This method is based on rankings with truncations and approval cutoffs.
Let X be the candidate approved on the greatest number of ballots. Let Y be
the candidate ranked on the greatest number of ballots.
If X and Y are the same candidate, then this candidate wins.
Otherwise, a ballot is drawn at random. If this ballot distinguishes these two
candidates, then this ballot determines which of the two wins.
Otherwise, the candidate Z ranked highest on the ballot gets to say which
member of the set {X,Y} is the winner
The rationale for the last "otherwise" is that a voter that was not interested
in distinguishing X and Y would probably trust her favorite candidate to make
that distinction for her., and if she did feel strongly that one was better
than the other, then she would probably want to rank at least one of them.
Note that the contest is between the least truncated candidate and the most
approved candidate, two candidates that should both be broadly tolerated by
the voters in comparison with the other candidates.
It seems to me that this method removes almost all incentive for insincere
order reversal, and most incentive for order collapse (except for the collapse
inherent in truncation).
To simplify things, voters that do not wish to rank candidates should have the
option of merely specifying which candidate is their favorite, which other
candidates they approve, and which of the rest they would not truncate. This
would give them as much power as anybody else in determining candidates X and Y.
A ballot that specified only these four levels would still have a decent chance
of distinguishing X and Y, as well, but if not, then the choice of the ballot
favorite Z would probably be satisfactory in the eyes of the ballot's voter.
Comments?
Forest
<<winmail.dat>>
---- election-methods mailing list - see http://electorama.com/em for list info
