Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Pawn Hashkey Size

Author: David Rasmussen

Date: 06:52:49 12/04/01

Go up one level in this thread


On December 04, 2001 at 08:32:49, Sune Fischer wrote:

>On December 04, 2001 at 07:33:08, David Rasmussen wrote:
>
>
>Hmm, how many different pawn positions do you get?

I don't know. It's not important (unless there's a bug). The important thing is
that every time I probe successfully and thus return "early" without actually
calculating the pawn score, there is a probability of about 5E-6 that the answer
is wrong. Likewise for Crafty.

>If you get millions it makes sense, but not if you only have a few thousand.
>The odds of a collision between a few thousand when you have 4 billion slots is
>not that great (albeit not neglible), it is the famous "birthday problem".
>
>I'm not clear on what you mean by:
>"And now, I am only talking about the final score value, not the pawn structure
>properties that Crafty also stores in the table."
>
>I only pawn hash the pawn structure scores (like double, passers, connected,
>isolated etc..) which is why I get so few different positions.
>In more than 999 times of 1000 I get a hit.
>
>-S.

What I meant is, that in Crafty, some properties of the pawn structure is saved
in the table, but an evaluation score for the pawn structure is also saved. On
doing this experiment, I only checked if the scores differed, not if any of the
properties differed. One can easily imagine two different pawn structures, that
have the same score, but where some of the properties differ. I don't catch
these with this experiment. I only catch the ones where the score differ. The
score is only one field of Hyatt's pawn hash entry. There are 5 others or so.
And these I don't check in this experiment. If I did, there might be even more
collisions.

/David



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.