Author: Uri Blass
Date: 08:36:18 10/31/03
Go up one level in this thread
On October 31, 2003 at 11:09:40, Christophe Theron wrote: >On October 31, 2003 at 10:54:04, Uri Blass wrote: > >>On October 31, 2003 at 10:40:04, Gian-Carlo Pascutto wrote: >> >>>On October 31, 2003 at 02:40:26, Johan de Koning wrote: >>> >>>>The SSDF is allowed to test anything they want. >>>>That's the bottom line, even though it is the top line in this case. >>> >>>In that case, I'd be nice to have Deep Sjeng 1.5 tested under Shredder 7 >>>Classic (UCI) with the Shredder7.bkt and identical book and EGTB settings >>>as Shredder 7.04 is using. >>> >>>I'm sure it would do well. The result would be pure garbage, but apparently >>>that wouldn't stop you. >>> >>>-- >>>GCP >> >>I do not think that the result is going to be pure garbage and if you ask them >>to do it they should do it. >> >>They should only not write Deep Sjeng in the rating list because it may be >>misleading but if they write >>Team B 2700 and write notes to the list >> >>Team B consists Deep Sjeng and Shredder7.04's book then I fail to see the >>problem. >> >>Uri >>Uri > > > >You fail to see the problem because you do not understand the spirit of the SSDF >list. > >The SSDF list is not a human rating list. > >The SSDF list is not an advanced players (human+computer) rating list. > >The SSDF list is not a "hybrid software" rating list (program X + program Y). > >The SSDF guys are not trying to find what could be the strongest combination of >hardware and software. For example they would not use hardware A and program A >to play the opening, hardware B and program B to play the middlegame, and >hardware C running program C in the endgame. It does not make sense to use different hardware for different stage of the games because nobody is using chess programs in that way. It make sense to use different software for different stages of the game. > >The problem here is that they are starting to do that, and they did not even >notice that they were opening a very dangerous door. > >They are using opening book A with book learner B and engine C. A and B from one >company, C from another company. > >The SSDF has always tried to rate PRODUCTS (free or not) relative to each other. What about rating of a product that you can get by buying programs from 2 companies? >The question that we want the SSDF to answer is: "how good is ChessMaster 9000 >relative to the products they have already tested?". > >The question is not "What rating could be achieved by a cyborg composed of parts >taken from ChessBase and other parts taken from ChessMaster 9000?". You can think of Cyborg composed of parts also as a product that you can buy. People are interested in the question what is the best software that they can buy. Not testing a Cyborg simply means preventing information from them. 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.