Author: Enrique Irazoqui
Date: 16:39:36 01/13/01
Go up one level in this thread
On January 13, 2001 at 13:04:52, Uri Blass wrote: >On January 13, 2001 at 11:26:32, Frederic Friedel wrote: > >>http://www.computerschach.de/tourn/cad2001/cad01a.htm > >This interesting position was from one of the games between Fritz6 and >Gandalf4.32g > > >[D]5k2/p4pp1/1p5p/1K2P3/5q2/P4P2/1P2Q1P1/8 w - - 0 1 > >Enrique claims that top programs should not play Qe4. Good remark, but unfortunately it wasn't mine. :) IM Hannu Wegner is writing some fine analysis. >The sad truth is that part of the top programs do not know much about pawn >endgames and some of them can avoid Qe4 only by Deep search. > >Only Deep search helps Deep Fritz to avoid Qe4 at depth 14. >Only Deep search helps Gandalf4.32g to avoid Qe4 at depth 13 and unfortunately >it could get only depth 12(Gandalf needs only a few more seconds in Enrique's >hardware to find that Qe4 is losing at depth 13 and immediatly it finds a better >move. Gandalf would need about one minute longer to drop Qe4. That's what I just posted on CSS' forum: If you start analyzing from scratch after 30... Kf8, Gandalf drops Qe4 in 119 seconds on a P750. Instead, let Gandalf compute after 29... Qf4+, wait as in the game until it reaches ply 13, force it to play, enter Fritz' move Kf8 and then Gandalf takes much longer to drop Qe4. That's why it played it in this game. In both cases, don't forget to erase the book.lrn file. Enrique Gandalf 4.32g - Irazoqui,E 5k2/p4pp1/1p5p/1K2P3/5q2/P4P2/1P2Q1P1/8 w - - 0 1 Analysis by Gandalf 4.32g: 31.Kc6 Ke7 ² (0.27) Depth: 2 00:00:00 31.Kc6 Ke7 32.Qb5 ² (0.27) Depth: 3 00:00:00 31.Qe4 ² (0.28) Depth: 3 00:00:00 31.Kc6 Ke7 32.Qb5 ² (0.28) Depth: 3 00:00:00 31.Kc6 Ke7 32.Kb7 Qa4 = (0.22) Depth: 4 00:00:00 31.Qe4 = (0.24) Depth: 4 00:00:00 31.Kc6 Ke7 32.Kb7 Qa4 = (0.24) Depth: 4 00:00:00 31.Kc6 Qa4+ 32.Kc7 Ke7 33.Qd3 Ke6 = (0.17) Depth: 5 00:00:00 6kN 31.Qe4 = (0.19) Depth: 5 00:00:00 6kN 31.Kc6 Qa4+ 32.Kc7 Ke7 33.Qd3 Ke6 = (0.19) Depth: 5 00:00:00 6kN 31.Ka6 = (0.19) Depth: 5 00:00:00 6kN 31.Kc6 Qa4+ 32.Kc7 Ke7 33.Qd3 Ke6 = (0.19) Depth: 5 00:00:00 6kN 31.Kc6 Ke7 32.Kb7 Qa4 33.Qa6 Qc2 34.Qxa7 Qxb2 = (0.14) Depth: 6 00:00:00 32kN 31.Ka6 = (0.16) Depth: 6 00:00:00 32kN 31.Ka6 Qa4+ 32.Kb7 Qd7+ 33.Kb8 a5 34.Qe3 b5 35.Qf4 ² (0.27) Depth: 6 00:00:00 32kN 31.Qe4 ² (0.28) Depth: 6 00:00:00 32kN 31.Ka6 Qa4+ 32.Kb7 Qd7+ 33.Kb8 a5 34.Qe3 b5 35.Qf4 ² (0.28) Depth: 6 00:00:00 32kN 31.Ka6 Qa4+ 32.Kb7 Qd7+ 33.Kb8 a5 34.Qe3 b5 35.Qc5+ Ke8 ² (0.27) Depth: 7 00:00:01 81kN 31.Qe4 ² (0.28) Depth: 7 00:00:01 81kN 31.Ka6 Qa4+ 32.Kb7 Qd7+ 33.Kb8 a5 34.Qe3 b5 35.Qc5+ Ke8 ² (0.28) Depth: 7 00:00:01 81kN 31.Ka6 Qa4+ 32.Kb7 Qd7+ 33.Kb8 a5 34.Qe3 b5 35.Qc5+ Ke8 ² (0.27) Depth: 8 00:00:01 181kN 31.Qe4 ² (0.28) Depth: 8 00:00:01 181kN 31.Qe4 Qxe4 32.fxe4 Ke7 33.Ka6 Ke6 34.Kxa7 b5 35.Kb6 ² (0.30) Depth: 8 00:00:01 181kN 31.Qe4 Qxe4 32.fxe4 Ke7 33.Ka6 Kd7 34.Kxa7 Kc6 35.Kb8 Kc5 ² (0.30) Depth: 9 00:00:03 327kN 31.Qe4 Qxe4 32.fxe4 Ke7 33.Ka6 Kd7 34.a4 Kc7 35.Kxa7 Kc6 ² (0.46) Depth: 10 00:00:10 1103kN 31.Qe4 Qxe4 32.fxe4 Ke7 33.Ka6 Kd7 34.a4 Kc6 35.b4 h5 ² (0.46) Depth: 11 00:00:21 2284kN 31.Qe4 Qxe4 32.fxe4 h5 33.Kc6 Ke7 34.Kb7 Kd7 35.a4 h4 ² (0.46) Depth: 12 00:00:59 6395kN 31.Qe4 Qxe4 32.fxe4 h5 33.Kc6 Ke7 34.Kb7 Kd7 35.a4 h4 = (0.15) Depth: 13 00:03:17 20797kN 31.Qe4 Qxe4 32.fxe4 h5 33.Kc4 g5 34.Kd4 h4 35.Ke3 g4 ³ (-0.68) Depth: 13 00:03:35 20797kN 31.b4 ³ (-0.67) Depth: 13 00:03:37 20797kN (Irazoqui, Cadaqués 14.01.2001) >Hiarcs7.32 also likes Qe4 and I guess that only deep search can teach it to >change it's mind. > >Qe4 is a losing move and humans have no problem to see because of the following >selective search: > >Qe4 Qxe4 fxe4 h5 Ka6 h4 Kxa7 g5 Kxb6 g4 a4 h3 when it is clear that the black >pawn queens first. > >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.