reversed rankings

[Blake Cretney]

On Wed, 28 Oct 1998 18:28:42   Mike Ositoff wrote:
>
>I'd like to answer the reversed-rankings nonsense, though I
>already have.
>
>Blake claimed that VA will pick the same candidate if you
>reverse all the rankings. Or at least implied that it will
>do so often under ordinary plausible conditions. Nonsense.
>

I never made any statement about which methods pass and which
fail this criterion.  In fact, Condorcet (EM) and IRO fail,
Schulze and Approval pass.  Margins or VA is irrelevant.

>With a 1-dimensional policy space, one of the twe most extreme
>candidates will win with a 1st choice majority if rankings
>are reversed, since everyone will have an extreme as last choice.
>
>Markus said that Schulze(VA) can't pick the same candidate
>when rankings are reversed. Do you disagree, Blake?

No.

>
>Blake says that if that happens, it means that VA says that
>the best candidate is also the worst. Nonsense again. MEthods
>don't pick the best or worst. Each voter has his best &
>worst, and the method picks based on certain more meaningful
>& concrete & verifiable criteria than picking the best.

Well, of course we differ on this point.  I think that the
point of election reform should be to get better government.
Do you have a different goal?  If so, how do you justify it.

I think you're right that whether the method picks the best
guess for best candidate is not verifiable.  It is, however,
falsifiable.  Clearly, if the method would pick the same
candidate for best and worst, it is wrong in one direction.

---
Blake



-----== Sent via Deja News, The Discussion Network ==-----
http://www.dejanews.com/  Easy access to 50,000+ discussion forums