[Election-Methods] Improved Approval Runoff

[Abd ul-Rahman Lomax]

At 04:01 PM 8/22/2007, Kevin Venzke wrote:
>To be clear, I thought you were claiming that any method (not just a
>Condorcet method) that allows such votes could simply be called a mix
>of Condorcet and Approval. That's why I brought up ER-IRV.

Oh. That is so blatantly false that it did not even occur to me that 
what I wrote could be read that way, though, of course, it is a 
possible reading, even a literal one. No, what I intended to write was:

Any *Condorcet* method which allows equal ranking and truncation is 
Condorcet/Approval, *not* referring to a specific method by that 
name, and not specifying exactly how the votes are counted and the 
winner determined, beyond it being a Condorcet method.

In fact, however, I also had in mind IRV, and IRV / Approval is pretty simple.


> > There are lots of Condorcet methods, of course. But allowing equal
> > ranking and truncation causes most of them to take on aspects of
> > Approval, and, of course, voters could elect to cast what are
> > essentially approval ballots. If all of them did, I would assume that
> > the Approval winner would win....
>
>If the method strictly satisfies Condorcet, then yes, the candidate with
>the most votes would have to win.

Right.

>However, MinMax (pairwise opposition), which does still determine its
>winner strictly from the strengths of pairwise contests, would not
>necessarily elect the approval winner on such votes. On the other hand,
>it satisfies FBC and LNHarm.

No claim is made that such a method is optimal. I do claim that 
allowing equal ranking and truncation generally improve methods.

("The Approval winner" is indistinguishable from a Condorcet winner 
in the context described, a Condorcet method that allows equal 
ranking and truncation. At least that is how it looks to me, superficially.)