[EM] Approval voting heuristics

[Bart Ingles]

Bart Ingles wrote:
> Your strategy of voting for n/2 candidates might be an easy heuristic,
> but if you want to use it I would add an exception for candidates whom
> the voter rates very high or very low.  So a quick and dirty rule might
> be something like "...vote for the half of the candidates you like best,
> but always vote for a candidate rated above 8 or below 2 on a scale of
> 0-10...", where the 8 and 2 figures would be replaced by whatever
> numbers best approximate the curve for the average utility formula.

Whoops.  I should have written "...but always vote for a candidate rated
above 8, and never vote for a candidate rated below 2..." (possibly 9
and 2).

I'm not sure all that is needed, though.  The basic strategy of voting
only for candidates whom you rate above average assumes you are using
von Newmann-Morgenstern utilities, i.e. you are indifferent between a
candidate with utility 0.5, and a 50-50 lottery between a favorite rated
1.0 and an adversary rated 0.0.  If voters tend to rate on a different
scale, it might be more effective to simply let voters use intuition.

In other words, in a three-way race with no polling info, we assume a
voter's favorite has a 50-50 chance of defeating his least favorite. 
Simply letting the voter decide whether he is comfortable casting an
additional vote for the 3rd candidate is probably the best way of
finding out whether that candidate is above or below 0.5 on a vNM scale.

If nothing else, approval voting strategy is easy to grasp, and there
are no important counter-intuitive strategies I am aware of.  In the
absence of polling information, approval voting guided by voter
intuition is probably better than the systems currently in use.