Ken,
Interesting. How exactly does one calculate the "correlation" of the
pairs of candidates?
Chris Benham
Ken Kuhlman wrote:
I've recently become interested in election methods, and have noticed
that while the Borda count has been shown to have many desirable
properties, it is customarily dismissed because of its susceptibility
to clones. The method proposed below, which I call "Correlated
Instant Borda Runoff," is basically a tweak of Baldwin to solve the
clone problem.
Individual ballots are scored according to the Borda count, and then
all possible candidate pairs are ranked according to correlation. The
Borda loser of highest correlated pair is eliminated, and the next
round proceeds with the remaining candidates.
Under this method, clones can be seen as a special case of correlated
pairs (they're perfectly correlated), and so the Borda loser of a
clone pair is eliminated immediately, before they can spoil the election.
For example, consider an election between A, B, and C, where B & C are
clones:
1:A>B>C
1:B>C>A
While Borda gives the election to B, CIBR recognizes BC as highly
correlated (and thus a clone), and so eliminates C and declares the
election a tie between A & B.
Note that if the clones are transposed on the second ballot, as below:
1:A>B>C
1:C>B>A
Then AB are clones as well, and the election becomes a three-way tie.
(The Borda property of declaring symmetrical elections ties is
retained under CIBR).
------------------------------------------------------------------------
----
Election-methods mailing list - see http://electorama.com/em for list
info
----
Election-methods mailing list - see http://electorama.com/em for list info
