[EM] RE : Re: Are proposed methods asymptotically aproaching some limit of utility?

[Abd ul-Rahman Lomax]

At 05:56 PM 3/13/2007, Kevin Venzke wrote:
> > The problem is that we have this idea of
> > exaggeration. But why would one exaggerate?
> > *Because they care.* In other words, it is not an exaggeration.
>
>You can similarly say that if I rob a bank at gunpoint, I must have
>genuinely needed the money.

This conversation is bankrupt.

Yes, if I rob a bank, at gunpoint or otherwise, I obviously have 
sufficient motivation to take the very substantial risks involved. 
What's the point?

(That the robber "genuinely needs the money' -- it would be more 
accurate, perhaps, to say that he sincerely believes he needs it -- 
is, in fact, proven by the fact of the robbery, to a reasonable 
expectation. And thus Range votes can be assumed to be sincere.

Just as someone might rob a bank on a lark, perhaps in some deranged 
condition, a Range Vote may have nothing to do with the opinions the 
voter actually holds. But we will normally assume that it does.

>I don't really mind if you want to define strategic voting out of
>existence. I don't think it sheds light on anything, though.

Just because you can't see that light doesn't mean it doesn't exist.

I have not defined strategic voting out of existence. Range, quite 
simply, does not encourage true strategic voting. Reverse rank order 
in order to gain a better outcome, *that* is strategic, insincere voting.

Range allows voters to express preference strength *as they define 
it*. I can vote pure Approval in Range; what I'm saying if I do is 
that I wish to support the candidacy of these candidates and I do not 
wish to support the candidacy of those. I say that we have become 
confused with utility analysis. In the end, voters will use the 
system to advance their own purposes -- which might include the 
welfare of society as a whole as a purpose, and it might not. We use 
utility analysis to help determine the effect of election methods on 
society, but that is not necessarily how voters will look at the method.

However, if an election method is being used to make an economic 
choice, it would be quite sensible to vote according to the estimated 
financial effect of each option. And this analysis, for overall 
benefit, could quite sensibly use simple summation. If we were 
talking "absolute Range" where true, non-normalized expectations 
could be used, summation would indeed be fully appropriate.

But, to me, Range looks like a method where I can support one set of 
candidates, fully, provide no support whatever for another set, and 
provide intermediate support for intermediate candidates. It is as if 
I am 100 voters, not one, voting 100 times in an Approval election.

(And it is interesting to look at the Majority Criterion from this 
point of view.... but I won't go there now.)

>I mean, it's trivial for me to imagine myself in a Range election with
>a variety of personal ratings for many candidates. Since I personally
>don't vote with enough uncertainty to want to undermine my own voting
>power (going to the polls is enough of an inconvenience), I would vote
>approval-style. And here you're basically saying you have enough
>confidence in me, some random voter, to trust that I must truly care
>deeply about this separation of the candidates into two sets.

Yes. I think that if you vote Approval style, you are dividing the 
candidates into two groups, and you are willing to support one group, 
fully, and not the other. It is true that this might not reflect much 
care, it might be simple disinterest, insufficient to go to the 
effort of rating candidates intermediately.

I'm utterly unoffended that you might vote this way. You might, under 
present conditions, use standard Approval "strategy." That is, vote 
for your favorite of the top two, and then for any candidates you 
equally approve or more approve. That this is called "strategy" 
points out the problem. It is not insincere.

Range allows voters *freedom* to vote intermediate votes. It does not 
require them to do so. If few take advantage of it, it may be argued 
that the system is too complicated, and that the range should be 
reduced. I'd approach it from the other direction: I'd start with 
Approval, which is terminally simple, and then add intermediate 
ratings, starting with one (probably the intuitively appealing range 
of -1, 0, +1). If enough voters want more resolution, we can go on up.

What I've become opposed to, even though I often think of ways to do 
it, is setting up a system that allows the expression of a preference 
with no rating difference. In the end, I'm thinking of Range, 
overall, as a system that maximizes voter satisfaction (another way 
of looking at utility analysis).

Many of us look at elections from the point of view of how partisan 
voters will behave. But there are large numbers of voters who are not 
particularly partisan. They, obviously, vote Democratic in one 
election and Republican in another. Among others, these are voters 
who will use intermediate ratings, and their behavior will be a heavy 
factor in who wins.