Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: hash table & draw detection

Author: Dieter Buerssner

Date: 16:35:02 05/02/02

Go up one level in this thread


On May 02, 2002 at 14:52:22, Uri Blass wrote:

>On May 02, 2002 at 11:48:59, Robert Hyatt wrote:
>
>
><snipped>
>>It doesn't.  Here is a better explanation...
>>
>>
>>
>>root----------YYY----------------------XXX--------------TIP
>>
>>root-------------------XXX-----------------YYY----------------XXX-------TIP
>>
>>You search the first path first.  And when you finish searching position
>>YYY, you store a score of +1.2...  Which represents what happens between YYY
>>and the TIP position.
>>
>>You search the second path next.  When you reach position YYY you retrieve
>>that +1.2 score and use it.  It is _wrong_.  Because position XXX will be
>>reached before the tip position is reached, producing a correct score of
>>draw.  But you are claiming +1.2 which is wrong...
>
>The only way to use hash tables without these problems seems to be to store
>a score for yyy in the hash tables only in cases that the last move before yyy
>is a conversion and to look at hash tables in order to cut the tree only after
>conversions(you can only use hash tables not after conversions but in this case
>using hash tables should be done only for better order of moves and not to cut
>the tree).

You can also use hash tables in the other cases for move ordering. I have no
numbers, but I think in the middle game move ordering is a significant part of
the strength increase of hash tables. Also, hashing the path since the last
irreversible move, would be enough, but would probably in most cases make hash
tables almost useless. One could have two hash keys, one including the path
since the last irreversible move, that can be safely used for cutoffs from the
HT. The other, without this path information, that would only be used for move
ordering. I would guess, that it is not worth the trouble.

>I do not say that it is a good idea but if someone wants to avoid the errors and
>still use hash tables then it seems to be the only way.

These errors are "self healing" (no idea, if this is a correct English phrase).
Sooner or later you will not have enough draft in the hash table to allow a
cutoff, and must search it.

A good reading about this is also Dennis Breuker's thesis, which contains a
chapter about this. http://members.ams.chello.nl/dbreuker/thesis/index.html

Regards,
Dieter



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.