Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The Book Problem, Open Book Format proposal, Request for comments.

Author: Dann Corbit

Date: 13:59:18 03/06/06

Go up one level in this thread


On March 04, 2006 at 20:01:52, Kenny Dail wrote:

>On March 04, 2006 at 15:05:59, Dann Corbit wrote:
>
>>On March 04, 2006 at 07:25:23, Kenny Dail wrote:
>>
>>>I kind of think the idea of a GUI book is currently a silly implementation and
>>>would like to see a better standard appear.
>>>
>>>I would not like it if I went to a tournament and the TD came and told me which
>>>moves I was going to play. For me at least, opening selection and preparation is
>>>part the game of chess, by forcing me to play someone else's opening lines, is
>>>more like playing a chess variant. I don't know of any chess organizations that
>>>allow "thematic" tournaments to be figured into a person's rating.
>>>
>>>To make a book useful, (GUI or otherwise) there needs to be input from the
>>>player. At the very least presenting the list of "acceptable" moves to the
>>>player and letting it select the one it likes. Thus the player can try to steer
>>>towards games that fit it's style.
>>>
>>>Ideally a good book setup will allow the engine to add new lines, or break from
>>>the book if it doesn't like any of the available lines. For example if the book
>>>supplied has only 1 e4  available, then engine should still be allowed to play 1
>>>d4, and add that line to the book for future reference.
>>>
>>>For a book that is going to be used by multiple engines, there shoulb be engine
>>>specific tags available, so each engine can keep it's own notes as well be able
>>>to access global stats.
>>
>>You make an interesting point.  There should be two different types of book
>>information:
>>Thematic moves (you MUST perform the requested move)
>>Suggested moves (the book will make a suggestion that you can take or leave).
>>
>>I remember that Crafty used to analyze book moves at 10% or so just to make sure
>>that it wasn't about to do something silly based on a bad stored move.
>>
>>Arena has a checkbox for "own book" so you can use an Arena book, or allow the
>>program to use its own book.
>>
>>The problem addressed by this thread is not necessarily about book utilization,
>>but in creating a better book for use by amateur chess engine authors and
>>amateur book authors.  I think that the Rybka book notes are revealing.  The
>>ChessBase format was chosen because currently, the book author is unaware of any
>>decent tools for making a good book in some other format.  *That* is the problem
>>that is being addressed.
>
>Well, then to try and put this into relevance, If the GUIs would cater to the
>needs of the engines as outlined above, then there wouldn't be a need to build
>engine specific books. Instead the engine could train the GUI. The author would
>then only have to use an API and supply a sample of games for the engine to
>parse into the book.

The database approach does not preclude that sort of usage.  The opening book
could just as easily be bolted to a GUI as to an engine.  The tool produced will
be totally generic.

What the output of the project will be is a permanent store for chess
information that is carefully spelled out to contain the things needed for good
decisions.

Whether it is a chess GUI or a chess engine or something else that uses this
information is not relevent to the project itself.



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.