Author: David Dahlem
Date: 11:21:02 08/08/04
Go up one level in this thread
On August 07, 2004 at 19:37:30, Graham Banks wrote: >On August 07, 2004 at 19:10:02, David Dahlem wrote: > >>On August 07, 2004 at 18:07:09, Graham Banks wrote: >> >>>On August 07, 2004 at 17:30:39, David Dahlem wrote: >>> >>>>On August 07, 2004 at 17:01:38, Graham Banks wrote: >>>> >>>>>Athlon XP2000 >>>>>256mb hash each (or closest allowed by each program) >>>>>3,4,5 piece tablebases >>>>>All using Fritz Powerbook tournament book (maximum variety, minimum 5 games for >>>>>a move to be played, no learning) >>>>>40 moves in 40 minutes repeating. >>>>>Ponder off >>>>>Tournament run under Deep Fritz 8 GUI >>>>> >>>>> >>>>> >>>>>STANDINGS (after 38 of 76 rounds) >>>>> >>>>>29.5 - CM9000 Judge II >>>>>24.0 - List 512 >>>>>23.5 - Gandalf 5.1 >>>>>23.0 - SOS 4 >>>>>22.5 - Thinker 4.6c >>>>>22.5 - Rebel 12 >>>>>22.0 - Pharaon 2.62 >>>>>21.0 - CM 10th Edition (default) >>>>>21.0 - Delfi 4.5 >>>>>18.0 - Anaconda 1.6.2 >>>>>18.0 - Deep Sjeng 1.6 >>>>>18.0 - Tao 5.6 >>>>>17.0 - Gothmog 1.0b10 >>>>>16.5 - El Chinito 3.25 >>>>>15.5 - Yace 0.99.87 >>>>>15.0 - Crafty 19.15 >>>>>14.5 - Little Goliath Revival >>>>>13.5 - Fruit 1.5 >>>>>13.5 - Slow Chess 2.93a >>>>>11.5 - Comet B68 >>>>> >>>>>The top three programs apart from CM9000 Judge II and CM 10th Edition will play >>>>>in the final stage of Super Tournament III at 40 moves in 120 minutes repeating. >>>>> >>>>>The games and crosstable should be up on Kurt’s website sometime soon. . >>>>>http://www.utzingerk.com/banks/st_03_qual.htm >>>>> >>>>>I will now be replacing Rebel 12 with Pro Deo 1.0 and rerunning the first 38 >>>>>games before proceeding into the second half of the tournament as I like to try >>>>>and keep all current program versions in my tournament. >>>>> >>>>>Comments welcome. >>>>>Graham. >>>> >>>>Hi Graham >>>> >>>>I really can't understand the poor showing of CM 10th Edition (default). For me, >>>>it plays much stronger than this. :-) >>>> >>>>Regards >>>>Dave >>> >>> >>>Hi Dave, >>>I don't think that CM 10th Edition will require as much tweaking as CM9000 in >>>order to get stronger settings. I'd obviously put the selectivity up to 12 and >>>I'd change the material values slightly. Apart from these, I can't see too many >>>weak areas of play - maybe pawn weaknesses might need altering upwards? >>>I'm very interested in the results you've had with CM 10th Edition (default) if >>>you'd like to share them. >>>Graham. >> >>Ok, here's a quick rundown according to EloStat. >> >> Program Elo + - Games Score Av.Op. Draws >> >> 1 Shredder 8 : 2486 92 87 43 64.0% 2386 34.9% >> 2 Shredder 7 : 2409 199 106 15 53.3% 2386 53.3% >> 3 TheKing 3.33 : 2384 55 74 79 44.9% 2419 39.2% >> 4 TheKing 3.23 : 2365 118 191 16 46.9% 2386 43.8% >> 5 Ruffian Leiden : 2005 382 218 5 10.0% 2386 20.0% >> >>Regards >>Dave > > >What time control? And did you alter selectivity or is everything apart from >hash at default settings? >Graham. Hi Graham What Selective Search setting are you using for TheKing 3.33? I don't really have to set the parameters because of my setup. This is the line i use in my wb2uci.eng file: InitString=cm_parm default\ncm_parm opk=xxxxxx\n This automatically uses the proper default settings. According to a post from Johan de Koning, the default SS setting should be 10, and not 9 as in TheKing 3.23. Here is what he said about the changes (i saved this to a text file): TheKing 3.33 changes ... Slightly increased mobility weight. Much increased R against K attack. Some minor changes for more accurate positional play. Selectivity scheme changed quite a bit (now default SS=10, probably 12...14 will be strong in comp-competition) Time usage more aggressive. Version number 0.10 higher. :-) ... Johan This info should be very useful for anyone creating custom personalities for TheKing 3.33. :-) Regards Dave
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.