Author: Uri Blass
Date: 12:39:52 09/03/01
Go up one level in this thread
On September 03, 2001 at 14:58:36, Joshua Lee wrote: >What are the best results anyone has got on this test set? > >Most top programs get 14 or 15 out of 33 however they all solve some problems >but change their mind. Get this Hiarcs 7.32 which is supposed to be the more >positional program did solve some positions the fast searchers couldn't but >still didn't do any better. > >I would like to see results for Nimzo 3 or 3.5, Nimzo 7.32,98, 2000, >Shredder,WChess, CSTal, Zarkov etc. I am beginning to believe that nobody is >going to do much better. This set is also only the tip of the iceberg as the >gelfer positions are just as hard if not harder.(Positional Chess Handbook the >reprint is going to be out in November 2001) The main problem with positional tests is that I doubt if the solutions are right. If programs cannot solve more positions with more time then it suggests that the solutions are wrong and the humans who looked at the positions simply did not search deep enough to find the right moves. I believe that it is better to use games of good correcpondence players in order to have a better positional test because good correspondence players give their computer to anlayze for hours and they understand the fact that humans do not know much. I am almost sure that I am going to be the Israeli champion in correspondence games(I have 4 out of 5 and I have a clearly winning position in 2 of the 3 unfinished games) I can give a lot of non trivial moves from my games as a positional test. I believe that if I do it you are going to get a better positional test. Note that I almost always choose to believe one of my programs but I need to choose which program to believe. I usually use deep Fritz but if I do not like Deep Fritz's ideas I try also other programs. I saved a lot of analysis from my correspondence games Here is one example of positional test from my game against the female world champion in correspondence games In this case I chose gandalf's move and Deep Fritz prefered the wrong Nxg4 even after many hours. Luba kristol - Uri blass [D]r2qkb1r/5pp1/p3p2p/4n3/1p1NP1P1/4B3/PPPQ3P/2KRR3 b kq - 0 1 Analysis by Gandalf 4.32h: 17...Nxg4 18.Rg1 Nxe3 19.Qxe3 Qc7 20.Qe2 Qf4+ 21.Kb1 g6 = (0.20) Depth: 7 00:00:02 250kN 17...Bc5 = (0.19) Depth: 7 00:00:04 500kN 17...Bc5 18.Qg2 Qh4 19.h3 0-0 20.Rf1 a5 = (-0.04) Depth: 7 00:00:06 1100kN 17...Bc5 18.Bg1 Qc7 19.Qg2 0-0 20.g5 hxg5 21.Qxg5 a5 = (-0.14) Depth: 8 00:00:15 2900kN 17...Bc5 18.Qf2 Nxg4 19.Qg3 Nxe3 20.Qxg7 Nxd1 21.Qxh8+ Kd7 22.Qxd8+ = (-0.17) Depth: 9 00:00:55 11600kN 17...Bc5 18.Qf2 Qa5 19.Kb1 0-0 20.Rg1 Nc4 21.g5 e5 22.gxh6 ³ (-0.27) Depth: 10 00:02:10 28950kN 17...Bc5 18.Qe2 Qb6 19.Rd2 0-0-0 20.Red1 Rd7 21.h3 Rhd8 22.Qf2 = (-0.24) Depth: 11 00:06:15 83400kN 17...Bc5 18.Qf2 Nxg4 19.Qg3 Nxe3 20.Qxg7 Nxd1 21.Qxh8+ Kd7 22.Qxd8+ ³ (-0.41) Depth: 12 00:16:26 221750kN 17...Bc5 18.Qf2 Qa5 19.Nb3 Bxe3+ 20.Qxe3 Qxa2 21.Qc5 f6 22.Rf1 ³ (-0.38) Depth: 14 02:26:59 1984700kN 17...Bc5 18.Qf2 Qa5 19.Nb3 Bxe3+ 20.Qxe3 Qc7 21.Rf1 0-0 22.g5 ³ (-0.33) Depth: 15 05:51:46 380732kN (Blass, Tel-Aviv 01.05.2001) Uri
This page took 0.01 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.