Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Problem with TT

Author: Tom Likens

Date: 22:20:13 07/11/03

Go up one level in this thread



On July 11, 2003 at 18:36:18, Grzegorz Sidorowicz wrote:
>
>> Normally, when you fail-low
>>there is no move to store into the hash table.  Some programs
>>do store a move for an upper bound node (Yace does it I believe)
>>but the more vanilla hash implementations usually don't.
>
>In your opinion what I must change in this code and why.
>I'm a little don't understend. I must store only moves for upper bound??

A position can only fail low (i.e. the score <= alpha) if all the
nodes below it failed high, which means that their scores all exceeded
or equaled beta so that no best move was backed up.  In this case you
don't have a 'best' move to store in the hash table.  Note, a fail-low
gives us an UPPER bound on the position.  In other words, the true score
could be worst but no better.

>>
>>Is this the best move found by the search?  Again if this is
>>an upper bound position then there is (probably) no best move
>>to store.
>
>Yes.
>If I good understand I must store bounds but not move index?
>If bestvalue is lower than original alpha then best move is
>fisrt move and it is probably not true.
>Well, maybe I don't need store nodes with
>condition bestvalue<=original_alpha. What do you think about it?
>I was try it but my program still can't find solution

Stated slightly different, if after searching all the moves at a
node the score is less than or equal to alpha then we failed low,
which means that every move tried in this position was refuted at
the next level.  Since we searched *all* the moves and they all
returned alpha (or less) we don't have any idea which one is best.

regards,
--tom



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.