On Mon, 2007-04-09 at 09:48 +0900, Darren Cook wrote: > > A few weeks ago I announced that I was doing a long term > > scalability study with computer go on 9x9 boards. > > http://greencheeks.homelinux.org:8015/~drd/public/study.jpg > > Hi Don, > This is very interesting. I envy your persistent energy and enthusiasm. > > If you'd stopped at level 6 or 7 the conclusion could be quite different > (they seem to have started to taper off). Are there are any plans to > continue with a level using even more playouts? I feel like I am > watching a soap opera and now I desperately want to know what happens > next week: heavy has started to tail off, but lite has suddenly pulled > out of its slump. (Though we could be seeing an anomaly due to the > strong versions not playing enough games yet?) > > I hope you will make the games available for download: I'd like to study > the games of the strongest version to see where it is strong and weak. > > Darren
I plotted with gnuplot using smooth bezier smoothing. Also, I don't believe there is enough data at the higher levels to give the most accurate curve. Still, you can get the general idea. I have waited a few weeks to get as much data as I have and I will continue to run this for a few more weeks. I have all the games and results available to anyone who wants to analyze it. I cannot run at higher levels, it would just take too long. However I could add another program to the mix. It would be fun to put a strong version of Mogo to form a 3rd curve. - Don > _______________________________________________ > 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/
