Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: software for creating opening books

Author: Dann Corbit

Date: 17:24:51 12/14/05

Go up one level in this thread


On December 14, 2005 at 19:42:39, Andrew Walker wrote:

>On December 14, 2005 at 19:10:15, Albert Silver wrote:
>
>>On December 14, 2005 at 18:45:13, Joseph Ciarrochi wrote:
>>
>>>Can anybody recommend software for creating opening books in the chessbase gui?
>>
>>I'd recommend the Chessbase GUI.
>>
>>                     Albert
>
>Are there any tutorials in existance on doing this? The chessbase documentation
>could be improved, especially in regards to the meaning of all the different
>colours and symbols you see when browsing or editing a book.

Start by creating a new openings book (.ctg) and then import PGN games (or
whatever) into it.

Here is some stuff from the help file:

Importing games to the openings book


--------------------------------------------------------------------------------

Menu: Edit – Openings book – Import games
This is a very powerful function. It allows you to build giant opening trees
with a minimum of effort. You can, for instance, import all the games of your
main database, or all from a specific variation (to create a highly specialised
book).

The games of the database will be merged to a tree of positions. Crucial for the
size of the resulting tree is the length of the variations which are put into
the tree. It is not very meaningful to put entire games into it, so you should
truncate them at a specific point.



The program allows you to limit the length of the variations in two different
ways.

You can specify that every variation should be “n” moves long.

It is more sensible to limit the variations relative to the ECO classification.
This results in long variations for theoretical main lines (ECO classification
position found late in the game) and short variations for side lines (i.e.,
early deviations).

Obviously, if you choose a shorter length – absolute or ECO-relative – the tree
will be compact. But then you are liable to lose some valuable information. If
you choose a very large value then the tree will be gigantic, and it will
contain a lot of superfluous information. A good practical value for importing
games is ECO-relative 20.

You can also include variations (i.e., analysis that might be included in the
games), but these are not included in the statistical analysis.

The openings book


--------------------------------------------------------------------------------

In the initial phase of a game, the program plays out of an “openings book”.
This can consist of millions of openings positions, and for each of these, the
program has full information on which moves were played, how often and with what
results. The program stores its own experience with openings variations it has
played. It learns from success and defeat.

The openings book is called the “tree” as the branching variations in the
opening resemble the structure of a tree. The files that make up a tree are
pretty large. The program requires about 80 bytes to store a position and all
the information associated with it. You can leave the files on the program CD.
However, in this case, the program will not be able to modify the book as it
plays games (i.e., it cannot learn). Thus, it is advisable to copy them to your
hard disk.

Some things to note:
    The size of the openings tree is limited only by the amount of disk space
available. It is important to note that access to the information is always
extremely fast, even if the files are truly gigantic.

    Existing openings trees can be easily extended, simply by importing games.
This means that whenever a series of high quality games are available, the
program can read them and extend its openings knowledge.

    The tree recognises all transpositions, even those that did not occur in the
games from which the tree was generated. Sometimes you may be able to switch
from one known position to another with a move that has never yet been played.

    The tree can handle reversed colours perfectly. For instance after 1.d3 d5
2.d4, the program is likely to reply 2...c5! and play the entire Queen’s Gambit
as Black.

    The openings tree is not just an instrument to make the program strong, it
is also a very powerful training tool for chess players. For this reason each
move can be marked with chess symbols like “!” and “+=“.

Note: If you are using Windows XP it is very important to keep all your
databases and openings books in the standard directory (My Documents\ChessBase)
suggested by Fritz. Otherwise the operating system will try to back up the giant
files every time you close the program.

Openings book, learning from a database


--------------------------------------------------------------------------------

The program learns from the games it plays. But you can also enhance the book in
other ways.

The tools for this are in the menu Edit – Openings book :

    Import book merges the contents of a different book into the current one.

    Import old book converts a book in the old (FBK) format and imports the
contents.

    Import games merges the games of a selected database into the tree,
including full statistics.

    Learn from database adds the statistics from a selected database without
actually importing the moves into the tree.

    Add priority analysis imports the variations of a specified database and
marks the moves as preferred tournament moves.

The program allocates “weights” to the individual lines of the book. These
weights control the probability with which the program will choose the move in a
game. The value can range between –125 to +125. The values are initially set to
zero. The weights are given in the “Engine” column in the book window.

Learning from a database
With this function the program learns from the experience of other players. It
looks at the games of a database and pretends it has played them itself. On the
basis of the results it changes the weights of individual variations.

Edit – Openings book – Learn from database
displays a dialog box in which you can set a number of learning parameters.



    Wins: Only won games are used, the program increases the weights of white
moves if White won and black moves if Black won.

    Losses: Only losses are considered. If White lost, the white moves are
reduced, and if Black lost, the black moves are reduced.

    White: The program evaluates the games from the point of view of White only.
The weights for white moves are increased if White won, and decreased if White
lost.

    Black: The weights for black moves are increased if Black won, and decreased
if Black lost.

    Player: Only considers the games of one specific player in the database.

These five options can be combined in any way you like. You can also restrict
the import to a certain number of games in the database by selecting "Games" at
the top of the dialog box.



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.