Author: José Carlos
Date: 02:43:13 10/25/02
Go up one level in this thread
On October 25, 2002 at 04:34:42, Uri Blass wrote: >On October 25, 2002 at 02:53:19, José Carlos wrote: > >>On October 25, 2002 at 00:10:20, Aloisio Ponti Lopes wrote: >> >>>SSDF engines list: >>> >>>1 Fritz >>>2 Shredder >>>3 Tiger >>>4 Tiger >>>5 Shredder >>>6 Fritz >>>7 Junior >>>8 Rebel >>>9 Hiarcs >>>10 Shredder >>>11 Gandalf >>>12 Fritz >>>13 Gandalf >>>14 Tiger >>>15 Gandalf >>>16 Junior >>>17 Tiger >>>18 Shredder >>>19 Fritz >>>20 Fritz >>>21 Crafty... >>> >>>and the list repeats itself until appears 29. Nimzo !; 38 SOS ! and 40. >>>Chessmaster. >>> >>>In my oppinion the list should have this format: >>> >>>1. Fritz >>>2. Shredder >>>3. Tiger >>>4. Junior >>>5. Rebel >>>6. Hiarcs >>>7. Gandalf >>>8. Crafty >>>9. Nimzo >>>10. SOS >>>11. Chessmaster, etc... >>> >>>... but I doubt that if matches like Fritz-Fritz or Tiger-Tiger, etc were >>>removed from the list the ratings and standings would not change. >>> >>>Please see my bullet/blitz rating at >>>http://www.geocities.com/Colosseum/Track/7733/rating.txt >>> >>>... and for you guys that love soccer I'll send the "FIFA-SSDF" list: >>> >>>1. Brazil A >>>2. Germany A >>>3. Argentina A >>>4. Brazil B >>>5. Italy A >>>6. Germany B >>>7. Argentina B >>>8. Brazil C >>>9. Italy B >>>10. Germany C >>>11. Netherlands A >>>12. England A >>>13. Germany D >>>14. Brazil D >>>15. Argentina C >>>16. England B >>>17. Spain A >>>18. Spain B >>>19. Portugal A >>>20. Croacia A >>>21. Korea (only one version!) >>>22. Japan A >>>23. Brazil E >>> >>>... and the list repeats itself untill we find 44. Paraguay; 59. Uruguay; 75. >>>Cameroon... and so on... >>>Of course we won't see Argentina A x Argentina B matchs at FIFA's list... >>> >>>A. Ponti >> >> Nonsense comparison. >> A football team changes players everyday, while a chess program is a single >>and independent entity. Fritz 4 and Fritz 7 are two different programs; >>computing their rating together is like computing two tenis players rating >>together: nonsense. >> >> José C. > >Tennis players can get better and the same for a team of program+programmer. By that logic you can also say that a team of tennis players can get better, and then not give rating for a single player, but for the whole team. >I see nothing wrong if people give rating only for one Fritz,one shredder,... >They can play a tournament when they always use the latest Fritz,Latest >Genius,.... If you want to consider a program's evolution then you can't just compute the first version's game for the rating of the last versions, but you need to compute different ratings for different versions, like in human chess: every 6 months we have a new list. In human chess, last rating is used to calculate variation. But old games (from old "versions" of the human) are not computed anymore. So if you want an "evolutionary" rating list for programs, that's ok, but then you must throw away old games once they belong to an old rating, like in human chess. However, considering that every program is different from each other (this is not just an opinion, it's a fact), I prefer (opinion) having all of them in a permanent list of games, which gives a better comparison between them, and additionally shows an easy measure of evolution. José C. >My opinion is that in that case they should announce about tournaments before >they play them so programmers who are interested can send an update for the >tournament. > >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.