Author: Dave Gomboc
Date: 15:58:04 03/18/02
Go up one level in this thread
On March 17, 2002 at 16:29:21, Dieter Buerssner wrote:
>On March 17, 2002 at 15:46:05, Dave Gomboc wrote:
>
>>On March 17, 2002 at 03:14:37, Lyn Harper wrote:
>>
>>> These chess programs never cease to amaze me. They can see a combo
>>> a dozen plies deep, and in this case miss the simple Rg2. I'm not s
>>> strong player, but Rg2 is pretty obvious. White gets a passer that
>>> can't be stopped, while black gets one that's harmless. What's the
>>> problem? Don't passers get the respect they deserve?
>>
>>Perhaps they do not fully appreciate how pitiful the knight is against the rook
>>pawn on the far side in their static evaluator.
>
>I wonder, if you all can see immediately, that the knight cannot stop the pawn?
>I can't ... Or do you also need some concrete calculation? Can the knight side
>gain some tempos by intermediate checks. Also (for an engine) some other
>questions must be answered: Is perhaps my own king in the way of the pawn, can
>it be forced into the way of the pawn? Or can I just count pawn moves to the
>promotion, and perhaps use some table, to see, at which squares of pawn/knight
>the position for the knight is helpless?
>
>Regards,
>Dieter
No, I can't see immediately that the knight cannot stop the pawn. But I can see
immediately that:
1) each of the two split passers ties down one of Black's pieces
2) the black king simply can't get over to stop the h-pawn
3) the black knight is going to have a very hard time stopping the h-pawn
4) the only safe and fast route to the kingside for the knight is
Nb5-c7-e8-g7, which is ineffectual since the pawn advances too
quickly
5) Therefore Rxg2 wins
That white doesn't need to take time out to go and munch the g2-pawn to free up
his bishop from patrol duty is just a bonus. A human doesn't need to do a
full-width search here to prove that white wins, which is why it is 'obvious'.
The tree of variations is practically linear, and the moves are so forcing that
it's a serious mistake to do a static eval at anywhere along the way once Rxg2
has been played.
Dave
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.