Author: Carmelo Calzerano
Date: 10:12:52 03/26/01
Go up one level in this thread
On March 26, 2001 at 12:43:57, Miguel A. Ballicora wrote:
>On March 26, 2001 at 10:47:15, Carmelo Calzerano wrote:
>
>>On March 26, 2001 at 10:41:32, Ulrich Tuerke wrote:
>>
>>>On March 26, 2001 at 09:53:55, Paul wrote:
>>>
>>>
>>>Do you think that this is simpler than just comparing the hash codes ?
>>>I don't.
>>
>>Of course it isn't; but it has to be implemented in a program that
>>doesn't implement hash tables yet...
>>
>Hi Carmelo,
>
>You don't need hash tables to implement a "hash signature based" method such as
>number 2 mentioned by Bruce Moreland. You just need to update a hash signature
>every makemove() with the Zobrist algorithm. It is not more difficult and it
>will pay off in the long run. I implemented it like that in my program long
>before having hash tables ("transposition/refutation tables").
>When I decided to include hash tables, I already had the functions that
>calculated and updated the hash signature.
>On the other hand, the algorith posted is not perfect at all. That worries me
>not because the strength of the program will decrease, but for debugging
>reasons...
I perfectly agree with you.
In fact I'm using HT in my own program (Leila) since the beginning; the
algorithm is needed by some of our friends in the g_6 (the italian chess
programmers group). I tried to convince them to implement a simple hash-based
scheme, providing some sample code too; but unfortunately, I had no success!
:-(
Thanks,
Carmelo
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.