[EM] Why the concept of "sincere" votes in Range is flawed.

[Abd ul-Rahman Lomax]

At 12:34 AM 11/26/2008, Juho Laatu wrote:
>One approach to sincerity is to compare voter 
>behaviour to the requested behaviour. In 
>Approval if the request is to mark all 
>candidates that one approves then placing the 
>cutoff between two main candidates is often 
>insincere.. Approval is a special method from 
>this point of view since it is often described 
>as requiring the voter to plan what is the best 
>strategic vote (where to put the cutoff).

It requires no such thing. Voters, however, will 
maximize their expected outcome if they vote 
optimally. And they vote optimally by making a 
sincere expression, and this is it: if they 
prefer any candidate to the expected election 
outcome, vote for that candidate. If not, don't. 
This is a sincere preference expression.

It certainly is not an expression of "approval," 
hence I have often stated that ballot 
instructions for voters should not use the word 
"Approve." The instructions *might* use the word 
"accept," but even that is doubtful. Voters 
aren't stupid. A statement could be like, "This 
election will pick one winner. You may vote for 
as many candidates as you choose to support, 
including a write-in candidate. The winner will 
be the candidate who receives the most votes."

Telling the voters how to vote is offensive. 
However, I'd presume that if Approval is 
implemented, there would be plenty of advice out 
there. Some of it might actually be good.

What is the "expected election outcome"? It's the 
average, approximately, of the frontrunners. 
I.e., one might assume that the election is a 
tossup between the two frontrunners. Literally, 
it's a tie. If you don't vote for one of them, or 
you vote for both, your vote wil not affect the 
result, it will be moot. If you vote for one of 
them, only, your vote will count. Vote as if you 
vote will count, for the candidate you sincerely 
prefer to that lottery. And vote the same way for all candidates.

Simple. Sincere. Actually applying it with 
precision, of course, difficult. But it's just 
one vote that the voter must decide, which has 
very little chance of being a deciding vote. 
Another way of putting it is, don't vote so that 
you will seriously regret your vote. If you 
bullet vote, and it turns out that if you'd have 
approved another candidate, that candidate would 
have won over a total jerk whom you detest, 
you'll regret your vote. If you vote for two, and 
your favorite loses, but the other one you voted 
for is acceptable, that's not much of a cause for 
regret, even if the race was close and you might think your vote decided it.

A single vote, in Approval, never, by itself, 
decides an election. Instead, it decides between 
a lottery and a election. That is, it converts a 
tie between two candidates into an election of 
one, or the election of the other into a tie. 
Therefore the possible regret from a vote is 
reduced to half the utility difference between the candidates.

It gets simpler if a majority is required, which 
is one reason why such requirements are 
brilliant. Did I mention that requiring a 
majority for any decision is basic to democratic 
process, and we only abandon it for practical 
reasons having to do with the difficulty of large-scale eletions?

When a majority is required, the question one 
might be answering with an Approval vote is "Do 
you prefer this candidate to having the election 
fail, necessitating further process?" If so, vote 
for the candidate. If not, don't. The answer to 
this question is dependent on expectations of 
what would happen if the there is majority 
failure. Majority failure can seriously improve 
results if it leads to a good runoff process. 
Might be a new nominee! Or the field is narrowed 
and one's favorite might have a better chance. 
With a majority requirement, one can more safely 
bullet vote, i.e., express a single preferred 
candidate. But, of course, if one would prefer to 
vote for an additional candidate to avoid the 
need to go to a runoff, then one can do so.

In any case, the vote, seen in this way, is a 
sincere expression of preference, and not voting 
sincerely doesn't produce any advantage for the voter.

It's important to realize that these expressions 
of sincere vote in Approval are also 
strategically maximal. They produce optimal 
expectation, there is no more powerful way to 
vote. The only way to improve them is to have 
better information about the election 
probabilities, i.e., to be better informed. Let's 
see -- Approval gives more power to voters who 
are better informed. Aha! I knew it! The method 
is elitist! Those mathematicians and political 
scientist who promoted Approval Voting, they just 
want to give an advantage to educated people, who 
are well-known to be liberal elitists, just like 
those biased newspaper reporters, the ones who 
actually meet and talk with the candidates, who 
want to lord it over all the rest of us common 
people with their superior knowledge. Down with 
all of them! To the barricades! Send them to the collective farms!

Ahem. Range and Approval reward sincere voting, 
if sincere is understood. The "insincere voting" 
that they allegedly reward is based on 
assumptions of what the votes mean that are 
inaccurate and essentially baseless. Certainly a 
vote under Approval doesn't mean that the voter 
"approves" of the candiate, in the ordinary 
sense. It's an unfortunate name for the method, 
for this reason. Count All the Votes is what I 
like to call it. Open Voting might be a nice name, what do you think?

Open Voting.

Yes, I like it.



>Another interesting case is Range. If one 
>requests the voters to mark their utilities then 
>sincere voting might be rare since strategic 
>exaggeration is practically always available and 
>an efficient strategy. One can describe Range 
>also as "Approval with ability to cast weak 
>votes". In this case it could be as sincere as Approval.
>
>I think these two Range description styles 
>actually refer to two different methods although 
>the voting and vote counting procedures are technically the same.
>
>Also possible request to normalize one's vote 
>vs. to just mark sincere utilities makes a difference in Range.
>
>In summary, Approval is close to being a method 
>where voters are expected to vote strategically, 
>and therefore voters need not be insincere when 
>doing so. In most methods (and environments) any 
>deviation from one's sincere opinion in order to 
>win in the election should however be classified 
>as insincerity (and as unwanted strategic voting).
>
>Juho
>
>
>
>--- On Wed, 26/11/08, Kevin Venzke <stepjak at yahoo.fr> wrote:
>
> > From: Kevin Venzke <stepjak at yahoo.fr>
> > Subject: Re: [EM] Why the concept of "sincere" votes in Range is flawed.
> > To: election-methods at electorama.com
> > Date: Wednesday, 26 November, 2008, 6:45 AM
> > Hello,
> >
> > --- En date de : Mar 25.11.08, Abd ul-Rahman Lomax
> > <abd at lomaxdesign.com> a écrit :
> > > What Approval sincerely represents from a voter is a
> > > *decision* as to where to place an Approval cutoff.
> >
> > But is it not true that what *all* methods sincerely
> > represent from a
> > voter are the decisions related to voting under that
> > method?
> >
> > If a decision makes sense in a given context, then that is
> > a sincere
> > decision. Is that not your stance?
> >
> > Kevin Venzke
> >
> >
> >
> > ----
> > Election-Methods mailing list - see
> > http://electorama.com/em for list info
>
>
>
>
>----
>Election-Methods mailing list - see http://electorama.com/em for list info