RE: [Bug-gnubg] Re: The importance of METs

[Albert Silver]

> Is this what has been touted as a 1.2% improvement? I would not
conclude
> that
> from those numbers.

No, you shouldn't. I hallucinated the size of the difference by a
decimal point, and it should be 0.12%. Some results were presented
showing the difference with Jacobs/Trice's MET and the difference was
smaller, so I asked the group a different question: If relatively
important changes are made to the MET (in the order of 1%) and the
results change so little, the importance of which MET is used seems
negligeable, in which case, one might as well keep the Woolsey table
using Neil's numbers. This has led to discussion on this and whether the
changes brought are so small.

> 
> > I expected the correlation to be much higher - I am surprised that
the
> MET
> > used influences the outcome of more than a quarter of matches
(although
> > these MET's are much more different than Snowie and mec26)
> 
> Better variance reduction may fix this. If I understand your
methodology,
> if the
> length of a game but not the result depends on the MET, then the rest
of

I was pretty busy yesterday and missed this. If this is true, then the
testing is not revealing one other change in result: from No Double/Take
to Double/Take as this won't change the length of the game immediately
in most cases (short of a sudden market loss the next roll).

                                                        Albert