Author: Frank Quisinsky
Date: 09:32:29 12/27/00
Go up one level in this thread
On December 27, 2000 at 11:51:27, Leo Dijksman wrote: >On December 27, 2000 at 11:28:54, Frank Quisinsky wrote: > >>On December 27, 2000 at 11:11:37, Leo Dijksman wrote: >> >>>On December 27, 2000 at 10:34:07, Frank Quisinsky wrote: >>> >>>>On December 27, 2000 at 09:35:52, Leo Dijksman wrote: >>>> >>>>>On December 26, 2000 at 18:06:02, Frank Quisinsky wrote: >>>>> >>>>>>On December 26, 2000 at 16:53:12, Jorge Pichard wrote: >>>>>> >>>>>>>On December 26, 2000 at 12:09:33, Frank Quisinsky wrote: >>>>>>> >>>>>>>>Hi, >>>>>>>> >>>>>>>>new in my CCE tourney are ... >>>>>>>> >>>>>>>>01. CB The Crazy Bishop 0046 (super results in the first 2 rounds). >>>>>>>>02. CB Nimzo 8.0 (little problems in the first 2 rounds). >>>>>>>> 4.0 : 4.0 against Crafty 17.14 >>>>>>>> 5.0 : 3.0 against Comet B.27 >>>>>>>>03. CB Deep Fritz T29c >>>>>>>> At the moment against Fritz 6a (standing so far 1:0 for Fritz 6a, next >>>>>>>> games over night) ! >>>>>>>>04. WB LambChop 10.5 >>>>>>>> 4.0 : 4.0 against the very strong Amy 0.7 ! >>>>>>>>05. WB Insomniac 0.69a >>>>>>>> 6.5 : 1.5 against AnMon 5.09 ! >>>>>>>> >>>>>>>>I think Insomniac is the strongest amateur program but we will see. I believe >>>>>>>>Insomniac plays around ~ 2525 ELO and is stronger than Goliath, SOS and Crafty. >>>>>>>>I think Nimzo Level ! >>>>>>>> >>>>>>>>More information under: >>>>>>>>http://amateurschach.in-trier.de/schach/cbase/cce_main.htm >>>>>>>> >>>>>>>>Best >>>>>>>>Frank >>>>>>> >>>>>>>Why don't you have another Tournament Section C where Gandalf 4.32 G play >>>>>>>against all others programs, I notice that you did not include Gandalf 4.32 Vs >>>>>>>Deep Fritz or Nimzo 8. >>>>>>> >>>>>>>Pichard. >>>>>> >>>>>>I play with Chess-Base native engines in tourney "A" and with WinBoard Engines >>>>>>in tourney "B". I will not make changes in the organisation. Furthermore, it is >>>>>>to many work to play with WB engines in tourney "A" ! I mean my tournament >>>>>>running around one year and this is enough ! >>>>>> >>>>>>Best >>>>>>Frank >>>>> >>>>>Do this mean that you on top of the page shows the ratings of Frits, Gandalf and >>>>>Junior while Gandalf has never played against the other top programs?? >>>>> >>>>>Best wishes, Leo. >>>> >>>>I mean that why I know a lot of results against other top programs. You can >>>>search more information in the CCE beta test forum or in my News Ticker. >>>> >>>>CCE Beta Test forum: >>>>http://f22.parsimony.net/forum41668/ >>>> >>>>News Ticker: >>>>http://amateurschach.in-trier.de/news.htm >>>> >>>>(here a lot of informations in older News Ticker pages). >>>> >>>>News Ticker Overview under: >>>>http://amateurschach.in-trier.de/news_top.htm >>>> >>>>Best >>>>Frank >>> >>>Hi Frank, >>> i mean that the rating that appears on top of the page are from 2 different >>>tournaments (CCE-A and CCE-B), so it looks like Gandalf is the nr.2 but actually >>>it has not played any game against the other top programs that counts for that >>>ratinglist, or do i see it totally wrong? >>>I do not mean that Gandalf is not strong or so (i have Gandalf myself and it is >>>strong!), only i think its not fair to mix up 2 different ratinglists thisway >>>(but thats my opinion offcourse). >>>Sorry for my bad english, Leo. >> >>Hi, >> >>oh I understand you ! >>Yes a little problem. But in tourney A and B plays 5 same progams in WB version >>and CB version. I think end of the tourney is the rating list to 90% perfect. >>When I understand the calculations system right is the good enough when each >>program plays against each programs. Good enough when I have 5 same programs in >>tourney a and b. >> >>When I added games between Gandalf against Junior, Hiarcs and Fritz in my rating >>list (I make the experiment last week) I have the same ELOs. So I think the list >>is now good but at the moment I have not enough games for all programs which >>started in my tourney. >> >>Best >>Frank > >Thanks Frank, > now i understand your system better, how many games do you think that all >Engines must have played each other before a ratinglist is good enough? Not easy to say. I think when I have from all engines 150 games is this for a rating list -+ 20-40 ELO good enough. I mean it is better than to have a list with 400 games for each engines with different hardware, different time contorls or the same programs against the same programs in different versions. For over 1 year I made a 40 moves in 40 minutes rating list with Kai Skibbe and Christian Koch (~ 5.000 games). The differents are not very big after ~ 120 games to 300 or 500 games. The only program with differents over 40 ELO after 200 games was Francesca. I mean one program from all this program in the list. The list and all this games are available on my webpage. >I have now about 60 engines that runs under the chessbase gui, i want to let >them play 16 games matches against each other with a time control of 40/40(2x)+1 >hour, do you think thats enough or should it be 24 game matches? Yes :-) But a lot of test time ! My CCE tourney running around 1 year with 18 programs in tourney A and 18 programs in tourney B. Best Frank
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.