Disapproval//Condorcet? (was Re: "votes against" in pair-defeats

[Steve Eppley]

Demorep1 wrote:

>I again suggest that a separate majority approval/disapproval vote
>should be done first.  Namely in the 40-25-35 example the supporters
>of C and N would presumably defeat D 60-40 (assuming that the C and
>N supporters would not vote to defeat each other). C would then beat
>N 65-35.

What do you mean by "a separate ... vote ... first?"  Do you mean
separate balloting where the voters will know which candidates are
disapproved (eliminated) before they cast their ranked ballots, or
do you mean that the tally would begin by looking at the disapproval
portion of the ranked+disapproval ballots?  

The latter would be more efficient, but the two ways might be voted
differently.

* *

With a good voting system the concern about fratricide will be gone,
so enough good candidates will be able to run that Disapproval
shouldn't be needed.

C is already the winner without the Disapproval step if Condorcet is
the method.  So what has been gained by including the step?

* *

There's a possible downside to using a disapproval filter.  While
it's true that the C and N supporters would likely disapprove D,
it's also possible that some of the C supporters would disapprove N
and some of the N supporters would disapprove C.  Combining these 
disapprovals with those cast by the D supporters, maybe all the 
candidates will be disapproved.

Here's an extreme example illustrating that:
  40: D > NOTB > C=N
  25: C > NOTB > D=N
  35: N > NOTB > C > D
  Every candidate is majority-disapproved, creating a power vacuum.

So then we'd be left wondering how many of those disapproval votes
were cast for tactical or capricious reasons, and whether the power
vacuum is better for society than electing C. 

---Steve     (Steve Eppley    seppley at alumni.caltech.edu)