Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: zobrist key table questions

Author: Tom Likens

Date: 06:31:28 05/31/04

Go up one level in this thread


On May 30, 2004 at 23:56:57, Russell Reagan wrote:

>On May 30, 2004 at 09:56:39, Tom Likens wrote:
>
>>And of course, if you want to start playing with the Hamming distance
>>it's easy.  Of course, there is no real consensus (as far as I can tell)
>>on wheter numbers with a larger Hamming distance are better or not- so,
>>as always you should experiment and see what works best for movei.
>
>Can we define what "better" is? Before we can develop a test to determine
>whether a larger hamming distance is better, what are we trying to achieve? Are
>we trying to spread the positions out more evenly over the hash table? Or is our
>objective to have fewer full 64-bit hash key collisions? Or something else? I've
>read these "larger hamming distance is better" threads in the past, but I never
>really thought about what "better" meant :)

Hello Russell,

My first take on this, would be to define "better" as fewer collisions.
It would be interesting to plot the number of collisions vs. the
average/minimum Hamming distance of the random numbers (of course, as
Dieter, is likely to point out, calling these numbers "random" after
sifting them to produce a certain Hamming distance is a bit of a stretch)!

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.