[EM] STV and weighted positional methods

[Kristofer Munsterhjelm]

Kathy Dopp wrote:
>> From: Kristofer Munsterhjelm <km-elmet at broadpark.no>
>> Subject: [EM] STV and weighted positional methods
>>
>> Lately, I've been busy trying to generalize STV, the method, to be
>> applicable to any weighted positional method. At first, this seems quite
>> easy to do. STV (discarding Meek) works like this:
> 
> Yikes Kristofer,
> 
> I would rephrase that to say "I've been busy trying to *change* STV,
> the method, to treat all ballots equally so that it does not exhibit
> nonmonotonicity, finds majority winners, etc...."
> 
> Let's not forget that STV, at least as Minneapolis is trying to
> implement it, and probably San Francisco as well, has all the same
> flaws as IRV does.
> 
> We need a radical change.
> 
> I see that what you are suggesting as a change to STV such as using
> the Borda method does seem to be a *lot* better than current
> implementations of STV, so I might recommend not calling it "a
> generalization of STV".

Thank you, but even if I can find a reweighting function, the method 
wouldn't be flawless. The transferable vote (quota-reweighted?) method 
using Borda would reduce to Borda-elimination (Baldwin) in the single 
winner case, and Baldwin is nonmonotonic. When there is one, it does 
elect the Condorcet winner, though, which is something IRV doesn't do.

My primary reason for the post (and the change/generalization of STV) is 
to find out what makes a good multiwinner method a good multiwinner 
method, and thus how to create good multiwinner methods. In a 
theoretical sense, what I wrote about is a generalization of STV, 
because its skeleton is like STV: if something's above the quota, elect 
and reweight, otherwise exclude. In practice, there may not be much 
STV-like left about it, but I thought I should explain why I said what I 
said :-)