Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Fruit 2 and endgame play

Author: Dann Corbit

Date: 11:18:08 01/13/05

Go up one level in this thread


On January 13, 2005 at 10:20:58, Dieter Buerssner wrote:

>On January 12, 2005 at 13:52:16, Dann Corbit wrote:
>
>>I think it wise to test at least the pv also, when the pv has 5 men or less in
>>it.  Even if it is 100 plies long, that is only 100 probes and completely
>>insignificant for time.
>
>What would you do with the info? Say search/eval shows score +3.2; the PV goes
>to TB and says draw? I don't see anything one could do here - it is too late.

The information will still be valuable for knowing if you should accept a draw
or should deepen your search with a fail low, I think.

>It
>is also likely, that if you remembered that position that gave you the 3.2
>score, and you hardcoded it for a 0.0 score, you'd get some slightly different
>PV, again with a 3.x score. Perhaps again leading to a TB draw, perhaps to a
>win. You'd only know it, if you'd probed during search.

You may use the TB score to indicate that you need to probe (at least) all the
alternative root or attractive root + 1 moves with (TBCOUNT) pieces or less on
the board.  Or it may be an indication that your search needs to extend.

>BTW, It is only one probe, because only the first time your reach a TB-position
>in the PV is interesting.

You may be missing many tablebase files and all the probes may fail except the
last one.

>Regards,
>Dieter



This page took 0.01 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.