Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: question about building a book for my program

Author: Jorge Pichard

Date: 09:31:03 04/08/02

Go up one level in this thread


On April 08, 2002 at 00:07:39, Uri Blass wrote:

>On April 07, 2002 at 22:07:30, David Dory wrote:
>
>>On April 07, 2002 at 10:15:21, Uri Blass wrote:
>>
>>>I think to build a small book for my program when the target is mainly to save
>>>time on the clock and not to learn.
>>>
>>>I do not like the idea of a big book because I do not want my program to blunder
>>>by playing a book move that the program never looked at it and I also never
>>>looked at it.
>>>
>>>I think to have the book as a text file when every line includes the following
>>>information:
>>>1)FEN of a legal position
>>>2)one book move by algebric notation(the target of the book is not to give a
>>>variety of possibility but only to save time so I am not going to give more than
>>>one move)
>>>3)maximal target time (should be dependent of the speed of the computer) when
>>>the book move is going to be played only when the time that my program expect to
>>>waste on the move is smaller than that number and in other cases my program is
>>>going to calculate.
>>>
>>>The strings in the text file should be ordered by lexisographic order from the
>>>smallest to the biggest.
>>>
>>>Every time my program plays a move in a game from a position that is not in book
>>>it should add the position together with the target time that it knows to it's
>>>book in order to prevent it from calculating the same position again.
>>>
>>>Every time the opponent plays a move my program should check if the position is
>>>in book because it is known that it is possible to get the same position in
>>>different ways so the fact that my program is out of book should not prevent it
>>>to search if the position is in book.
>>>
>>>Are there programs with free source code that do it in that way and what is the
>>>simplest program that does it?
>>>
>>>Even if there is only a free program for another thinking game that is not chess
>>>I may be interested to see the relevant source code.
>>>
>>>Uri
>>
>>When you say:
>>
>>>Every time the opponent plays a move my program should check if the position is
>>>in book because it is known that it is possible to get the same position in
>>>different ways so the fact that my program is out of book should not prevent it
>>>to search if the position is in book.
>>
>>I'd suggest saving that info to a file for later checking into, but I'd sure not
>>put it into the opening book. Seems you could dilute a good opening book into a
>>mine field! YMMV
>>
>>I think putting the FEN for every position in your opening book is more than you
>>need. Just store the moves, and your program will show you the position, on
>>screen. Are you thinking of finding the next move by searching through the
>>FEN's,? Haven't seen any like that, but it sounds creative. Curious to see what
>>get's recommended to you.
>>
>>
>>
>>Dave
>
>Yes
>I think to search through the FEN's and if they are ordered by lexisographic
>order then it should not take a long time.
>
>There is a reason that I do not like the idea of remembering the moves.
>The same position can happen in a different order of moves and I want to know
>that I am in book not only after 1.e4 e6 2.d4 d5 but also after 1.d4 e6 2.e4 e5
>and also with opposite sides after 1.d3 e5 2.e3 d5 3.d4

In the same way that transposition works, for instance I have forced 1.h3 as the
first move to CTP 14.9 and depending on black's response it will converted it to
a known opening by transposition.

Pichard.


>If I store moves I need to remember all of these games and if I store FEN's then
>I need to remember only one FEN.
>
>Uri



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.