Re: [Bug-gnubg] Ratings

[Jim Segrave]

On Sat 16 Dec 2006 (11:43 -0300), Albert Silver wrote:
> Hi all and merry Christmas,
> 
> I received the following question regarding the tutorial, and included my 
> reply.
> 
> >1. Comparing some information with the "GNU Backgammon Handbook" on
> >http://www.gnubg.org/win32/gnubg/gnubg.html#Analysing I found a
> >difference between the definition of error rates in those two sources.
> >
> >You state:
> >0.0000 - 0.0017 Supernatural
> >0.0017 - 0.0062 World Class
> >0.0062 - 0.0083 Expert
> >0.0083 - 0.012 Advanced
> >0.012 - 0.018 Intermediate
> >0.018 - 0.026 Casual Player
> >0.026 - 0.035 Beginner
> >0.035 - infinity Awful!
> >
> >The GNU Backgammon handbook says:
> >0.000 - 0.002  Supernatural
> >0.002 - 0.005  World Class
> >0.005 - 0.008  Expert
> >0.008 - 0.012  Advanced
> >0.012 - 0.018  Intermediate
> >0.018 - 0.026  Casual Player
> >0.026 - 0.035  Beginner
> > > 0.035  Awful!

>From the source:, analysis.c


const char *aszRating [ RAT_UNDEFINED + 1 ] = {
  N_("Awful!"), 
  N_("Beginner"), 
  N_("Casual player"), 
  N_("Intermediate"), 
  N_("Advanced"), 
  N_("Expert"),
  N_("World class"), 
  N_("Supernatural"), 
  N_("N/A") };


static const float arThrsRating [ RAT_SUPERNATURAL + 1 ] = {
  1e38f, 0.035f, 0.026f, 0.018f, 0.012f, 0.008f, 0.005f, 0.002f };

which make the figures the handbook vaues you quote above (0.000-0.002
= Supernatural, for example)



-- 
Jim Segrave           address@hidden