Author: Sune Fischer
Date: 03:34:16 06/09/03
Go up one level in this thread
On June 09, 2003 at 00:50:29, Mike S. wrote: >Recently it has been critized that SSDF doesn't test King 3.23 (yet), it has >been discussed, explained why not, etc.etc. > >(Purists of 40/2h shouldn't continue to read :-)) > >I think nowadays, 10m+10s can provide a perfectly valid comparison too. On this >rating list created by Mr. Klaus Wlotzka for CSS Magazine, > >http://www.computerschach.de/rangliste/ This is fine, and the best way to quickly make a rating list I think. IMO it will be a good approximation for longer games too, but it isn't and will never be a rating list for long games. >(For the details of the test conditions, click "Testbedingungen" in the menu.) > >You can see King 3.23 has achieved rank #4 just behind Fritz, Shredder and >Tiger. King is running with Sel. 12 there (and 128 MB hash, rest default), so I >guess it could do a few points better with additional setting changes, i.e. >higher king safety. OTOH, I doubt that King could jump to rank #3 by that, which >is 17 Elo away. 17 Elo is a big distance nowadays in top computer chess. > I think the engine should always be tested (when officially trying to measure the programs strength) using the default values, or whatever values suggested by the author (should be the one and same probably). The reason is that finding optimal values is not trivial at all, in fact it may be one of the biggest challenges in computer chess. If the author gives the user the option of changing parameters, then that is fine and good fun, but if the users are able to improve on the default values, then that means the programmer has not done his job well enough and as such the engine doesn't deserve to play with these improved settings in getting an official score. Sounds like a strange opinion perhaps, but think of this: Suppose all engines had all parameters public for every one to edit. Now all these engines would have to play under equal conditions in getting a rating, right? Well that means if one engine is improved by the users, then all the other engines should have the same chance, or the conditions are simply not equal. I don't believe that any chess program has 100% optimal parameters, so this would practicly be possible to do for any engine. It isn't right to hand over the job of optimizing to the users, this is like 50% of chess programming, so it would be cheating IMO. -S.
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.