On Wed, 2007-01-10 at 16:38 -0800, steve uurtamo wrote: > in absolute terms, the time issue doesn't matter until > some piece of code is good enough to beat a dan-level > player on a 19x19 board at *any* physically realistic time > constraint. which hasn't yet been demonstrated. the super > slow motion tournament would be a good way for us to notice > when this happens.
Chrilly correctly points out that one of the programs tested played a lot faster than most of the others. Even if none them can beat a dan level player, what does this have to do with the big time advantage? My programs benefit enormously from extra time and if someone tested it playing at 1 second to another program playing at 1 minute I would probably view the test as unfair. Are you saying this would be fair because neither program can beat a 1 dan player? I don't get it. I like your idea of playing slow motion tournaments. Unfortunately, it's difficult getting humans involved to compare. But perhaps your stone handicap idea gives us a way to make rough comparisons. - Don > (i.e when program X can give program Y 9 stones when > program Y has a 30 minute time limit and program X has > a 24 hour time limit, and they're normally much closer in > strength when they both play at 30 minute time limits). > > as an example, if any program could give gnugo 9 stones > under these circumstances, it might be good evidence > that we're within a factor of 50x in speed of being capable > of beating a 1d player. > > s. > > ----- Original Message ---- > From: Don Dailey <[EMAIL PROTECTED]> > To: computer-go <computer-go@computer-go.org> > Sent: Wednesday, January 10, 2007 7:24:38 PM > Subject: Re: [computer-go] Gnugo vs commercial programs > > Chrilly, > > The computer go guys don't think of performance as a function of time, > only as a kind of absolute, it plays good or it doesn't. > > Us computer chess people are used to thinking of it as a function of > how fast the computer is and how much memory (along with how well the > code is written of course.) > > The UCT programs, assuming they are properly coded and bug free, all > play > perfectly, but what really counts is how well they play given some > time constraint. Again, most of the computer GO people do not think > in > those terms. > > It's very encouraging to me that Hiroshi reported the thinking times, > I think he understands that it is relevant. > > - Don > > > > > > On Wed, 2007-01-10 at 21:01 +0100, Chrilly wrote: > > You must test with Gnu-Go level 16.. This is according to Stefan Mertin by > > far the best mode. But it takes sometimes quite a long time till Gnu-Go > > makes it move. > > In your experiments Gun-Go played very fast. You played fast Blitz and > > Gnu-Go had a big time handicap (besides Handtalk, which plays Ultra-Blitz). > > > > Chrilly > > > > > > ----- Original Message ----- > > From: "Hiroshi Yamashita" <[EMAIL PROTECTED]> > > To: "computer-go" <computer-go@computer-go.org> > > Sent: Wednesday, January 10, 2007 6:10 PM > > Subject: [computer-go] Gnugo vs commercial programs > > > > > > >I tested Gnugo against some commercial programs. > > > > > > Gnugo is 3.7.10. > > > Level is default with --never-resign and --komi 6.5 option. > > > Commercial program is max level. > > > > > > All game is Japanese rule and komi is 6.5. > > > > > > gnugo wins losses winning rate average score > > > GinseiIgo5 (KCC Igo ) 11 55 0.17 -31.3 points > > > Saikouhou3 (Haruka ) 13 44 0.23 -41.4 points > > > TuyoiIgo4 (Go4++ ) 27 54 0.33 -11.4 points > > > ShudanTaikyoku3 (Handtalk) 29 37 0.44 - 4.5 points > > > > > > > > > GinseiIgo5 ... KCC Igo, published in 2004. > > > Saikouhou3 ... Haruka, published in 2002. Latest version. > > > TuyoiIgo4 ... Go4++, published in 2003. Engine is 2002 version. > > > ShudanTaikyoku3 ... Handtalk, published in 1999. > > > > > > These are not latest version except Haruka. > > > All game records are here. > > > http://www.yss-aya.com/gnugo_vs_result.zip > > > > > > Average expended hours. > > > KCC 17m51s Gnugo 3m14s, Opteron248(2.2GHz) > > > Haruka 11m53s Gnugo 4m28s, Opteron248(2.2GHz) + AthlonXP 2100+(1.73GHz) > > > Go4++ 4m59s Gnugo 2m18s, Opteron248(2.2GHz) > > > Handtalk 2m42s Gnugo 5m22s, AthlonXP 2100+(1.73GHz) > > > > > > > > > Appendix. > > > > > > GnuGo 3.5.4 (January, 2004 version) test result. Level is default. > > > > > > gnugo wins losses winning rate average score > > > ValueIgo3 (KCC Igo ) 4 26 0.13 -36.6 points > > > TuyoiIgo4 (Go4++ ) 11 19 0.37 -6.5 points > > > ShudanTaikyoku3 (Handtalk ) 12 18 0.40 -3.8 points > > > AI Igo2004 (ManyFaces) 18 12 0.60 +11.7 points > > > > > > ValueIgo3 ... KCC Igo, published in 2003. same GinseiIgo2PW(2001?) > > > AI Igo2004 ... ManyFaces, published in 2003. engine is 2003? > > > ------------------------------------------- > > > > > > Regards, > > > Hiroshi Yamashita > > > > > > _______________________________________________ > > > computer-go mailing list > > > computer-go@computer-go.org > > > http://www.computer-go.org/mailman/listinfo/computer-go/ > > > > _______________________________________________ > > computer-go mailing list > > computer-go@computer-go.org > > http://www.computer-go.org/mailman/listinfo/computer-go/ > > _______________________________________________ > computer-go mailing list > computer-go@computer-go.org > http://www.computer-go.org/mailman/listinfo/computer-go/ > > > > > > > ____________________________________________________________________________________ > Do you Yahoo!? > Everyone is raving about the all-new Yahoo! Mail beta. > http://new.mail.yahoo.com > _______________________________________________ > computer-go mailing list > computer-go@computer-go.org > http://www.computer-go.org/mailman/listinfo/computer-go/ _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
