[EM] Comments--few voters, 0-info

[Forest Simmons]

On Wed, 14 Mar 2001, MIKE OSSIPOFF wrote (among other things):

> I was surprised & disappointed to find out that, with very few voters
> the above-mean strategy isn't valid in Approval. I kept denying it
> for a long time, not because it was something I didn't want to believe,
> but because it really seemed contrary to what I expected.
> 

Approving everyone above mean doesn't maximize expected utility, but it
does maximize the probability that the winner will have above average
utility.

I suggest that we start with this, and then if the mean is below the
median, use the median as the cutoff instead, so that we are not diluting
our voting power. If there is one big gap, cutoff everything below that
gap as well. Then refine with gut feelings and other methods of
inspiration. 

Forest