Author: Uri Blass
Date: 13:27:29 02/20/01
Go up one level in this thread
On February 20, 2001 at 13:01:44, Chessfun wrote:
>On February 20, 2001 at 08:19:09, Vincent Lejeune wrote:
>
>>On February 19, 2001 at 19:48:05, Chessfun wrote:
>>
>>>
>>>4rk2/4qp2/5n2/p3NBQp/2P1b3/1P4P1/P3K3/8
>>>
>>>This position is from my Valentine tourney from round 20
>>>Gambit 1.0 v Shredder 5. Gambit choose the rep by 48. Qh6
>>>instead of what Shredder expected and appears to be winning
>>>48. Nd7+ Qxd7 49. Bxd7 Nxd7 50. Qxa5 Bg6+.
>>>
>>>After manually playing the moves Gambit sees white winning
>>>but even to depth 18 will not play 48. Nd7+
>>
>>
>>[d]4rk2/3n1p2/8/Q6p/2P1b3/1P4P1/P3K3/8 b - - 0 50
>>
>>
>>Position from your line, before 50..Bg6+
>>
>>Here I would definitivenely consider 50..f5 to let the bishop on the diagonal
>>where the 3 white pawns could advance
>>(but f5,Bg6+,f6 or Ne5 are in same score range (+0.50 to +0.60) for hiarcs 7.32)
>>
>>
>>Now the score of 2 engines in this position :
>>
>>Crafty 18.02 sees black better : -1.40 (50..f6)
>>hiarcs 7.32 sees white better : +0.50 (50..f5 and later 50..f6)
>>
>>
>>
>>This position is very difficult to estimate "R+N+B+2p" against "Q+3 linked
>>passed pawns+1p" and the 2 kings are active
>>
>>Here the help of a strong correspondance player who can analyse deeply would be
>>interresting !
>
>Actually what I found odd was Gambit sees the line the move before
>as can be seen from it's PGN. 47.Re4 { Score: 0.20 depth: 14 Re4 Bxe4
>Nd7+ Qxd7 Bxd7 Nxd7 Qxa5 Bf5+ Kd2 Bg4 a3 Re2+ Kc1 Ne5 Qd8+ Kg7 c5 Kh7 Qf8 }
>And making the first few moves in the line which are pretty much forced isn't
>very deep in terms of search depth 48. Nd7+ Qxd7 49. Bxd7 Nxd7 at this point it
>sees white winning, yet still plays the rep instead at 48. You would think if it
>sees the few moves ahead to the position after 49....Nxd7 which it evals as a +
>for white it would have played 48. Nd7+.
I do not find it as odd because I know that tiger is a preprocessor so the
evaluation function is dependent on the root position.
I guess that tiger see the line that it evaluated as score +0.20 when it played
47.Re4 but the evaluation of it is negative so it prefers the repetition line.
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.