Author: Uri Blass
Date: 08:26:57 11/29/04
Go up one level in this thread
On November 29, 2004 at 10:56:13, Joachim Rang wrote: >On November 29, 2004 at 08:20:38, Franz Hagra wrote: > >>see http://f23.parsimony.net/forum50826/messages/113535.htm >> >>r1b1n2r/1p2q3/1Qp1npk1/4p1p1/P1B1P3/2P1BNP1/1P3PK1/R3R3 b - - 0 22 >> >>best move is not Nf4 - better to play first Nd6 and then Nf4 >> >>This is a main problem test positions - there is always the possibility that the >>solution move must not be played imidiatly and in this case with nowadays >>testsuits and GUI's its not possible to test the fact that the engine is >>"understandig" the position and here motives! >> >>eg: an engine finding Nf4 and not Nd6 and than Nf4 would have better test >>results = the engine seeing less earns more points! >> >>Hagra > >Thanks Hagra for your confidence in my hurried analysis. Nvertheless it is not >yet proven, that Nd6 is better, perhaps both moves are equal. It's a very >complicated and unclear position. I would appreciate any analysis regarding this >position. > >But I thank you for pointing out the quality of the WM-Test, where in the first >50 Positions only one position was found to be unclear. I too think this is a >pretty good percentage for a testsuite. > >regards Joachim 1 position out of 50 that was proved to be unclear does not prove that the rest are clear. I do not know if people looked at the rest of the positions to decide if they are clear or not clear. I remember that in the past I looked at part of the problems and was not convinced that the solutions are correct. I was not convinced also about the opposite but as far as I remember I could not teach yace that the solutions are correct in a short time by going forward and backward. 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.