[EM] IA-MPO restricted to approval style ballots

[Forest Simmons]

This Implicit Approval Minus Max Pairwise Opposition is attractive to me
for the same reason I liked Kevin Venzke's "Median" that he suggested on 5
March 2003, which he alluded to in his first EM message less than a month
earlier: namely it had an uncanny cunning at finding the "center
candidate." the one whose max distance (in some sense) to any other
candidate was minimal.

It turned out that "Median" failed the Plurality Criterion, so we set it
aside with a promise that some day we would revisit the basic idea to see
if it could be rescued.

IA-MPO is one idea in the same spirit that still violates Plurality, but
not the modified version that makes more sense when you think about it in
the geometrical context of Keven's idea.

The modified version says that if A is ranked top above C on more ballots
than C is ranked, then C should not win.

IA-MPO satisfies this version.  And under perfect information it would
satisfy the more stringent version in practice, since under perfect
information the C supporters would vote C (or C>A if allowed) instead of
C=A.

I think that we should spend some time in reconsidering Kevin's idea in
general, and the IA-MPO idea in particular.

Forest
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20150602/67144b2f/attachment.htm>