[EM] Comment on MJ discussion (Jameson Reply)

[Michael Ossipoff]

I'd said:

Exactly. Your letter-grades encourage sub-optimal voting.

Jameson said:

"Zero-info" optimal strategy is to vote on an absolute scale such that for
recent elections you would have given equal numbers of each grade A-D and
twice that number of Fs. (Or slightly more sophisticated: give the same
score distribution as the electorate as a whole in recent elections).

1. A strategy should be more clearly expressed than that.

2. Our official public elections are most decidedly not 0-info elections.

3. Your claim about the 0-info optimal strategy needs substantiation,
verification. You're stating what you claim the 0-info strategy is,
but you need to justify your claim.

Now, I don't claim that this is a proof, but consider Weber's formula
for "strategic value" (S). (Actually strategic value was Merrill's
term,and Weber used a different term, it seems to me).

Here is candidate i's S:

Si = Pij(Ui-Uj), summed over all j.

Pij is the probability that, if your ballot can change the win between
2 candidates, they'll be i and j. Ui and Uj are the utilities of i and
j, as perceived by you.

If a candidate's S is positive, then you'd approve hir in Approval.

Of course MJ isn't Approval, but it's similar in that you're
individually helping the various candidates.

It's plausible that if a candidate's S is positive, then you want to
help that candidate. Help hir a little? If a positive S makes you want
to help hir, then it makes you want to help hir as much as possible.

So you top-rate hir.

The positive S means that you want to help hir. It doesn't mean that
you want to _maybe_ help hir.

If you rate hir less than max, then you're maybe helping hir. That
isn't what you want. So rate hir max.

So I suggest that the optimal strategy in MJ is to top-rate the
candidates whom you'd top-rate in Score--The candidates whom you'd
approve in Approval.


Jameson continued:


This
is encouraged by the grade scale

[endquote]

That's what I said: Your letter-grades encourage suboptimal voting.


Jameson continued:

, in fact more so than the equivalent
zero-info optimal strategy for Range (which is scaled Von
Neumann-Morgenstern utilities, something that behavioral experiments show
people are pretty bad at giving in any self-consistent manner).

[endquote]

My answer to that is nearly the same as my answer to your claim about
MJ strategy:



1. Our official public elections aren't 0-info

2. You'd need to substantiate your claim.

But, in this case, when we're talking about Score, rather than MJ.
More is known for sure.

It's well known that the optimal Score strategy is to top-rate the
candidates you'd approve in Approval.

In Approval, in a 0-info election, approve the above-mean candidates.

For reasons described above, when I was discussing MJ, if you have
reason to help a candidate in Score, then you have reason to help hir
all the way, by max-rating hir.


Jameson continued:

So for the
zero-info case, letter grades encourage optimal voting.

[endquote]

1. Our official public elections are not the 0-info case.

2. Your 0-info strategy claims need substantiation.

I'd said:

> So B&L have discovered that Approval fails IIAC? :-)
>
> Here's the brief, simple, precise and clear IIAC definition that i've
> heard:
>
> Removing a losing candidate from the ballots and from the election,
> and then re-counting the ballots, shouldn't change the winner.
>
> Approval and Score pass.
>
> [end of IIAC definition]
>
> It sounds as if you, and B&L, must be using a more complicated and
> wordier definition of IIAC.
>

Jameson replied:

No. It's fewer words:

 Removing a losing candidate from the election shouldn't have changed the
winner.

It's no longer mathematically precise

[endquote]

I daresay.

I asked for a complete and precise definition.

As you said, you haven't given one.

Maybe your indication of what you want could be better worded as:

If an election is conducted, and then another candidate is added to
the ballots, and the election repeated, the winner shouldn't change.

Do you think that there's a method for which the winner would never be
changed when a new candidate is added and the election repeated?

In any case, even with that better wording, it still isn't a defined
criterion. It's merely a hint at what you want for a criterion.

Of course criteria have a premise and a requirement. The criterion
that you want is clearly about what will happen if a new candidate is
added, given certain preferences and strategies on the part of voters.
My first impression is that your criterion's premise needs to
stipulate something about voters' preferences, and how they vote in
relation to their preferences.

But I advise you not to bother, because, as I asked above: Do you
think that there exists a voting system with which the winner will
never change if a new candidate is added to the ballots and the
election is repeated?

I suggest that that question can be answered without a precise and
complete criterion being written.
.
I'd said:

> A ballot that rates Favorite at
> top, and Compromise somewhat lower, could change the winner from
> Favorite to Compromise. That sort of thing won't happen with Approval
> and Score.

You replied:

With MJ, as with Approval and Score, it may happen that after the election,
you wish that you had voted otherwise. Gibbard proved that this can happen
with any non-dictatorial electoral system, so that's no surprise.

[endquote]

Nonsense.

MJ's failure of Mono-Add-Unique-Top isn't shared by all methods, and
is not an inevitable consequence of the Gibbard-Satterthwaite theorem.

You're confusing two different things.

The G-S theorem applies to Approval and Score.

Approval and Score don't fail Participation, Mono-Add-Top, or
Mono-Add-Unique-Top.

MJ fails Participation, Mono-Add-Top, and Mono-Add-Unique-Top.


Jameson said:


Yet your
expected regret is actually LOWER with an "honest" MJ vote than with any
given zero-information Approval or Score vote.

[endquote]

That's an unsupported statement.

In fact the statement's meaning isn't entirely clear.

So word it with a clearer meaning, and then substantiate it.

Jameson said:

 So as far as I can see, the
complaints about the formal valence of your vote within the mechanics of
the election system, are just words.

[endquote]

Everything that is said is words.

Substantiate your claims about strategy and
utility-expectation-maximization (if that's what you're referring to).



>
> Meaningless, because the voter doesn't know where the median is.
>

With experience and a modicum of polling, voters will be able to have high
confidence that the top two medians will fall in certain ranges.
Distinctions outside the intersection of these ranges will be safely
expressive. That is, you can safely draw a distinction between your
favorite at A and your favored frontrunner at B, or your nightmare at F and
your disfavored frontrunner at D, if you wish to.

[endquote]

For one thing, that, too, needs substantiation.

For another thing, wouldn't it be nice if voters had good predictive
information. If you need that reliable predictive information to
justify MJ, then you should reconsider that justification.
you're right here.

I'd said:

> If you do otherwise [other than rate at extremes], that's where you get "maybe".

Jameson replied:

The set of optimal votes always includes a large number of votes which are
not extreme.

[endquote]

So you claim, without any substantiation.

Jameson continued:

If you consider that votes have an expressive as well as an
instrumental purpose, then in fact the extreme vote is essentially never
the truly "optimal" one.

[endquote]

It goes without saying that, with any voting system, you can, if you
want to, cast an expressive vote instead of an instrumental vote.

So let's clarify that when I spoke of "optimal strategy", I referred
to optimal instrumental strategy. In fact, it's safe to say that
that's the understood meaning when people speak of optimal strategy.

Incidentally, it's well-established that nearly everyone wants to vote
instrumentally in our actual official public elections.

Jameson continued:

You will argue that since before the election is over you never know which
votes will be optimal, aside from the extreme ones

[endquote]

No I won't. I argue that only extreme votes are optimal, given the
lack of predictive information.

You're now using "optimal" to mean "giving the best result achievable,
given a specific configuration of other people's votes".

Because that configuration can't be predicted, that meaning for
"optimal" isn't the one that we use.

Instead, we speak of what is optimal given whatever preferences and
impressions you have before the election, at the time of voting.

Jameson said:

, that only the extreme
ones are truly optimal. This is wrong in two ways. First off, before the
election is over you never know for certain which of the extreme votes will
be optimal.

[endquote]

Again, you're using a different meaning for "optimal", different from
how we usually mean that word--As I described immediately above.

Jameson continued:

Secondly, the median is a very statistically-robust measure, so
if the bulk of voters behave in any minimally-predictable fashion, you can
set confident outer bounds on the winning and second-place medians, and
still "optimally" draw expressive distinctions outside those bounds.

[endquote]

As I said, wouldn't it be nice if we had reliable predictive information.

Anyway, why should we vote inbetween ratings, just because we hope
that they'll have the effect that we want, when we can be _sure_ to
have the effect that we want, by voting extreme ratings?

As I said, you claim that voters will be discouraged from voting
optimally because voting suboptimally might or might not achieve the
same result.



>
> We live in a technological society. Among some people, there's a
> tendency to worship science. Anything that;s more complex is felt to
> likely be better. That's MJ's mystique.
>

Bullshit.

[endquote]

An expression of your opinion.

Jameson continued:

And not a productive line of argument, even if it were true.

[endquote]

I disagree. Psychological explanation can be productive, when we're
trying to find out how (or if) people can be reached.


> It's just complicated enough that it's easy to obfuscate (for oneself)
> what's going on, and whether it's an improvement. Given the need to
> worship technology, and the consequent love for complexity, it's easy
> to be tempted to deceive oneself that MJ must be doing something
> good--even if one can't say what it is.
>
> So I suggest: If you can't clearly articulate how MJ brings
> improvement, and if you can't answer my criticisms of it, then maybe
> you could reconsider your beliefs about it.
>

If you can't move outside your own framework enough to give a better
account than the above of why my arguments appeal to me, then your
counterarguments have little chance of being convincing.

[endquote]

Ok, I was just presenting my best theory about the appeal and mystique of MJ.

You (Jameson) said:

I certainly try to
understand the internal validity of your arguments, and I think I often
succeed. If you read even just this one email, you'll see several cases
above where I grant the logic of what you're saying, and then say why I
don't think it means what you think it does.

[endquote]

As I said, it would be great news if a method as easily-counted as MJ
could really significantly alleviate the chicken dilemma.

For that reason, I am not prejudiced against MJ. It would be great if
you could show MJ is all that you say it is.

But, so far, I just haven't heard any demonstration that MJ is
anything other than a more complicated, more imperfect, copy of
Approval.   ...one that fails Mono-Add-Unique-Top.


>Third, the median satisfies majority by grade. If a majority says X has
> grade B, and B is the highest grade used, then X wins. This is a
> protection against "crankiness", as someone (I don't remember his name)
> proposing median voting for budget calculation problems said. If you use
> the mean, then someone who is very loud will get his say
> disproportionate to his number. Range voting advocates say this is an
> advantage, but in a political system, if a majority gets overruled, it
> could easily try to regain its "right" by less peaceful means.

That's the same old argument repeated. You're saying that MJ reduces
the responsiveness to extreme ratings because inbetween ratings might
or might not achieve the same effect. Sorry, but "might or might not"
won't do. MJ doesn't discourage extreme rating.

> MJ advocates seem to have a _moral_ belief about sincerity. They
> believe that it's somehow wrong, or dishonest, to vote optimally, in
> one's best interest.
>

You keep repeating that we believe this, even though I've clearly stated
before that I don't.

[endquote]

Ok. Duly noted.

So then, there is some other reason why you want to encourage
"sincere" (utlilty-proportional) ratings. Fine.


Jameson said:

There's nothing individually immoral about voting
strategically. The reason that you want a voting system which doesn't
excessively encourage strategy is that a greater proportion of "honest"
(non-extreme) votes enables a socially-better outcome on average.

[endquote]

In Score, if everyone rates sincerely, SU will be maximized. But
inbetween rating is suboptimal, unless it's called-for by uncertainty
about a candidate's deservingness of max instead of min, or by SFR.

But, even if utility-proportional rating maximizes SU in Score, that
doesn't mean that it does so in MJ.

Maybe MJ's failure to always help whom you want to help lowers its SU,
in comparison to Score. What do the simulations say about the SU
difference between Score and MJ, when everyone rates
utility-proportional. Let's guess: I bet that Score does better.



>
> Hence the MJ-ist's desire to discourage optimal voting. Of course you
> won't discourage it, even if you try.


Unsubstantiated assumption.

You think that MJ will discourage optimal (extreme) rating because
suboptimal (inbetween) rating might or might not achieve the same
result.

No.



URL: <http