Peter Zbornik wrote:
Hi Richard,
maybe a second round could take place between the winners of competing
methods, say Schulze winner vs Approval or Bucklin winner or any
other endorsed method. This would allow for election-methods
"competitions", and could address potential weaknesses of each method.
When one of the method would generate a "bad" winner, then the other
method could still give a "good" winner. For instance, if the Schulze
method would generate a winner noone has heard of before (the dark
horse), then in the second round, when he meets the Bucklin or
Approval winner, he or she might lose the second round, after new
details of his/her political past come to public knowledge due to the
increased attention given. The election methods should be different
for this approach to work. The obvious downside of this approach is
increased complexity and thus less public support. With this method
IRV might be used as one of the methods, or the old method could be
used against the new method.
I'd like to add to this that plain old top-two runoff can have the
honest CW win in a game-theory equilibrium if everybody communicates
with everybody else, and there are only a few candidates. This is
interesting, because Plurality has no such equilibrium. Perhaps an
advanced runoff method could have such "candidates in honest X-sets will
win under complete information" (X being Smith, Landau, whatnot)
equilibria. It does seem that runoff methods can improve upon resistance
to strategy, at least, because the last round is honest among the two
candidates that remain.
----
Election-Methods mailing list - see http://electorama.com/em for list info
