[EM] Responses to some of Forest's ideas

[Richard Moore]

Forest Simmons wrote:

> 
> I vaguely remember that there was a sudden realization that the above mean
> criterion was necessary but not sufficient for optimizing expected utility
> in a zero-info environment.


I think it's necessary and sufficient for zero-info, large 
populations.


> 
> I also remembered that you made excellent progress on the small population
> case, and (prematurely?) gave you credit for an asymptotic solution in the
> large population case :-) 
> 
> Forest


I did suggest a formula, I just never came up with a good 
test to see if it's accurate. It asymptotically approaches 
the mean for large populations so it should be reasonably 
close for populations of say a couple dozen, but it probably 
has some error for really small populations (say, less than 
10). Also, it depended more on the number of "yes" votes 
that are expected to be cast, rather than the number of 
voters. I don't have it handy but if I come across it in my 
saved e-mail I'll post it again.

Richard