Look what the cat dragged in!

[Mike Ositoff]

> 
> DEMOREP1 at aol.com wrote:.
> > 
> > The example could even be more extreme--
> > 49 H W
> > 48 S W
> >   2 W H
> >   1 W S
> > with the same result.
> 
> This is indeed an extreme example, and shows how easily an unknown
> candidate with a few million to spend could walk away with an election
> under pairwise methods, at least if there is no other safety mechanism
> present.

No it doesn't. It shows a CW who beats both other candidates.
An unknown spending a few million dollars could get a majority
to rank him over both other candidates?

If you're suggesting that the pairwise methods are more succeptible
to campaign spending, then you must demonstrate that, in order
for your claim to have any credibility. Let's not start those
wild shots in the dark again.

My problem is with paid-off voters. I don't want a rich candidate
to win for no other reason than because lots of voters traded
their vote for a jug of wine. That's why, as Charles pointed out,
we'd need assurance that the ranking coudn't be used as a
signature. Aproval, with fewer possibilities, makes it easier
to avoid that problem.


> 
> In recent years, a number of wealthy private individuals have bought
> double-digit results in high profile U.S. elections.  Ross Perot in '92
> was probably the most striking example, until he dropped out because the
> "CIA threatened to disrupt his daughter's wedding."

Ok, then require public campaign financing, with no private
spending or contributions. But we've thoroughly discussed your
claim that voters can't be trusted to vote in their own best
interest when they elect a middle CW. They're adults. They
might not especially like Middle, but they consider him better
than the other extreme. That's for them to decide, not for
you to decide for them.



> 
> > 
> > I repeat my standard mantra-
> > nonpartisan nomination and elections for executive and judicial elections,
> > having a YES/NO vote on candidates in such elections, using Condorcet on
> > majority YES candidates, repeatedly dropping the lowest number ranked YES
> > choices to break any tie.
> 
> I believe you are on the right track by including some sort of absolute
> approval requirement, at least if you insist on going with Condorcet,
> but what if no candidate has a majority YES vote?

Y/N voting complicates strategy, and is inconsistent with the
relative nature of the choice.


> 
> Seems like you could replace relative ranking with some sort of absolute
> "rating" system (with a fixed number of values, like 0 - 10).  You could
> then treat the votes as approval votes, starting with the top level and
> them adding in successively lower levels as necessary until one or more
> candidates achieves a majority (or some other threshold).  You could

By not letting people vote 2nd choice over 3rd choice as strongly
as they can vote 1st choice over 2nd choice, you retain the
current strategy problem of FPP. That's elementary.

But adding votes starting from the top of the rankings, til
someone gets a majority, that's Bucklin. Very close to Approval
in merit. Not as good as VA, but Bucklin has precedent for use
in the U.S. (including in San Francisco). Bucklin has the
advantage of being as easy as IRO to count, but being a lot
better. But, not being significantly better than Approval, 
I wouldn't suggest it unless it's historical precedent would
greatly help its acceptance, and make it more winnable than
Approval. I stand by my previous statement that any rank
proposal isn't such a good idea, due to inevitable controversy
about the count rule (look at this list). Then, when it's in
use, everyone dissatisified with a result (and someone always
will be) will show that their count rule would have elected
their favorite, and the debate would be continually renewed.


> then use whatever method you want on the candidates who meet this
> threshold (or just use the greatest approval vote).  This would
> eliminate the separate YES/NO vote.
> 
> There would still be no guarantee that any candidate would get even a
> minimal level of approval from a majority of voters, so some other
> threshold would be needed.  Maybe you could quit adding levels when more
> then half of the positive votes have been added.
> 
> > 
> > I must berate Mr. Davidson for putting his catty comments on the ER list when
> > discussion of specific methods are supposed to be on the EM list.   Many ER
> > folks presumably have no great interest in the details/defects of various
> > election reform methods.
> 
> I wouldn't be too concerned about the ER folks.  Neither list is any
> place for pussies.

Demorep meant that ER is mostly for news, and not so much for
involved or protracted method debates. Discussion of IRO is
relevent on ER now, it seems to me, since it's being proposed
in Santa Clara & England, but Don needs to rein  himself in a bit,
and stick to the facts (and be more careful to decide what he's
trying to say and to evaluate how well he's saying it, before
posting it to ER).




> 
>