[Bug-gnubg] 0-ply cubeful equities

[Joern Thyssen]

Hi,

I've modified how gnubg calculates cubeful equities to more closely
match what I've earlier implemented for match play.

I played a 250-game money session of the new code versus the old code
(cube 0-ply, chequerplay 0-ply). The final score was 284-240 or 0.18
ppg. The luck adjusted result was 19.5 points or 0.08 ppg.

Michael, can you re-check some of the positions you mention in your
gammonline posting? I'll let you know when I've committed the code.


Jørn