[EM] Approval vs ??

[Bart Ingles]

I don't disagree with Mike's post, but would point out that the
differences between Condorcet and Approval with regard to SU are fairly
small.

With results averaged over many runs, Condorcet generally has the edge
over Approval.  Although in Merrill's "Making Multicandidate Elections
More Democratic", Approval edged out Condorcet in the three-candidate
case for both the Random Society and Spatial Model runs.  With more than
three candidates, the two were very close (within a couple of percentage
points) regardless of the number of candidates in the Spatial Models,
with both considerably better than the runoff-based methods.

When looking at individual "worst-case" outcomes, Approval seems better
at avoiding the "hated middle" situation, most pronounced when there are
three candidates.  With more candidates, there is less of a difference. 
Granted, since neither is a Duvergerian method, there will be a tendency
for more and better candidates to enter the race, making it difficult
for a "hated middle" candidate to win anyway.  But not always-- suppose
the race is for a minor office where not many people are willing to run
even in the absence of Duvergerian barriers.  Or (with Iraq or
Afghanistan in mind) suppose the electorate is divided into factions
with mutually-exclusive ideologies, where it may not be possible to
define a suitable consensus candidate let alone elect one (I don't
really know if this is the case in those countries, but just suppose...)

Anyway, it seems to me that there are probably situations where either
of Approval or Condorcet are better suited than the other.  To me it
makes the most sense to start with the simpler of the two, and introduce
the other wherever the need has been actually demonstrated.

Bart



MIKE OSSIPOFF wrote:
> 
> I prefer Condorcet to Approval because Condorcet receives and counts all of
> the pairwise preferences that anyone wants to vote. That's why Condorcet is
> better at electing the CW, and meets SFC, GSFC & SDSC. And it's why
> Condorcet does better at SU (Though Approval also does very well at SU,
> significantly better than IRV).