[EM] Jameson: MJ, optimal voting, Strong IIAC

[Michael Ossipoff]

Jameson:

> But the criterion's premise stipulates optimal voting. Voting to
> maximize one's utility-expectation. That's extreme voting.

Unproven assertion. One which I believe is based on sound logic but faulty
assumptions, and is therefore false.

[endquote]

Sure, the matter of what way of voting is optimal in MJ must come
before any comparisons of frequency of Strong IIAC failure.

You said that you'd rather stipulate realistic voting instead of
optimal voting, in Strong IIAC's premise. But optimal voting, voting
to maximize one's utility-expectation, is simply-defined. Defining
realistic voting would be prohibitively difficult  and subjective, for
a criterion. Besides, realistically, people _do_ vote to maximize
their utility-expectation, based on their assumptions and beliefs.
Everything said by voters confirms that.

Anyway, certainly, as you said, the first question is the question
about what kind of voting is optimal in MJ. Until that is established,
it would be premature to discuss how often the methods fail Strong
IIAC.

Are you sure that it's better for a method to less frequently violate
Strong FBC? Why? ordinary IIAC, as I defined it, makes obvious sense,
in terms of consistency and responsiveness. Even though it isn't
crucially important, it still makes more sense to pass it than to fail
it. But why would it be better for a method to less often fail Strong
IIAC? Why shouldn't people make different choices, among a different
candidate-set?


You said:

I would prefer to use "realistic voting" rather than "optimal voting" for
this criterion.

[endquote]

Above, I told why that would be problematic.

You said:

However, I believe that you are wrong for both [realistic & optimal]

[endquote]

One nice thing about u/a elections is that the strategy is a lot
simpler to determine. There just can't be any doubt, can there, about
the u/a optimality of extreme-rating in MJ?

I think that MJ's optimal strategy is the same as that of Score, u/a
or not, but I don't have a rigorous proof. MJ's non-u/a strategy might
have to remain unresolved, unless someone can supply a proof. In the
meantime, its u/a strategy is obvious and uncontroversial.

And, if MJ's advocates and its opponents can't prove what MJ's
non-u/a, non-0-info strategy is--Doesn't that count against MJ as a
proposal for our official public elections if you don't believe that
our official elections are u/a?

So there are two possibilities: Either MJ's strategy is the same as
that of Score, or it's unproven _what_ it is.

Either way, are you sure you want to propose MJ?



...We certainly don't have 0-info elections, as I said earlier.
> In fact, we have non-0-info u/a elections,


Unproven assertion. I believe that for over half the electorate, the
information limits are more salient than the U/A aspects.

[endquote]

Nearly all voters firmly believe that the winner will always be a
Democrat or a Republican, and that it can't be otherwise, no matter
how they vote (because, in their belief, nearly all the other voters
will always vote only for Dem or Repub).

There is nothing 0-info about that. Information needn't be valid.
False information is still information. There is no lack of definite
winnability-information. The virtually universal belief is that only
the Democrats and Republican are winnable. No 0-info there.

Proof? Ask anyone. You yourself vote Democrat because you think that
only the Dem or Repub can win. I don't, but nearly all Dem>Repub
voters do, even if Dem isn't their favorite.

(Note that
earlier I said that absolute rating was the optimal strategy in the 0-info
limit,

[endquote]

That is a class 2 claim--Unlikely, and wouldn't help MJ even if it
were true (because our elections aren't 0-info).

You said:

*and* that it continued to be an optimal strategy with limited
information for twice as long as for score or probabilistic approval. I
believe that for the majority of voters in the majority of real elections
it will still be optimal.)

[endquote]

A class 3 claim. Unlikely to be true, but would help MJ's value if it
were true. (Because if sincere valuation were sufficient strategy,
then strategy would be simplified).


> Suppose that some set of voters prefer X to Y, and Y to Z. But their
> utility difference for X vs Y is very, very small in comparison to
> their utility difference for X & Y vs Z. Their optimal strategy in MJ
> is to top-rate X and Y, and bottom rate Z.


That depends on their expectations for the medians of X, Y, and Z. In
particular, if they expect at least two of those medians to be below the
second-to-top grade, and the strength of that expectation is greater than
the ratio of the expected instrumental utility of voting to the utility of
an expressive vote (which is almost certain, because the instrumental
utility of voting is infinitesimal)...

[enquote]

Whoa! When we speak of optimal voting, we're talking about
instrumental voting. That's fair, because it's very well established
that nearly all voters vote instrumentally. Again, ask anyone.

And though your vote, as an individual, isn't important, it's still
true that if you belong to a sufficiently large set of voters who
believe and voter similarly, then how you and your set vote can and
probably does affect the election result. So vote Green. What if
everyone more progressive than the SleazeDemocrats voted for what they
want? The problem is that you don't think that the others of your
preference-set will do other than compromise. Are we all giving it
away to a corrupt "compromise" because we're following eachother?



You continued:

, then they will optimally vote Y at
second-to-top.

[endquote]

See above. Nearly all voters vote instrumentally.

You continued:

You can even build a quantal-response-type model in which
this is instrumentally optimal.

[endquote]

Build one.

> Now Z withdraws. Now there
> are only two candidates. Those voters' optimal strategy is now to
> top-rate X and bottom rate Y. If that set of voters is large enough,
> that could change the winner from Y to X.
>

That example works at least as well for Approval or Score; in fact, better,
because neither objection above (expressivity or quantal-response) applies
to either of those.

[endquote]

I don't deny that Approval and Score fail Strong IIAC. I don't know
why Strong IIAC should be passed. I haven't heard any demonstration
that MJ will pass it more often.

But I agree with you that Strong-IIAC-failure-frequency is a premature
topic until we establish what MJ's optimal strategy is.

>
> Why would MJ fail Strong IIAC less often than would Approval and Score?
>
> In particular, in our non-0-info u/a elections?
>

I don't accept this assumption. Obviously I realize that elections are
non-0-info, but I believe they are in practice closer to 0-info than they
are to u/a for most voters in most elections.

[endquote]

See above. Ask anyone--Nearly everyone votes instrumentally.

As for u/a, that's harder to demonstrate. Because nearly everyone is
entirely sure that the winner must always be Dem or Repub, then Repub
wouldn't have to be u/a unacceptable in order for thus-believing
Green-preferrers to vote Dem.

For _me_ the elections are certainly u/a. Maybe, for anyone believing
the media, anything other than a Republocrat would be completely
unacceptable, just because the tv says so.


Mike Ossipoff