[EM] One more thing: If voters just assumed p(R) = p(D)=.5

[Michael Ossipoff]

But what if what voters expect of the outcome isn't influenced by the
special 3-candidate defection strategy that I spoke of--What if voters
merely assumed that the Democrat and the Republican both had .5 probability
of winning?

Then, as Lomax said, his Green Democrat would have reason to approve the
Green, by better-than-expectation strategy. 

Lomax meant that example to refute my suggestion that preferrers of Middle
of 3 candidates don't have reason to approve anyone else. If they don't,
that makes it easy to estimate the numbers of preferrers of the 2 extreme
candidates, from their approval count totals.

But suppose some Green Democrats would approve Green for the reason
described in the 1st paragraph above?

I've said earlier that, when Approval gives non-Republocrat parties and
candidates the viability that is genuinely theirs, statisticians and honest
poll-takers of all political persuasions will be very interested in
determining the numbers of preferrers of the various candidates and parties.
Under the conditions described in the 1st paragraph above, these people and
others would estimate the relevant preference numbers based on the available
data, including approval count totals, polling results, and both together.

I emphasize, regarding my statement that my statements about Middle
preferrers would have no reason to approve anyone else, that that statement
was intended to apply only to elections with only 3 candidates. 

My discussion of results based on better-than-expectation strategy is more
general in its application.

I reply fully to questions and objections, and some of these messages have
been posted in order to completely reply to Lomax's example.

Mike Ossipoff