Author: Uri Blass
Date: 23:54:32 10/02/01
Go up one level in this thread
On October 02, 2001 at 20:52:13, TEERAPONG TOVIRAT wrote: > >Dear Dr.Hyatt, > >I think the number of total _possible_ positions in chess is beyond >the question. However,I'd like to know how they calculate it ? > >A few months ago,I tried to work out the number in _checkers_ >game. My logical assumption is that if the hash signature is too >small there would be a high incidence of type 1 collision and if >the hash signature is large enough the collision would be 0. >So,my experiment started with 32 bit hash table ,after 1 million >entries are fully occupied I started counting the incidence of >type 1 collision. Each time I got type 1 error,I increase another >bit to the hash table.I saw the incidence is lowered by half too. >Until one point the incidence is zero ,of course after a few >games test. > >Could I claim that at that point the hash signature is approximately >equal to the number of legal(possible) positions? I do not know what is type 1 collision and I also do not understand what hash tables can do with the number of legal positions because it is known that the search trees of programs contains only a very small part of the number of legal positions. Uri
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.