Qxd5 2.cxd5 Kg8 3.d6 Kf7 4.f5 a4 5.d7 Ke7 6.f6+ gxf6 7.h5 a3 8.h6 a2 9.d8Q+ Kxd8 -+ (-4.60) Depth: 14/35 00:00:05 1738kN, tb=13 1...Qxd5 2.cxd5 Kg8 3.Kf5 a4 4.Ke6 Kf8 5.d6 Ke8 6.f5 a3 7.f6 gxf6 8.h5 b5 9.Kxf6 -+ (-5.30) Depth: 14/35 00:00:05 1807kN, tb=13 1...Qxd5 2.cxd5 Kg8 3.Kf5 a4 4.Ke6 Kf8 5.d6 Ke8 6.f5 a3 7.f6 gxf6 8.h5 b5 9.Kxf6 -+ (-5.70) Depth: 15/35 00:00:06 2392kN, tb=31 1...Qxd5 2.cxd5 Kg8 3.Kf5 a4 4.Ke6 Kf8 5.d6 Ke8 6.f5 a3 7.f6 gxf6 8.h5 a2 9.Kxf6 a1Q+ 10.Ke6 -+ (-6.35) Depth: 15/35 00:00:07 2474kN, tb=31 1...Qxd5 2.cxd5 Kg8 3.Kf5 a4 4.Ke6 Kf8 5.d6 Ke8 6.Kd5 a3 7.Kc6 Kd8 8.f5 a2 9.Kxb6 Kd7 10.f6 a1Q 11.fxg7 Qxg7 -+ (-6.70) Depth: 16/35 00:00:09 3341kN, tb=67 1...Qxd5 2.cxd5 Kg8 3.Kf5 a4 4.Ke6 Kf8 5.d6 Ke8 6.Kd5 a3 7.Kc6 Kd8 8.Kxb6 Kd7 9.Kc5 a2 10.Kd5 a1Q -+ (-6.79) Depth: 17/39 00:00:12 4741kN, tb=150 1...Qxd5 2.cxd5 Kg8 3.Kf5 a4 4.Ke6 Kf8 5.d6 Ke8 6.Kd5 a3 7.Kc6 Kd8 8.Kxb6 Kd7 9.Kc5 a2 10.Kd5 a1Q 11.h5 -+ (-6.76) Depth: 18/39 00:00:18 7488kN, tb=369 1...Qxd5 2.cxd5 Kg8 3.Kf5 a4 4.Ke6 Kf8 5.d6 Ke8 6.Kd5 a3 7.Kc6 Kd8 8.Kxb6 Kd7 9.Kc5 a2 10.Kd5 a1Q 11.h5 -+ (-7.16) Depth: 19/41 00:00:37 16536kN, tb=804 1...Qxd5 2.cxd5 Kg8 3.Kf5 a4 4.Ke6 Kf8 5.d6 Ke8 6.Kd5 a3 7.Kc6 Kd8 8.f5 a2 9.f6 gxf6 10.Kxb6 a1Q 11.Kc5 Kd7 -+ (-7.28) Depth: 19/41 00:00:39 17628kN, tb=935 (Blass, Tel-Aviv 21.06.2004) > >[d]7k/6p1/1p6/p2Q4/2P1qPKP/8/8/8 b - - 0 50 > >After the queen exchange this is the position: > >[d]7k/6p1/1p6/p2P4/5PKP/8/8/8 b - - 0 51 > >White cannot stop the passed pawn on a, but the black king can stop the d-one. >Why some evals work better than others here? > >Jaime I think that it is not simple to see by evaluation that black is winning. For example in the following position the black pawn is not unstoppable. [D]7k/8/1p4p1/p2P4/5P1P/6K1/8/8 b - - 0 1 Some engines including commercial programs and crafty evaluate it as better for black at small depthes. New game, 7k/8/1p4p1/p2P4/5P1P/6K1/8/8 b - - 0 1 Analysis by Crafty 19.13: 1...Kg7 -+ (-6.85) Depth: 1/3 00:00:00 1...Kg7 2.d6 -+ (-6.12) Depth: 2/3 00:00:00 1...Kg7 2.d6 Kf6 -+ (-6.32) Depth: 3/3 00:00:00 1...Kg7 2.d6 Kf6 3.d7 Ke7 -+ (-5.89) Depth: 4/6 00:00:00 1...Kg7 2.f5 gxf5 3.Kf4 a4 4.Kxf5 -+ (-5.12) Depth: 5/7 00:00:00 1...Kg7 2.f5 gxf5 3.Kf4 a4 4.Kxf5 -+ (-5.12) Depth: 6/8 00:00:00 2kN 1...Kg7 2.f5 gxf5 3.Kf4 a4 4.Kxf5 Kf7 -+ (-5.22) Depth: 7/10 00:00:00 4kN 1...Kg7 2.f5 gxf5 3.Kf4 a4 4.Kxf5 Kf7 5.d6 -+ (-4.49) Depth: 8/11 00:00:00 8kN, tb=1 1...Kg7 2.f5 gxf5 3.Kf4 a4 4.Kxf5 Kf7 5.d6 -+ (-4.89) Depth: 9/14 00:00:00 14kN, tb=1 1...Kg7 2.f5 gxf5 3.Kf4 a4 4.Kxf5 a3 5.d6 Kf7 -+ (-5.32) Depth: 9/14 00:00:00 19kN, tb=1 1...Kg7 2.f5 a4 3.fxg6 Kf6 4.h5 a3 5.d6 b5 6.h6 ยต (-1.03) Depth: 10/15 00:00:00 55kN, tb=11 1...Kg7 2.d6 Kf6 3.f5 gxf5 +- (4.45) Depth: 11/19 00:00:01 195kN, tb=28 1...Kg7 2.d6 Kf6 3.f5 gxf5 4.d7 Ke7 5.h5 a4 6.h6 a3 7.h7 a2 8.d8Q+ Kxd8 9.h8Q+ +- (4.22) Depth: 12/23 00:00:01 245kN, tb=42 1...Kg7 2.d6 Kf6 3.f5 gxf5 4.h5 a4 5.h6 a3 6.h7 Kg7 7.d7 Kxh7 8.d8Q a2 +- (4.32) Depth: 13/24 00:00:01 371kN, tb=68 1...Kg7 2.d6 Kf6 3.f5 gxf5 4.h5 a4 5.h6 a3 6.h7 Kg7 7.d7 Kxh7 8.d8Q a2 9.Qf6 +- (4.49) Depth: 14/29 00:00:02 549kN, tb=151 1...Kg7 2.d6 Kf6 3.f5 gxf5 4.h5 a4 5.h6 Ke6 6.h7 Kxd6 7.h8Q b5 8.Kf4 Kc5 9.Qc3+ Kd5 10.Kxf5 +- (6.10) Depth: 15/29 00:00:04 1393kN, tb=230 1...Kg7 2.d6 Kf6 3.f5 gxf5 4.h5 Ke6 5.h6 Kxd6 6.h7 b5 7.h8Q Kc5 8.Qc3+ Kb6 9.Qd4+ Kc6 10.Kf4 a4 11.Kxf5 +- (6.28) Depth: 16/32 00:00:06 2599kN, tb=353 1...Kg7 2.d6 Kf6 3.f5 gxf5 4.h5 a4 5.h6 b5 6.h7 Ke6 7.h8Q Kxd6 8.Qd4+ Kc6 9.Kf4 Kb7 10.Qd6 Ka7 11.Kxf5 +- (6.41) Depth: 17/33 00:00:11 4763kN, tb=469 1...Kg7 2.d6 Kf6 3.f5 gxf5 4.h5 Ke6 5.h6 Kxd6 6.h7 Kd5 7.Kf4 Ke6 8.h8Q Kd5 9.Qe8 Kc5 10.Qe5+ Kc4 11.Qc7+ Kb5 12.Kxf5 +- (6.60) Depth: 18/36 00:00:18 8588kN, tb=708 (Blass, Tel-Aviv 21.06.2004) I think that it is better to be careful and not give unstoppable pawns big bonus unless you are really sure that the opponent does not promote first My evaluation is relatively simple and based on my memory I do not give bonus for unstoppable passed pawn if the opponent has pawn that is more advanced. Uri
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.