Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: position, avoid move

Author: Andrew Dados

Date: 23:00:43 12/11/01

Go up one level in this thread


On December 11, 2001 at 20:47:13, Will Singleton wrote:

>On December 11, 2001 at 16:54:50, Will Singleton wrote:
>
>>On December 11, 2001 at 14:13:10, Andrew Dados wrote:
>>
>>>On December 11, 2001 at 10:59:55, Will Singleton wrote:
>>>
>>>>[D]4k2r/1nrqbppp/p3b3/1pppP1NQ/8/1NP1B2P/PP3PP1/R4RK1 w k - 0 1
>>>>
>>>>Most programs don't immediately see that Nxh7 puts the knight in a bad spot.
>>>>Can this problem be addressed by root analysis?
>>>>
>>>>(Beowulf might be doing something like that, since it never considers Nxh7.)
>>>>
>>>>Will
>>>
>>>consider this:
>>>
>>>[D]8/6pN/5p2/8/8/8/8/8 w - -
>>>
>>>The only square N can escape through is c8...
>>>So you may treat N on a/h 7th rank as vulnerable as bishop and just set your
>>>squaretable value for N to - 1 pawn there (or more). If search can see N will
>>>leave h7 safely - you are pawn up; if it stays there - it is probably in
>>>trouble.
>>>
>>>-Andrew-
>>
>>I do something similar for B & N, but only for the root move.  Perhaps your
>>method is simpler and better.
>>
>>Will
>
>I tried your approach, it didn't work.  Somehow the pv has the knight moving off
>the square, but I didn't really analyze it.  Seems like it should work, but it
>didn't for me.
>
>How does your prog handle the position? (Assuming your approach is implemented.)
>
>Will

I lost some files in HD crash, so my proggy is currently not functional. But it
was doing ok in similar tests. When I get it back to work I will test that one.

-Andrew-



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.