[EM] Approval vs Condorcet

[robert bristow-johnson]

> On 03/22/2024 12:58 AM EDT Michael Ossipoff <email9648742 at gmail.com> wrote:
> 
> 
> Robert—
> 
...
> To most everyone, OPOV is synonymous with the Plurality, Vote-For-1 (VF1), voting-system.
> 
> …& yes I admit that, by that definition, Approval indeed violates OPOV.
> 
> …& I admit that Condorcet violates it too.
> 

Certainly FPTP and IRV and Condorcet count people differently, but what they're doing is counting people, some group of people against some other group of people and saying that, if every person's vote counts equally, then the candidate that has more persons voting for him/her wins the election.

Approval voting is counting marks.  As if marks have equal rights.

There is a reason I began, in my paper, with this very old ruling from the North Dakota Supreme Court:

"The theory of cumulative voting... rests upon a false or fictitious premise. It assumes that the computation of the number of marks placed upon a ballot in favor of a candidate should determine whether he is elected, when in fact the marks are, and can only be, representative of persons possessing certain qualifications [citizens having franchise]. ...
The placing of marks upon the ballot is only a method of enumerating
persons, ...  Our system of government is based upon the doctrine that the majority rules. This does not mean a majority of marks but a majority of persons possessing the necessary qualifications and the number of such persons is ascertained by means of an election ..."

This is really trimmed to show just the bare essential.

Now, measuring Approval on an Approval ballot is, in a degenerate case, essentially the same measure as measuring Approval on a Score ballot, where the degree of Approval is limited to two levels.

Now, what the Ordinal methods do are, instead of asking the electorate about how much they Approve a candidate, they ask "Do you *prefer* this candidate over this other candidate?"  We voters are partisans, not judges.  We cannot be nor should be expected to be measuring (and awarding) worth of a candidate.  We're trying to get some candidate elected and not elect some other candidate.  That selfish motivation is our motivation as partisan voters.  And we have that right, along with the secret ballot.  The thing that keeps this fight fair is that our votes count equally.  (Or *should* count equally.) So...

> > On Sat, Mar 23, 2024 at 18:27 Michael Ossipoff <email9648742 at gmail.com> wrote:
> > > Approval maximizes the number of people pleasantly-surprised, &/or the number who get something they like.
> > > 

If the stakes are high and you Approve both A and B (but you like A better) and then you find out that the top-two Approval candidates are A and B, but B wins, you might not be very happy that you Approved B.  Especially if it was close.

So Approval, like FPTP, loses information about how voters would vote in different contingencies.  But the ranked ballot preserves that information, but it doesn't ask, nor should we care, what degree of preference a voter has for one candidate over another.  It should not matter.  If I enthusiastically prefer Candidate A and you prefer Candidate B only tepidly, your vote for B should count just as much as my vote for A.  We all get it right when there is only A and B in the race.  But when there is C in the race, the Ordinal ballot doesn't force the voter to make tactical considerations like a Cardinal ballot inherently requires.  (But bad tallying methods can make voters regret their vote.)

Then, to preserve the equality of our votes, that each person's vote has equal effect at getting their candidate (or their contingency candidate) elected is that at the end of the day, if more voters mark their ballots *preferring* Candidate A to Candidate B (this information can be lost in Approval), then if Candidate B is elected, that would mean that the fewer voters that marked their ballots preferring B had cast votes that had more effect, that counted more, than the votes coming from the larger number of voters marking their ballots preferring A.

Condorcet violates One-person-one-vote *only* if there isn't a Condorcet winner.  Then Arrow and Gibbard prevail and, no matter what method is used, the election is spoiled.  There is a candidate that can be removed which will result in changing who would be elected.  Can't be avoided but we should do the best that we can do in that rare pathological circumstance.

My opinion is that when no Condorcet winner exists, the easiest sell to the public (and legislature that might enact such law) is either just the Plurality (of first-choice votes) winner or perhaps the runoff winner between the top two first-choice vote getters.

To disincentivize sophisticated strategic voting, probably Schulze or Ranked Pairs is best.  I prefer margins over winning-votes.

--

r b-j . _ . _ . _ . _ rbj at audioimagination.com

"Imagination is more important than knowledge."

.
.
.