Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Which program has the best Opening Book ?

Author: Peter Berger

Date: 10:31:15 12/12/00

Go up one level in this thread


On December 12, 2000 at 13:12:28, Jorge Pichard wrote:

>On December 12, 2000 at 13:01:15, Amicitia Stone wrote:
>
>>On December 12, 2000 at 12:40:43, Jorge Pichard wrote:
>>
>>>Is it Shredder 5, Nimzo 8 or Chess Tiger ?
>>>
>>>Pichard.
>>
>>
>>Hmm...
>>The question you just asked, is sort of like saying which program is best. There
>>is no answer to such a vague question.
>
>I don't agree with that, I simply asked which  program has the best Book, we all
>know that Chess Tiger 13.0 and Shredder 5 are better program than Nimzo 8.

I seriously doubt that it is true that we all know that. Nimzo 8 simply seems to
have less fans ; the Nimzo results posted look very promising at tournament time
controls.

>I just finished a small test in which Fritz 6a using Nimzo 8.ctg or Opening book
>beat Chess Tiger 13.0  8 games out of 15 at G/30. But in another test using
>Fritz 6a original Book Chess Tiger 13.0 won by 8.5 out of 15 at G/30. Therefore,
>the Opening Book is irrevelant to the strength of the program.

This might have been a bad test for the question discussed as both books ( Fritz
and Nimzo ) were prepared by the same author : Alex Kure.

A good test might be to use the general.ctg of Fritz 5.32, Fritz 5 or Hiarcs.ctg
if you have one of those which were (or are said to be) automatically generated
books made out of grandmaster games . So then you could pair Fritz with one of
these books against Tiger and then compair to the result with the Fritz6-book or
the nimzo.ctg . For test purposes it would be best if you could turn off book
learning for the competitors but as I don't know how to do that with the Tiger
book it might be best to just leave the learning as it is ( which probably will
show the effect a little less clearly ) .

You would have to play a huge number of games to get a reliable result if the
ELO difference were about 30-50 points it seems you would have to play at least
about 100 games with each constellation to see this trend .

I don't want to convince you to do these time-consuming tests I just think I'd
like to see the results if someone did it .

Another good comparison would be Crafty with its own book against Crafty with
one of the Kure books btw ; I suspect the trend would be much clearer here.

greetings.

pete



>
>Pichard.
>>
>
>>Nimzo is known for its strong book, however all the books you mentioned are made
>>highly skilled individuals. All are very good.
>>(and very big) :)



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.