[EM] Approval reducing to Plurality

[Kevin Venzke]

Hi,

In my simulations involving repeated Approval polling prior to having the
election, scenarios tend to eventually settle on two frontrunners. (If
there are three candidates and one is closest to the median voter, the 
two frontrunners will be the middle candidate and one of the flanks.)
This means strategic voters will vote for exactly one of the frontrunners
(i.e. not neither and not both). Makes sense...

If the polls can't settle on two frontrunners, such as if there's a cycle
and nobody can remain content with one way of voting irrespective of
how others vote, then the result tends to seem arbitrary, and I bet there
would be complaints that some candidate spoiled it.

So will votes tend to be Plurality-style? Maybe, maybe not. It doesn't
matter that much as far as I can see, assuming the polling was accurate.
One of the non-frontrunners (as perceived at the time of the election)
winning should be very unusual.

(Maybe ask instead: Will Approval *polls* resemble Plurality polls, both
in how they are asked and how voters respond to them?)

Will Approval just turn into Plurality? Not really, because in Plurality
once you have two frontrunners, this pretty much can't change: There's
no way for a third candidate to break in. In Approval it can: If the
third candidate can win, the polls should reveal this prior to the 
election, changing who the frontrunners are perceived to be.

The way I see the difference is like this: If you have three candidates
trying to run, all trying to run near the center but with one between the
other two (and the center one being *most *likely to be the median voter's
choice) then Plurality will never elect the middle one, and Approval will
never elect one of the flanks (and may pick somewhat arbitrarily which one
can't win, if they were close in strength).

Kevin Venzke