Author: Kurt Utzinger
Date: 02:59:43 06/14/03
Go up one level in this thread
On June 14, 2003 at 05:45:59, m.d.hurd wrote:
>On June 14, 2003 at 05:24:34, George Tsavdaris wrote:
>
>>On June 14, 2003 at 04:59:09, Kurt Utzinger wrote:
>>
>>>English:
>>>We are going to start a long tournament at 120'/60 under the conditions
>>>mentioned below. As you will see, we want to merely test the strength of the
>>>engines. Therefore, the remis.ctg will be used instead of the own books. And
>>>book learning is of course deactivated. Furthermore, the programs will have no
>>>access to the EGTB so that we can see the proper endgame knowledge of the
>>>engines. Hints about the outcome of this tournament are welcome. We are
>>>publishing first results after 10 rounds.
>>>
>>I think not letting programs have access to endgame tablebases is clearly
>>wrong! Because tablebases have been created for using them with engines,
>>to improve their endgame play. And furthermore some programmers may not include
>>code for some positions, because with the help of EGTB the engine will play
>>properly.
>>I honestly don't understand why you decided that. It's like not letting
>>in a human tournament, the people to have some advice from their coach or
>>not letting them to read the new book from Kasparov about endgames, because
>>you want to see their real strength.
>>Because if Shredder 7.04 for example is able to play the endgame with 2710 ELO
>>and Chess Tiger with 2690 with the help of EGTB and without EGTB Shredder
>>would play it with 2680 and CT with 2675 then we have an unfair competition
>>if we wont let engines use EGTB.
>
>I agree with you, I would have thought that most people use their engines with
>EGTB and therefore suggest that Kurt re-considers. However ce la vie.
>
>regards
>
>mike
Hi Mike
The fact that most people use their engines with EGTB
is well known. But that's just the problem: in this way
nobody can estimate the real endgame knowledge/strength
of an engine. And this is what we are interesting to learn
more about it.
http://f11.parsimony.net/forum16635/messages/49266.htm
"The surprising conclusion of these tests is that I could
not observe an increase of playing strength in Yace Paderborn
due to bitbases and even tablebases."
Kurt
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.