Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Crafty Book Implementation

Author: David Rasmussen

Date: 02:56:27 06/12/01

Go up one level in this thread


On June 11, 2001 at 20:52:07, Robert Hyatt wrote:

>>I see now. That is of course a problem. But it would still be nicer, if this was
>>somehow prevented, so one could take full advantage of the opening knowledge in
>>the book. This would allow one to make transpositions into book positions from
>>non-book positions.
>
>
>It can do this.  What my current book logic requires is that the previous
>position must be a legitimate book position that is followed by the current
>book position (to allow the transposition.)  If this fails, crafty will have
>to search normally.  But if It then finds a move that returns it to a book
>position, all will be ok as it will then transpose in for its "next" search
>since this and that position are known book positions.

I know, but then you make the transposition "unintentionally" by search. I think
a scheme where a positions doesn't depend on it's parent position is nicer, as
the engine can use the book knowledge also in the case I mention. I don't really
see how the "going back" in the book, can be a problem, as long as one checks
during book move selection, that this move doesn't get one back in a position
that the game has already been in, that is, repetition check the book moves.



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.