[EM] Desperate to use Pareto?

[Markus Schulze]

Dear Mike,

you wrote (2 Jun 2000):
> Markus wrote (2 Jun 2000):
> > I want you to remember that the Ossipoff Subcycle Method (= your
> > favourite method until August 1998) actually violates the Pareto
> > criterion.
>
> What's the Ossipoff Subcycle Method? If you mean what I used
> to call the "subcycle rule", it was a way to avoid subcycle
> fratricide. It was never my favorite proposal. I never considered
> a proposal. The subcycle fratricides it avoided included subcycles
> that weren't clone sets.
>
> Later, I defined another method for that purpose, years ago.
> I liked it better than the subcycle rule.
>
> It seemed to me at the time that it did as well by lesser-of-2-evils
> standards as the other methods that I liked, with the added
> protection against defeats in subcycles affecting which subcycle
> turns out to be the one containing the winner.
>
> You pointed out that it could violate Pareto. But that didn't
> seem important, because I was more interested in majority rule
> between the elements of the main cycle, than the Pareto Criterion
> within a subcycle. It violated the Pareto criterion if there was
> a unanimous defeat within a subcycle. So what? If that subcycle
> should be the winner, with respect to the main cycle, ensuring
> that defeats within the subcycle won't spoil that win seems more
> important than whether Pareto is violated within the subcycle.
> I felt that what happens in the main cycle is more important.
>
> I don't know if either of those subcycle-protecting methods
> would meet BC, or would strictly meet MDC or GSFC. It doesn't
> matter, since they were never proposals.
>
> This was all just hypothetical, since I never proposed those
> methods.
>
> ***
>
> In any case, your bringing up methods that were defined but
> never proposed, and which haven't even been discussed for almost
> 2 years, seems unexplainable. I said that Pareto is useless because
> it is failed by so few methods. You named one method that failed
> Pareto, a method that was never proposed. Are you saying that
> Pareto is an important criterion because a never-proposed method
> that hasn't been discussed for 2 years fails it? Then let me
> clue you in, Markus: the subcycle rule isn't really a burning
> issue.
>
> As I said, a criterion that is failed by as few methods as
> Pareto is is pretty much useless for comparing voting systems.
> Sure, you can find uses for it if you assiduously search the
> archives. Another thing I recommend to you: You can use Pareto
> to compare Sequential Pairwise to other methods. Good luck, and
> have fun.

There is no reason to feel accused. I only wanted you to remember
that the question whether a given election method violates the
Pareto criterion can be difficult. I could also mention Saumur's
Smith//RandomCandidate.

Markus Schulze
schulze at sol.physik.tu-berlin.de
schulze at math.tu-berlin.de
markusschulze at planet-interkom.de