The "Turkey" problem (Re: [EM] 2-rank and N-rank Condorcet)

[Rob Lanphier]

Kevin Venzke wrote:

>Specifically, it seems that a compromise could be elected more easily [with "two-rank Condorcet"] than under
>Approval, while a rogue/turkey would have a harder time being elected than under Condorcet (at least to an extent).
>
It seems like two-rank Condorcet makes things incredibly complex, and 
creates more problems than it solves.

However, I agree that there's a general fear of the "turkey" factor.  
I've been mulling this one for a while, and I think someone (maybe me) 
has posted this solution to the list before, but if not, here goes 
again.  It seems that the simple solution is to have a primary election 
using Approval, and a general election using Condorcet.  This would have 
two benefits:
1.  it would winnow the field to the point where sensibly constrained 
debates could occur
2.  it would ensure a core following for all of the candidates.

Of course, the difficult part is determining how to do the primary.  
There's three ways that make sense to me:
1.  Have a candidate quota of X candidates.  The X candidates with the 
highest approval rating advance to the general eleciton
2.  Pick a percentage quota Y.  All candidates (no matter how many) who 
exceed Y% of the vote advance to the general election.
3.  A combination of #1 and #2, where some minimum number of candidates 
is set.

I think I like #3.  My gut tells me that X should be something like 5, 
and Y should be something like 25%.  That means that all of the 
candidates who exceed 25% approval rating advance to the general 
election, and that if there aren't at least five, then the top five move 
on, regardless of their approval.  This round would weed out most of the 
turkeys, while allowing for a general election in which the turkeys make 
themselves obvious in debates and the rest of the campaign.

This method seems the easiest.  Primary campaigns do serve a role in 
focusing the population on a few candidates who can then be vetted in 
depth.  Even though I'm a huge fan of a more open field, I like the idea 
of having a reasonably small field of cnadidates.

Strategy for the primary is a little hard to calculate, but the reason 
for not having a cap on the number of candidates is to make the strategy 
less of an issue.  A cap would create a zero sum game where an 
acceptable compromise candidate would potentially knock out a favorite.  
I think that with this system, the proper strategy is to vote for the 
most favored of the two frontrunners, and all candidates preferred to 
them (same as a general election with Approval).  My only concern with 
not having a cap is that it might lead to "attack of the clones", where 
a party would run as many candidates as possible, in order to pack the 
field with party candidates.

An alternative is some sort of Coombs primary, Condorcet eleciton.  
However, I'm getting too tired to think about that.  ;)

Rob