Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Without a book and with their own book, this is a great idea !

Author: Drexel,Michael

Date: 16:59:28 12/29/02

Go up one level in this thread


On December 29, 2002 at 19:30:00, Uri Blass wrote:

>On December 29, 2002 at 19:04:47, Drexel,Michael wrote:
>
>>On December 29, 2002 at 17:51:06, Jorge Pichard wrote:
>>
>>>On December 29, 2002 at 17:48:59, Jorge Pichard wrote:
>>>
>>>>On December 29, 2002 at 17:13:19, Joshua Haglund wrote:
>>>>
>>>>>On December 29, 2002 at 14:07:14, Lieven Clarisse wrote:
>>>>>
>>>>>>I disagree, testing without an opening book is a good test for chess engines!
>>>>>>
>>>>>>lieven.
>>>>>
>>>>>Disagree...
>>>>>
>>>>>I believe there should be a rating list for all the programs without a book; 100
>>>>>rounds bullet, blitz, and standard time controls.
>>>>>
>>>>>Then we'll find the best engine I believe. If it can find the best move in an
>>>>>unorthadox opening A00, B00, C00, D00 E00... it should find the best moves for a
>>>>>program with a book. I think chess tiger 15 is the new king of the mountain.
>>>>>
>>>>>For now... ;)
>>>>>
>>>>>Joshua
>>>>
>>>>
>>>>I Believe that the SSDF should be testing programs by using the first 20 games
>>>>without a book and the remaining 20 with their own book, this will give us an
>>>>idea of how good some programs really are.
>>>>
>>>>Pichard
>>
>>I totally disagree.
>>look at the games which Cristophe Drieu posted. if you would play completly
>>without any book the engines would play always the same (bad) variations.
>>no A00,B00,C00,D00,E00 just B00 or B01.
>>it would be interesting what the programers Cristophe or Stefan have to say
>>about all this. if they say I havent done anything to improve the play of my
>>engine within the first 10 moves for many years now (which is what I guess),
>>then it is meaningless to let them play without opening books.
>>
>>a good idea would obviously be:
>>to create an opening book which has a wide range of variations from A00-E99 in
>>it. no variations should last longer then till move 9 or 10 and should not lead
>>to great disadvantages for both sides. no dubious gambit variations for example.
>>
>>the SSDF should use such a book for all engines.
>>this will never happen of course.
>>
>>Michael
>
>I do not think that the ssdf should do it.
>
>If the target is to find the strength of the engines without books then it is
>possible to start from opening like 1.a3 a6 or 1.a4 a5 and to continue in that
>way

It makes not much sense to play 1.a3 a6 first. there would be not much
difference I guess. if SSDF would play without opening book (or with 1.a3 a6
opening book)then programers would probably start to implement opening books in
the engines (if that is allowed).
the point is: if SSDF would decide not to allow the use of opening books any
more, then they have to give programers time to improve play during
first stage of a game.


 >but it is not the job of the ssdf but of the people who are interested in it.
>
>I also think that the games that were played were not 120/40 and I expect the
>engines to play better in serious games even without book.
>
>I doubt if tiger15 may play 1.e4 d5 2.ex5 Qxd5 3.Nc3 Qe6+ in 120/40
>
why not. Look at this:

Analysis by Shredder 7:

1. ² (0.69): 3...De6+ 4.Le2 Dg6 5.Sf3 Dxg2 6.Tg1 Dh3 7.Tg3 Dd7 8.Se5 Dd4 9.Sb5
Db6
2. ± (0.74): 3...Dd8 4.Sf3 Sf6 5.Lc4 e6 6.0-0 Ld6 7.d3 0-0 8.Lg5
3. ± (0.76): 3...Da5 4.Sf3 Sf6 5.Lc4 e6 6.0-0 Ld6 7.Sb5 0-0 8.Sxd6 cxd6 9.d3 Sc6
10.Le3
4. ± (0.79): 3...Dd6 4.Lc4 Sc6 5.d3 Sf6 6.Sge2 Lg4 7.0-0 e6 8.Sb5 Dd8 9.Lf4
5. ± (0.87): 3...Df5 4.Sf3 Sc6 5.Ld3 De6+ 6.Le2 Sf6 7.0-0 Dd6 8.d3 a6 9.Le3 Lf5
10.Sh4 Ld7
6. ± (1.00): 3...De5+ 4.Le2 Sf6 5.Sf3 Dd6 6.0-0 a6 7.d3 Sc6 8.Le3 Lf5 9.Sh4 Ld7
10.Te1
7. ± (1.08): 3...Dd7 4.Lc4 Sc6 5.Sf3 Sf6 6.d4 e6 7.Sb5
8. ± (1.17): 3...Dc5 4.d4 Db4 5.Sf3 Sf6 6.a3 Dd6 7.Lc4 Le6 8.Sb5 Dd7 9.Lxe6
Dxe6+ 10.Se5
9. ± (1.17): 3...Dd4 4.Sf3 Db4 5.d4 Sf6 6.a3 Dd6 7.Lc4 Le6 8.Sb5 Dd7 9.Lxe6
10. ± (1.21): 3...Dg5 4.Sf3 Dh5 5.d4 Lg4 6.Le2 Sc6 7.h3 0-0-0 8.0-0
11. +- (6.45): 3...Lg4 4.Sxd5 Lxd1 5.Sxc7+ Kd8 6.Sxa8 Lxc2 7.Lc4 e6 8.d3 Sc6
9.Lf4 Sb4 10.Tc1 Sxd3+ 11.Lxd3

depth 12/30 0:02:00








>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.