[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [gnugo-devel] First Monte Carlo code ready for inclusion

From: Ben Lambrechts
Subject: Re: [gnugo-devel] First Monte Carlo code ready for inclusion
Date: Mon, 17 Dec 2007 10:49:55 +0100
User-agent: Thunderbird (Windows/20071031)

> So the correct commands are: (after I rebuild gnugo with modified board.h)
> tcl.exe cgos3.tcl GnuCvs-1-MC ***** "gnugoMC.exe --mode gtp --cache-size
> 512 --chinese-rules _--max-level 1_ --capture-all-dead
> _--positional-superko_ --monte-carlo _--mc-games-per-level TOTEST_"
> and
> tcl.exe cgos3.tcl GnuCvs-10-MC ***** "gnugoMC.exe --mode gtp
> --cache-size 512 --chinese-rules --level 10 --capture-all-dead
> _--positional-superko_ --monte-carlo _--mc-games-per-level TOTEST_"

For GnuCvs-1-MC I would suggest both "--level 1" and "--max-level 1".
After looking through the code I'm not convinced that it does the
sensible thing with only "--max-level 1".

Otherwise it looks good.

Well, "--cache-size 512" probably doesn't make much good. That cache
isn't used by the Monte Carlo code and its size has turned out to be
rather uncritical after the involved code was last revised, long ago
now. I never bother changing the default size myself. On the other
hand, unless you're short on memory, it probably doesn't hurt either.

This are my conclusions after running several bots on CGOS:

   * 1000 games per level
         o L1 :: 1698
         o L10 :: 1813
   * 8000 games per level
         o L1 :: 1777
         o L10 :: 1938
   * 10.000 games per level
         o L1 :: 1736
         o L10 :: 1827

I don't know why the games at 10k games per level are weaker than the ones at 8k. Maybe it is because with the 10k-ones, I didn't rebuild gnugo with MAX_BOARD 9.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]