On Sat, 2008-11-01 at 21:48 -0400, Don Dailey wrote: > I also tried large bonuses and so far 0.20 works better than 0.10 by > about 20 ELO after 2,500 games. I tried 0.40 and it was significantly > weaker, so this can definitely be overdone. The final result is 131 > ELO for the 0.20 value. I will probably try 0.15 just to see what > happens. I prefer to use lower values if it is all the same and I know > that 0.10 is weaker than 0.20. > > I can't try but so many things, so I'm optimizing and testing for 2000 > playouts, which is a reasonable level for a PDA device. It's plays > very quickly of course on a PC. > > On CGOS, looking at the bayes page it is rated 1475 based on 162 games, > which compares well to ControlBoy (which is really anchorman) which has > attained a rating of 1538 based on 74,000 games. But ControlBoy does > 5000 playouts and this does only 2000. I guess I should put that one > up too at 5000. > > - Don
I'm curious, how do you do your automated testing? I tried using gogui-twogtp -auto with jrefbot and drefbot. When I look at the resulting .dat files, there's no indication of which side won each game. I know final_status_list is implemented. Is anything else required to make that work? Here's a sample summary file: #GAMES ERR DUP USED RES_B ERR_B WIN_B ERRW_B UNKN_B RES_W ERR_W WIN_W ERRW_W UNKN_W RES_R ERR_R WIN_R ERRW_RUNKN_R 200 0 0 200 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 j _______________________________________________ computer-go mailing list [email protected] http://www.computer-go.org/mailman/listinfo/computer-go/
