Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Hashkey collisions (typical numbers)

Author: Robert Hyatt

Date: 11:22:27 04/07/04

Go up one level in this thread


On April 07, 2004 at 12:37:47, Renze Steenhuisen wrote:

>No, I do not
>

You _must_.  A position X can have a best move for white to move or for black
too move, depending on how it was searched.  You _must_ factor in side-to-move
or you will have huge problems in endgames where you will miss zugswang
problems.  I simply factor in side-to-move by XORing the hash signature with a
wtm value...  In my case if it is btm, I complement the hash signature.

I'll bet this is producing your illegal hash moves...




>>On April 07, 2004 at 10:24:41, Renze Steenhuisen wrote:
>>
>>Do you store 1 extra bit in hashtable who has the move?
>>
>>>Alright, I promised to get back on this...
>>>
>>>I tested my do_move and undo_move by comparing the incremental hashkey with the
>>>freshly generated hashkey, and guess what: NO PROBLEM! The incremental system is
>>>working just fine.
>>>
>>>Now what?
>>>
>>>I guess I'll be checking my retrieve_data_from_transposition_table() code
>>>again...
>>>
>>>What could be wrong there?
>>>
>>>Any other ideas?
>>>
>>>Cheers!
>>>  Renze



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.