[EM] PR-STV with approval based elimination

[Kristofer Munsterhjelm]

Dan Bishop wrote:

> You could use Plurality (with vote-splitting between equally ranked 
> candidates) to determine surpluses and a different method to determine 
> eliminations.  For example,

[snip]

> So the winning set is {Andre, Escher, Gore}.  Coincidentally, the same 
> as the CPO-STV result.

Yes, you could do that. However, this leads to a great incentive to rank 
the candidate you preferred second, with some nonsense candidate in 
first place. The can't-win candidate shields you from being 
downweighted. For instance:

DQ/2 + 1: X1 > A > B > C > ...
DQ/2 + 1: X2 > A > B > C > ...
(...)

If A, B, or C wins, then these two blocks (a Droop quota plus two worth) 
contributed to the winning candidate's victory, yet they're not 
downweighted.