A while back I said on EM that Bucklin met FBC, but Markus posted a Bucklin
FBC failure example. I don't know if I'd said that ordinary
non-equal-ranking Bucklin passes FBC, or if Markus's demonstration was based
on an incorrect definition for a majority in Bucklin.
It seems to me that, in Markus's example, Bucklin failed because of how the
upranking of favorite to 1st place affected what consitutes a majority. But
a majority in Bucklin is a vote total greater than more than half of the
voters, as opposed to more than half of the votes cast. As I said, I don't
know which explanation explains that failure example.
It certainly does sound as if ERBucklin(whole) passes FBC. That's
encouraging that such a well known method as Bucklin has an obvious version
that meets FBC.
Shall I call it ERBW? It meets SDSC, but it loses SFC compliance. Bucklin,
not being a pairwise-count method, has an easier handcount. But that doesn't
seem an important issue, because public elections are where FBC compliance
is really needed. Bucklin's FBC compliance is a nice added benefit for
organizations that want rank-balloting but don't do a computer count.
Bucklin probably doesn't ever need strategic lower ranking or power
truncation, due to being stepwise instead of pairwise.
I don't know which is better. Certainly MDDA (which I call "Majority
Choice") has a briefer and simpler definition, and that's probably the
decisive factor, since the two methods' merit seems similiar. Criteria-wise,
MDDA seems better too, since it meets SFC.
Mike Ossipoff
_________________________________________________________________
Dont just search. Find. Check out the new MSN Search!
http://search.msn.click-url.com/go/onm00200636ave/direct/01/
----
Election-methods mailing list - see http://electorama.com/em for list info
