Polyglot opening book specification

Discussion of chess software programming and technical issues.

Moderators: hgm, Dann Corbit, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Michel
Posts: 2210
Joined: Sun Sep 28, 2008 11:50 pm

Polyglot opening book specification

Post by Michel » Tue Nov 11, 2008 10:29 am

I have written out the Polyglot opening book specification. You can find it on my Toga website

http://alpha.uhasselt.be/Research/Algebra/Toga/

There are already some tests. When I find a bit of time I will also provide sample code.

In any case feedback is appreciated.

Michel

Felpo

Re: Polyglot opening book specification

Post by Felpo » Tue Nov 11, 2008 10:44 am

Thanks a lot,
I just needs some good random for Zobrist Key generation, I think using exactly that strategy in an engine code would reduce errors in comparison with an "home made" random generator. Am I wrong ?
And, furthermore, is there some permission needed to grap these random and using in an open source engine as mine ?
Thanks

Felpo

Re: Polyglot opening book specification

Post by Felpo » Tue Nov 11, 2008 10:52 am

Sorry for my question before...
I just noticed that at the beginning of your link you talk about copirights...

krazyken

Re: Polyglot opening book specification

Post by krazyken » Tue Nov 11, 2008 3:01 pm

I did not see mention of color support in the book spec. So with this format it is not possible to specify opening to be played only as white?

ernest
Posts: 1928
Joined: Wed Mar 08, 2006 7:30 pm

Re: Polyglot opening book specification

Post by ernest » Tue Nov 11, 2008 4:11 pm

Michel wrote:I have written out the Polyglot opening book specification.

Any comment by Polyglot expert Marc Lacrosse?

Dann Corbit
Posts: 12035
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: Polyglot opening book specification

Post by Dann Corbit » Tue Nov 11, 2008 7:56 pm

I find the polyglot book specification lacking.

It's better than a lot of book formats, but I think this is better:
http://cap.connx.com/openingbook.sql

ER/Win model:
http://cap.connx.com/OPENINGBOOK.ER1

Even if you do not like this format or think it is too complex, I think you will immediately see things that it contains that are completely lacking from the polyglot format.

If you add some of the ideas from this model to the polyglot model, I think it will improve it.

Michel
Posts: 2210
Joined: Sun Sep 28, 2008 11:50 pm

Re: Polyglot opening book specification

Post by Michel » Tue Nov 11, 2008 8:15 pm

I have posted a new version as I had forgotten to indicate how castling moves are represented.

See here

http://alpha.uhasselt.be/Research/Algebra/Toga/

Michel

Michel
Posts: 2210
Joined: Sun Sep 28, 2008 11:50 pm

Re: Polyglot opening book specification

Post by Michel » Tue Nov 11, 2008 8:35 pm

I did not see mention of color support in the book spec. So with this format it is not possible to specify opening to be played only as white?
I don't quite understand what you mean. The color is encoded in the position.
Can you explain?
If you add some of the ideas from this model to the polyglot model, I think it will improve it.
I have no intention of inventing a new opening book format. Nobody would use that
anyway.

The PG format is of course not perfect but

(0) It is not bad either.
(1) It is trivial to parse.
(2) It is understood by a number of important programs in the computer chess world.
(3) There are some good books already available (thanks Marc Lacrosse).

This being said it is possible to extend the PG format in a backwards compatible
way by adding "magic entries".

Michel

Marc Lacrosse
Posts: 511
Joined: Wed Mar 08, 2006 9:05 pm
Contact:

Re: Polyglot opening book specification

Post by Marc Lacrosse » Tue Nov 11, 2008 9:50 pm

ernest wrote:
Michel wrote:I have written out the Polyglot opening book specification.

Any comment by Polyglot expert Marc Lacrosse?
Hmm I have no qualification for the programming aspect : I am not competent.

From the book-user or book-maker point-of-view I am very happy that someone helps develop PG books use.

From my little experience : Polyglot books are really outstanding :
... they are small
... the book-making options are very sophisticated and powerful, and book-making heuristics is extremely efficient
... we were missing book-editing functions, but with latest SCID versions by P George we have now a powerful interactive way for modifying move weights in PG books

Michel just announced on the WB forum that he would work on additional book-editing utilities.

The main annoying point with Polyglot as a book-manager lies in the fact that it is a uci-to-winboard adaptor. So if you use PG as a book provider for an UCI engine, the engine becomes a winboard one and its use becomes difficult in uci-only GUIs. There are some tricks for circumventing this point (see my site) but this is still a bit of a pain.

One thing that is still missing (as Bob Hyatt pointed) is a kind of book-learning function. For what regards myself I am extremely cautious regarding automated book tuning and book learning. But it is true that there is place for development here.

Record ratings on ICC in all timing categories were achieved with polyglot books by the more brilliant exponent of this format, Nick Carlin, who also got excellent results on Playchess and in Freechess events.

This clearly demonstrates how high is the potential of these books

I am pretty comfident that with new more powerful and easier to use tools Polyglot books should become a de facto standard.

Marc

PS Main place to discuss these topics seems to be the Winboard forum.

Alessandro Scotti

Re: Polyglot opening book specification

Post by Alessandro Scotti » Tue Nov 11, 2008 9:59 pm

Thanks a lot Michel, I can now try to add support for PG books in Hamsters! :idea:

Post Reply