thinking about it, it may be that Range Voting (and its variants) is the ONLY
FBC-obeying
voting method which is immune to candidate-cloning (with voters assumed to have
slight
preferences among the clones) and disregarding non-deterministic election
methods.
That would be a very interesting theorem. I wonder if it is true. Note that
it is essential
here that the score range be a CONTINUUM.
Range voting with discrete score set like {1,2,3,4,5} is vulnerable to cloning
if the voters decie to non-equally score clones, but with continuum scores
with the voters assumes to make arbitrarily small score perturbations, range is
generically immune.
wds
----
election-methods mailing list - see http://electorama.com/em for list info
