[EM] To Bill Lewis Clark, re: Approval, CR, & IRV

[Bart Ingles]

Dgamble997 at aol.com wrote:
> 
> Bill you wrote in part:
> 
> >> To gloss over lots of details it is generally presumed that voters
> >> behave rationally and vote strategically to maximise the utility of
> >> their outcome in an election.
> 
> >I think that presumption is almost undoubtedly false, in the general
> case.
> >I suspect it's based on some misleading observations:
> 
> I'd agree almost entirely with the points you've made about strategic
> voting and the likelihood that people will use strategic voting. Most
> people on the EM list, I think, will disagree with you (particularly
> Approval supporters). This is because Approval starts to fail badly as
> a method when people stop using strategy and start approving all the
> candidates they like.

I'm not sure how you define "like", but if the voters simply approve all
candidates they consider to be "better than average", then approval
comes off pretty well (as well or better than IRV in terms of electing
sincere Condorcet winners-- certainly more stable in this regard-- and
it beats IRV hands-down in terms of social utility).

In fact "better-than-average" is the only approval strategy Merrill uses
in "Making Multicandidate Elections More Democratic".

Bart