Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Question for Top Programmers ........ You know who you are

Author: Dagh Nielsen

Date: 09:45:10 11/26/05

Go up one level in this thread


On November 26, 2005 at 12:02:23, chandler yergin wrote:

>Also, this way I won't leave anyone out
>;)
>Please understand I am a Player only I don't use Playchess,
>nor do I Program. I use Chessbase & Software to help me improve my game.
>So my 'perspective' is from this viewpoint only. I am not familiar with
>many of your Programming terms, so please cut me some slack.
>You Gentlemen are Master Players & Programmers.
>Question on Opening Books.
>Many play On-Line and do testing of Engine vs Engine etc.
>Is it not the purpose or goal to have the strongest Opening Book possible?
>If so, why do many make & play their own shallow opening books?
>Using Mega99 Database it is easy to Drag & Drop into the Standard Opening
>Book that comes with the Program, Truncating the games 20 moves beyond
>the ECO Classification. Install Opening Key and you have up to date
>Opening Theory. You also have over 56,000 Opening Key ECO Classifications.
>What am I missing here?
>Thanks in advance for your thoughts & comments,
>Chan
>
>
>
>
>--------------------------------------------------------------------------------

I'm not a top programmer, but an engine player on playchess with a customized
book. The short answer why I don't do as you suggest is that it will not give a
very good book. Some problems:

1) GM games are far from flawless.
2) There is more to opening theory than what has been played or published. A
book like yours will play variations that are just bad and that my engine
opponents know how to beat.
3) Engines don't play one of the colors very well in some kind of positions
(positions that GMs can play well). You would want to avoid these variations.
4) Compared to known and published opening theory, there are many, many new and
equally good moves deep in the different variations, moves that your engine
opponents will find and prepare. Hence, you would want to prepare your book more
thoroughly than just "published theory" in whatever variations you play, so that
you are not outprepared by the opponent's homework.

So instead of your suggested approach, I manually tune and expand my book. This
does not give me a shallow book, on the contrary, I often play games where I do
not leave book until 10 or more moves out of "known" opening theory.

In a sense, engine book cookers are opening researchers. That research can
easily improve on known theory, and hence the benefit of doing it this way.

Regards,
Dagh Nielsen



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.