Re: [Bug-gnubg] SGF Double Analysis Format

[Guido Flohr]

Hi,

On Tue, 2012-08-21 at 20:25 -0400, Ryan McKinnon wrote:
> ;B[double]DA[E ver 3 2C 1 0.000000 1 0.572881 0.222155 0.024300
> 0.139598 0.006174 0.243145 0.527005 0.572881 0.222155 0.024300
> 0.139598 0.006174 0.243145 0.516940]BM[2]

Here are my findings:

"E": Evaluation analysis (as opposed to "R" for rollout)
"ver 3": SGF version, currently always 3
"2C": 2-ply cubeful
1: deterministic
0.000000: noise
1: prune

See WriteEvalContext() in sgf.c.

Following are two arrays (for both players?) with seven probabilities
and equity values.  There are symbolic constants for the indices into
these two arrays in eval.h:

#define OUTPUT_WIN 0
#define OUTPUT_WINGAMMON 1
#define OUTPUT_WINBACKGAMMON 2
#define OUTPUT_LOSEGAMMON 3
#define OUTPUT_LOSEBACKGAMMON 4

#define OUTPUT_EQUITY 5 /* NB: neural nets do not output equity, only
                           rollouts do. */
#define OUTPUT_CUBEFUL_EQUITY 6

Pure coincidence, but I'm currently working on displaying GNUBG analysis
results in Gibbon.  Work in progress, but you can see my current state
in git, see mainly:

http://git.savannah.gnu.org/cgit/gibbon.git/tree/src/gibbon-analysis-view.c
(starting with gibbon_analysis_view_set_move(), the rest is Gtk stuff)
http://git.savannah.gnu.org/cgit/gibbon.git/tree/src/gibbon-met.c
http://git.savannah.gnu.org/cgit/gibbon.git/tree/src/gibbon-met.h

Hope that helps!
Guido
-- 
Империя ООД                         | Imperia OOD
ул. „Княз-Борис-I“ № 86, София 1000 | ul. "Knyaz-Boris-I" № 86, Sofia
http://www.imperia.bg/