Not logged inRybka Chess Community Forum
Up Topic The Rybka Lounge / Computer Chess / How can we know which of the opening books is the best ?
- - By bil-124 (**) Date 2009-07-09 18:59
Hi all
actuality im little confused about chess oeining books , thay are a lot and every where so how i know this is good and this bad?
and is there any relation between  x engine and x opening book ? ( i mean if i use x opening book which belong for example for rybka and use it with fritz or shredder is it will be the same strength and achieve the same goal ? or some books are made especially to work for specific engine ?
what is the way to test these books ? is there any?

thanks
bil124
Parent - - By Permanent Brain (*****) Date 2009-07-09 19:31 Edited 2009-07-09 19:34
Opening book tests and -tournaments are being done all the time. You cannot possibly have missed all these threads? The idea is simple: The books are the only difference between the opponents, e.g. same engine, (on) the same computer, etc. Then, many matches are being played with very many games between them, and then we see which book scores highest.

Which the best book might be, is also a matter of definition. But for online engine sport, what people are most interested in is, which book scores best combined with Rybka 3. Because Playchess machine room competition is at least 80% R3 vs. R3, probably more. It's like a running gag. Which is better, Rybka or Rybka? :lol:

Another definition could be, what is the most comprehensive and up-to-date master's openings reference, without having engine competitions in mind at all.

I think that relations between specific engines and related books are generally overestimated a lot. Still, an engine specific book (designed for that engine and tested with that engine) can include some "tuning", which means for example, normal continuations are deactivated if the book author finds that the engine will "notorously" handle the resulting positions badly, after the book. A simplified example would be: A gambit where you should make use of a dynamical attack potential, but instead the engine tries to win the pawn back, only. So this gambit is not suitable for this engine, but maybe for another. - But I think these are marginal details in an ocean of possible book moves and variations.

(Except books are (anti-)optimized and play such narrow repertoires that 1 from 2 games each, is a Sicilian Najdorf. I refuse to understand why that is being done, even if it would provide best ratings. It's comparable to the difference between a Nascar oval track and a good Formula One racing track, somehow.)

Also, there were large book tests (many games!) in the past, first with engine X + book X versus engine Y + book Y, and then X+Y vs. Y+X, but also X+Y vs. Y+Y etc.; all reasonable variants. Either there was no significant impact of these "wrong" assignements, or it even happened that an engine scored better with another book than with it's own. So, I think this whole engine specific tuning is only interesting if you like to explore specific openings and variations (and why which move was edited, etc.) but not something which has a significant impact in result statistics.
Parent - - By The Truth (**) [de] Date 2009-07-09 20:16

> Which the best book might be, is also a matter of definition. But for online engine sport, what people are most interested in is, which book scores best combined with Rybka 3. Because Playchess machine room competition is at least 80% R3 vs. R3, probably more. It's like a running gag. Which is better, Rybka or Rybka? :lol:


We are testing books not engines and simply want to have the best entity there is. If you want to go on a weaker engine go ahead, and prepare to lose tons of games. A strong line in the book is a strong line no matter what you do and will win if played right, it is not an issue of engines. This issue is so 2003 and a typical weakling's excuse.

>Another definition could be, what is the most comprehensive and up-to-date master's openings reference, without having engine competitions in mind at all.


In your dreams because there is no such thing.

>(Except books are (anti-)optimized and play such narrow repertoires that 1 from 2 games each, is a Sicilian Najdorf. I refuse to understand why that is being done, even if it would provide best ratings. It's comparable to the difference between a Nascar oval track and a good Formula One racing track, somehow.)


If you are a good book maker and have good a book you can go against any repertoire, enough said.
Parent - - By Permanent Brain (*****) Date 2009-07-09 20:45

> In your dreams because there is no such thing.


You must have misunderstood something. It's possible and not difficult to do. Of course, I mean only a reference for known opening moves and repertoires as they were played in public games, not something hidden in the "secret labs" of the GMs, still.

Maybe not entirely up to date, but up to May 2009, I recently found very good public domain chess game collections, for download. It is easy to make a CTG book from them. Since this game collection is being updated, I could even keep that reference book better up-to-date by repeating the process. But I don't do it because I can live without the opening novelties from the most recent months. :lol:

But it is possible. Based on even bigger data masses - but maybe with a little delay due to production time - ChessBase releases such reference book trees annually. I have no clear overview at the moment but I think there are several different products of this kind (Power Books, etc.).

> This issue is so 2003 and a typical weakling's excuse.


I don't need any excuse for not being interested in this circus.

As for the books, from recent results I have the impression that every 12 hours, another one becomes the "best" book.

P.S. Is book making being done ONLY for the Playchess machine room, nowadays?! Jeroen Noomen's opinion on this:

"For whom is the Rybka 3 opening book interesting?

Although the Rybka 3 book is especially designed for the Rybka 3 engine, it might also be used as a theoretical guide on topical lines by:

- Correspondence chess players
- Tournament chess players
- All other chess players interested in the latest theory on the most popular opening lines of today
- ChessBase users
- Chess enthousiasts interested in playing engine-engine games
- Fritz, Junior, Hiarcs or other ChessBase engine users, who want to update their opening books"
Parent - - By The Truth (**) [de] Date 2009-07-09 21:15

>You must have misunderstood something. It's possible and not difficult to do. Of course, I mean only a reference for known opening moves and repertoires as they were played in public games, not something hidden in the "secret labs" of the GMs, still.


The ones hidden in the 'secret labs' of the GMs are their combined analysis with rybka. I am very sure GMs like Anand or Kramnik etc. have rybka as their analysis partners. And this is what most serious chess and computer chess player wants and not just some reference of human games which contain lots of tactical mistakes.

>But it is possible. Based on even bigger data masses - but maybe with a little delay due to production time - ChessBase releases such reference book trees annually. I have no clear overview at the moment but I think there are several different products of this kind (Power Books, etc.).


Power Books?! Oh please, they are next to useless. Not even a good product for human and computer chess.

>I don't need any excuse for not being interested in this circus.


Yes, you don't need one because it is obvious you know nothing about computer chess.

>P.S. Is book making being done ONLY for the Playchess machine room, nowadays?! Jeroen Noomen's opinion on this:


Jeroen Noomen's book would have not existed if it was not for Playchess engine room. Please learn to read the credits: http://www.rybkachess.com/index.php?auswahl=Rybka+3+book

"The Playchess engine room, what you guys do is amazing and without you this opening book would never exist!"       -Jeroen Noomen
Parent - - By Uly (Gold) [mx] Date 2009-07-09 21:31

> Yes, you don't need one because it is obvious you know nothing about computer chess.


Wow, I'll beat Permanent Brain to saying it:

<plonk!!!>
Parent - - By Permanent Brain (*****) Date 2009-07-09 23:20
The astonishing thing is, that instead of simply disagreeing with me, which is not a big deal, some people need to claim that I "know nothing about computer chess." :grin: Yes, sure. I know nothing about it, I am not a serious chess player, I make weakling's excuses for not making a super book for Playchess (which is of course an absolute obligation for everyone). I write totally senseless postings all the time, using a random text generator, since 1999, and magazine articles on the topic throughout many years before.

It was all a giant Turing test. :lol:
Parent - - By Jonas (****) [at] Date 2009-07-10 02:16
i already heard that too...
as long as im not a coder i must admit that i dont know very much about it.
there should be a 'how much do you know about computer chess' quiz.
Parent - - By Permanent Brain (*****) Date 2009-07-10 03:33
Good idea. But if online, the challenge would be to find questions which Google cannot easily answer. Also, if someone has access to this DVD:

http://tinyurl.com/maye8a

then he has almost all of my pre-2000 computer chess knowledge available, too.
Parent - By bil-124 (**) Date 2009-07-10 11:04
thanks dear
Parent - By bil-124 (**) Date 2009-07-10 11:05
Yes, you don't need one because it is obvious you know nothing about computer chess

what about you ?? what you know ???
Parent - By bil-124 (**) Date 2009-07-10 11:05
Yes, you don't need one because it is obvious you know nothing about computer chess

what about you? what you know ?
Up Topic The Rybka Lounge / Computer Chess / How can we know which of the opening books is the best ?

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill