Peter Zbornik wrote:
Hi Kristofer,
thanks,
so is it right to state, that:
"The only advantage of Contingent vote before Schulze in terms of
satisfied criteria is in the case of three candidates, where the
Contingent vote satisfies Later-No-Harm"?
It's not exhaustive; my point is that Contingent vote in the three
candidate case is like IRV. Thus, the criteria that Contingent possesses
and Schulze does not is exactly equal to those that IRV does and Schulze
does not - in the three-candidate case.
Other criteria IRV passes are: Plurality, Independence of Clones (if the
correct tiebreaker is used), Later-no-Help, Condorcet Loser, and Mutual
Majority.
Schulze fails both the LNH criteria (Later-no-Harm and Later-no-Help).
It passes Plurality, clone independence, CL, and MM.
IRV is also resists Burial strategy to a greater extent than do
Condorcet methods. Some Condorcet methods are particularly good at
resisting burial (BPW, Smith,IRV, first preference Copeland) as far as
Condorcet goes, but to my knowledge, they are still more susceptible to
that than IRV is.
In any case, if you're going to go down the simulated runoffs path, I
think IRV beats Contingent vote in all respects but summability. I
further think that going in the Condorcet direction would be preferrable
to trying to simulate runoffs..
----
Election-Methods mailing list - see http://electorama.com/em for list info
