[EM] PR-STV with approval based elimination

[Raph Frank]

On Fri, May 1, 2009 at 3:56 PM, Jonathan Lundell <jlundell at pobox.com> wrote:
> The strategy is effectively identical in both versions, despite the
> difference in form. That is, in both your mechanisms it seems apparent on
> the face of it that a voter would not have a motivation, strategic or
> otherwise, to choose a different approval set in the two methods.

Right.  However, later no harm is preserved for the ranked ballot.

One of the concerns raised with ranked ballot methods is that people
won't rank more than 1.  If they do that, then you lose the benefits
of ranked ballots.

This is not a concern here.

For the approval vote, I agree that later no harm doesn't apply, but
it never applied to approval voting anyway.

It would allow other factions potentially to decide which of your
candidates get elected.

If the 2 votes are linked so ranks must be consistant with the
approval vote then later no harm is lost as your ranked ballot
determines who you can and can't approve.  However, it means that
voters must decide if they want to influence their own faction's
elimination ordering or a different one.