Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: repetition detection & hashing now with text

Author: Robert Hyatt

Date: 13:04:34 01/16/02

Go up one level in this thread


On January 16, 2002 at 15:26:29, David Hanley wrote:

>On January 16, 2002 at 14:08:31, Robert Hyatt wrote:
>
>
>>But suppose you have _already_ repeated it twice?  You entered this line
>>thinking the score was +x.xx due to the hash entry but before you reach that
>>+x.xx position you are forced to step thru a 0.00 position that ends the game
>>instantly...  and incorrectly according to the hash score of +x.xx
>
>Ah.. So let me rephrase for my own edification.
>
>A position has been repeated twice already, and the program though it was good
>before and put it in the hash table.
>
>*NOW* it sees a win from that position where it didn't see it before, and plays
>to it, seeing from the hash that it can win.
>
>I'm still confused i guess, beause if it didn't hash it as a desirable line
>before, why would it think it would be now?  Wouldn't it have to create a new
>winning line *past* the hash move?  Wouldn't this always re-search that position
>in question, triggering the fraw detection?
>
>dave


Remember that not _all_ programs consider two-fold repetition as a draw, they
wait for the 3rd repetition to match the rules of chess.  This means that most
programs that handle three-fold correctly will not be too concerned about
repeating the second time as they know that doesn't hurt, just so long as they
don't repeat the third time.  But due to hashing, you can see a draw that is
not there, or miss a draw that is, because hashing only hashes _position_
information, yet the actual game value depends on positions of all the pieces
_plus_ information about all the moves played in the game so far.




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.