[EM] MMPO[IA>=MPO]

Forest Simmons fsimmons at pcc.edu
Thu Oct 17 14:42:44 PDT 2013


As Kevin noted  MMPO[IA>MPO] doesn't always elect from Smith.  Here's an
example:

26 A
24 C=A
24 C=B
26 B

All three candidates have greater Implicit Approval than Max Pairwise
Opposition.

Candidate C has the MinMaxPairwiseOpposition, so is elected by
MMPO[IA>=MPO].

But Candidate C is the ballot Condorcet Loser, so cannot be in Smith.

Yet it seems likely to me (judging from the ballots) that C (though not a
ballot CW) is more likely than A or B to be a compromise, center candidate.

Indeed it is not hard to imagine (in the face of approval style ballots and
a disinformation campaign about the viability of only A or B, for example)
how the above ballot set could arise from true preferences of

26 A
24 C>A
24 C>B
26 B

in which C is the CW.  If approval voters were (falsely) convinced that A
and B were the two front runners, and then used the strategy of voting for
their preferred of the front runners and whomever they liked better, there
would be an approval tie between A and B, the same that would happen in
Plurality voting with the same disinformation.

I believe that is essentially what happened in the (Plurality based) 2000
US presidential election, with the tie being broken (unconstitutionally) by
the Supreme Court.  Informal pre-election polls and mass popular rallies
favored Nader over both Gore and Bush, but the corporate debated and
corporate media conspired to cover up Nader's popularity.  Then after the
election, Nader suddenly lost his popularity, not because of any change in
his stances on the issues, but because of the false perception that he was
a "spoiler," which was just a self-fulfilling prophecy of the corporate
disinformation campaign.

Under MMPO[IA>=MPO]  Nader would have won, even if not quite under
Approval, as in the above scenario.

In other words, MMPO[IA>=MPO], even with approval style ballots is a better
remedy to the spoiler problem than Approval (which does always elect from
ballot Smith if not actual Smith).  I think that this was Kevin's idea on
one of his initial posts to the EM list years ago.  He saw that MMPO was
better at discerning central candidates than Approval, even when limited to
approval style ballots.  He abandoned the idea after seeing that MMPO by
itself failed Plurality.  But now we have fixed that!

So far I don't see any bad side of MMPO[IA>=MPO].

Forest
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20131017/a38ebc2a/attachment-0003.htm>


More information about the Election-Methods mailing list