Author: Uri Blass
Date: 05:01:29 12/16/00
Go up one level in this thread
On December 16, 2000 at 04:22:27, Georg Langrath wrote: >This is a very interesting position in my opinion. I showed it in this forum a >couple of years ago. But there must be some new members in this forum since >then. It is from a match between Spassky and Tal 1959. >[D]6k1/8/p3r2p/1p1pPpp1/1n1pP3/1Pq4P/PR4PB/3QN1K1 b >Tal played Qxb2 and lost. The only right move here is dxe4 >It was showed in a German magazine for chesscomputers I think for about 12 years >ago. None of the chesscomputers on that time managed to find the right move, >even if you let them analyze for days. Today many programs find the right move >in a couple of seconds. But not all. Chessmaster 6000 didn’t find it overnight >on my computer (P2 450 mhz). It persisted in playing Qxb2. >On the other hand Tiger ll doesn’t even consider Qxb2, which I find remarkable. >It finds dxe4 at once and doesn’t change its mind. >Very interesting is that this is a question of better programs. I mean if a >program today find the right move in seconds, the same program had found the >right move in days if it was in a chesscomputer of 1988. >It is also very interesting to play eng-eng after the right move and the wrong >move. >I think that also many readers of this message will find this position very >amusing. > >Georg I think that this is a question of evaluating connected passed pawns. I think that Tiger evaluates them too much and this is the reason that it does not consider Qxb2 as the best move. I remember a game when it lost inspite of clearly winning evaluation against Gandalf because of this reason(it was KQBPPPP vs KQBPPPP endgame if I remember correctly and Tiger could win a pawn but did not want to do it because it did not want to be without the connected passed pawns). I suggest that you try to increse the evaluation of pawns for chessmaster. I guess that the personality with 1.1 for a pawn instead of 1 for pawn is going to find it and if you increase the value of pawn more than it then it is going to be faster in finding it. I did not try it and it is only a guess. I expect Junior5 to be faster than Junior6 because Junior5 has a bigger value for pawns relative to other pieces. 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.