[Bug-gnubg] Extending match equity tables

[Guido Flohr]

Hi,

I have to extend match equity tables to length 100 and looked at the
corresponding code in matchequity.c, function ExtendMET().  There was an
assertion that can never fail:

    g_assert ( 6.0f * rSigma > nScore1 - nScore0 );

nScore1 is always less or equal nScore2, and rSigma is always positive.

I have fixed that inconsistency in revision 1.83 of matchequity.c.  If I
see it correctly, the only difference to the previous versions are
caused by rounding errors.  A review by someone more familiar with the
code will still be welcome.

However, when I set MAXSCORE to 100 (in my code, not in gnubg), the
assertion fails.  Does anybody know where I can find a description of
"David Montgomery's extension algorithm" as mentioned in the comment to
the function?  Or can anybody explain the details of the algorithm?
Where does the factor 6.0f come from?

If I see it correctly, I could solve my problem by extending the series
of standard deviations arStddevTable[].  But where do the numbers come
from?

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