STV for party candidate lists?

[Mike Ositoff]

On Wed, 29 Jul 1998 DEMOREP1 at aol.com wrote:

> Demorep wrote (28 Jul 1998):
>  Whether or not some sort of other criteria is violated is irrelevant.
> 
> Mr. Schulze wrote-
> 
> Why?
> ---
> D- Any tiebreaker (AFTER doing Condorcet head to head math for N = 1, N = 2 or
> N = whatever) violates some sort of criteria (using some sort of manufactured
> bad case example).   As I just wrote-- Current legislative bodies are a very

What? You're saying that if there are _any_ criteria unmet, 
like Arrow's mutually incompatible criteria, then all criteria
are irrelevant? Have you been listening to CVD?

Some of us feel that a violation of the Condorcet Criterion
is a violation of expressed public wishes. Some of us feel
that majority rule, and the LO2E problem are important, and
some of us have pointed out the importance of GMC in regards
to those standards.


> [bad] approximate representation of the total voters in very many countries,
> etc.
> 
> To break ties, Bucklin (to get winners) and/or Reverse Bucklin (to get losers)
> is simple enough for most (99 percent +) voters (who do not stay up all day
> and night worrying about strategic and/or sincere/insincere voting).   

As for Reverse Bucklin, you never did say what its advantage was.

As for Bucklin, its defensive strategy requirement is obvious,
and doesn't require staying up all night: Try not to vote
a ranking that extends past the alternative that you're likely
to need as a compromise, the alternative that you & voters like
you can give a vote-majority to.

Though Bucklin is better than IRO, and is perhaps the best
method short of EM's best (being marginally better than Approval),
it does routinely require defensive strategy and the predictive
knowledge that makes well-informed strategy possible. So it
isn't at all as good as EM's best.

> To paraphrase U.S. Navy Admiral David G. Farragut in his August 5, 1864 attack
> on Mobile, Alabama during the U.S. Civil War ---
> 
> Damn the tiebreaker criteria, full speed ahead (for election reform to get rid
> of minority rule).

What? Getting rid of minority rule sounds like, if not a criterion,
at least a standard that could motivate a criterion, to measure
compliance with that standard. As I said, the Condorcet & 
GMC criteria protect majority rule (thereby getting rid of
minority rule, as you requested).

This idea of forging ahead without regard to criteria explains
how bad methods get advocates.


Mike Ossipoff

 
>