Author: Terry McCracken
Date: 21:13:50 03/07/02
Go up one level in this thread
On March 07, 2002 at 06:01:04, Chris Carson wrote: >On March 06, 2002 at 23:31:14, Terry McCracken wrote: > >>On March 06, 2002 at 13:37:44, Chris Carson wrote: >> >>>On March 06, 2002 at 11:02:21, Terry McCracken wrote: >>> >>>>On March 06, 2002 at 07:16:27, Chris Carson wrote: >>>> >>>>>The SSDF has the proper ELO calibration for programs vs GM's. >>>>>http://w1.859.telia.com/~u85924109/ssdf/ >>>>> >>>>>See Tony's page for Human vs Machine results (look at calibration games): >>>>>http://home.interact.se/~w100107/welcome.htm >>>>> >>>>>and (my list on Tony's Page): >>>>> >>>>>http://home.interact.se/~w100107/manmachine.htm >>>> >>>> >>>>Chris been here and it's all good and nice, but until the F.I.D.E. allows the >>>>same rights for machines as it does humans, you'll not see enough games to >>>>prove that today's programmes are GM's or GM strength, let alone 2700+! >>>> >>>>Terry >>> >>>All data points to this. You should ask the people who think it is lower to >>>provide data to back up the claims (tournament/match conditions in public the >>>same as these tournaments and matches). >>> >>>Just saying it is lower does not make it so. >> >>Which to that I'll add, "Just saying it's higher does not make it so!:> >> >>Well Chris, I guess I'm a Super GM! I beat RC4 two games in a row! >> >>And I didn't use "Anti-Computer Play" either!:o) But I did force it down to >>a winning endgame. >> >>I nailed Deep Fritz 6 without anti-computer play as well in 27 moves, Checkmate! >> >>Sorry I haven't the game scores, but then again people may think I fabricated >>the games!;-) >> >>Terry > >SSDF Calibration Human vs Computer = 318 games >Carson list Human vs Computer = 211 games > >Total = 529 games > >Pleenty of games, plenty of GM norms. Statistically significant and a very >powerful study. > >Play matches and tournaments in public against the top programs on top hardware >as these games were. > >If you will notice, the programs can be beaten, but not consistently and they >are consistently winning matches against the top GM's. The programs win a lot >more than they loose. > >If yhou win a game, enjoy it, the program will win more though. Maybe? I'm not in form anymore and actually out of practise! Ten years ago it was pretty easy to beat standalones and early PC software, but those days are gone:( Although I'm hesitant to say they are at the level of GM's let alone Super GM's it's really just a matter of time before they will beat us all and all the time! At least 99.999% of all chessplayers, all the time:) It will be rendered academic, probably by 2010. This is a conservative estimate. Terry
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.