[EM] Variations on IRV

[Forest Simmons]

Tom, I want to apologize because upon re-reading this previous message it
sounds kind of arrogant.

I don't want to be guilty of discouraging exploration of new ideas. Even
in the unlikeliest places there are some gems of knowledge and great
insights awaiting.

I am curious, though, to know if you agree with me as to the facts of
iterated IRV.

Also, when I said that PAV is the best PR method for an half dozen or
fewer winners, I should have added the qualification "in my opinion" since
we haven't made any comparisons on the basis of objective criteria as of
yet. (The size limitation is due to computational complexity, exponential
in the number of winners.)

I guess grading too many finals brought out the worst in me.

Just be glad you're not one of the students in the last class to get
graded.

Peace,

Forest

On Sat, 24 Mar 2001, Forest Simmons wrote:

> Tom, I am curious about something that doesn't quite add up in my mind.
> 
> You know about recursive, iterative methods for improving IRV (as well as
> other methods).
> 
> You agree with me that each iteration is an improvement.
> 
> You are aware that with each iteration the method has a greater chance of
> picking the Condorcet winner if there is one. 
> 
> You know that in a finite number of iterations IRV will reach the
> Condorcet winner if there is one. 
> 
> (Correct me, if I am wrong on any of the above assumptions.)
> 
> What puzzles me is how you can waste time working for IRV and its little
> variations in view of the above facts (if you agree that they are facts).
> 
> If you are going to require voters to use preference ballots, why don't
> you just support some version of Condorcet that allows equalities as well
> as inequalities?
> 
> Just Wondering,
> 
> Forest
> 
> 
>