Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: A question on Opening book creation?

Author: Uri Blass

Date: 14:45:32 08/30/05

Go up one level in this thread


On August 30, 2005 at 12:46:11, J. Wesley Cleveland wrote:

>On August 30, 2005 at 12:21:10, Gopikrishna wrote:
>
>>Gerbil creates opening book by hashing the position and comparing the hash
>>values. But tscp does by getting moves from structures and comparing it with
>>book lines. I guess this approach wont make it slow right? Is there any
>>difference in performance? Since I would like to decide on which way to go...
>>Thanks.
>
>The hashing method allows the program to find transpositions, which is generally
>good, but can have odd side effects. E.g 1. e4 e5 2. Nf3 Nc6 3. Ng1 Nb8 ?!?!

odd side effects is not because of comparing hash values but because of trying
to use hash for tranpositions in not book positions.

I also compare hash values and have not that problem.

After 1.e4 e5 2.Nf3 Nc6 3.Ng1 black is out of book because the hash key of the
position after 3.Ng1 is not in the book so Movei does not look at the book.

After 2.Nf3 the hash key of the position is in the book so Movei looks at the
book and find a move(can be Nc6 or Nf6 or other moves if they are in the book)
and it plays the move.

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.