[EM] Re: Rewording Strategy A (BF(1st))

[MIKE OSSIPOFF]

Russ said:

Perhaps some of the folks here missed my point about Approval strategy.
Well, at least one person has, but he is a fruitcake.

I reply:

Great that Russ has risen above his ranting :-)

But that's still unoriginal, since I'd used  "fruitcake", to refer to Russ's 
anti-evolution arguments.

Russ continues:

Two different questions can be asked about any voting system:

1. What is the best voting strategy?
2. How effective is the best voting strategy?

I reply:

Russ's question, from the 1st time he's asked it, and the 3 or 4 subsequent 
times he's asked it, has been about the effectiveness. That his question was 
about that has never been doubted.

Russ continues:

Let's apply these questions to plurality as an example. The strategy is
simple to formulate, but how effective is the strategy? It is very
effective if you prefer one of the two major parties, but otherwise it
is not very effective.

I reply:

"Prefer" to what? Sometimes "prefer" is used just to mean "to like", and 
maybe that's what Russ means.

When you believe that a certain 2 parties will be the frontrunners, then 
vote for whichever of those you prefer to the other. That is effective, for 
the purpose of maximizing your expectation in the election.

Or maybe Russ is saying that you don't prefer either of the 2 likely 
frontrunners to the other. If you believe that those will be the top two, so 
that any tie or near-tie will be between them, and if you have no preference 
between them, then truly you have no effective way to vote--based on those 
assumptions about your predictions and preferences. But that's hardly the 
fault of any strategy.

Russ continues:

You've stated what is perhaps the best strategy for Approval, but how
effective is that strategy? Well, in some cases it may be very
effective, but in other cases it may not be.

For example, what if *three* parties are equally popular? Let's take the
classic Democrat (D), Republican (R), and Green (G) case. Suppose they
are approximately equally likely to win. And suppose your own order of
perference is G>D>R. Who do you approve?

Well, you could play it safe and approve both G and D. But then you will
be disappointed in your choice if it turns out that G could have won had
more G>D>R voters not approved D.

Then again, you could be bold and only approve G. But then you will be
even more disappointed if R wins but D had a good chance if only more
G>D>R voters had approved D.

Do you see the dilemma here?

I reply:

Yes, Russ, it's the basic strategy situation of Approval, and we've been 
discussing it here quite a bit.

  This is unbelievable. Any  capable newcomer would have caught on by this 
time.  Ok, everyone, Russ has announced his discovery that, in Approval, the 
G voters can regret voting for D, if G could have  otherwise won--or regret 
not voting for D, if D needed their vote to beat R.  Gads, that changes 
everything we were saying! Now we have to rewrite everything :-)

Delusions of grandeur.

Russ continues:

The problem is not that you don't have a
formula to guide your strategy.

I reply:

But there is a "formula" for that situation: Vote for the above-mean 
candidates.

Russ continues:

The problem is more *fundamental* than
that. The problem is that *no* formula can help. Your best strategy is
more or less a coin toss.

I reply:

For one thing, in those cases where a random choice is really called for, 
then that random choice could be considered a "formula'. Or, some formulas 
could result in a random choice, conceivably. But though there are ways to 
use  a coin, , in an approximate way, to choose from among 3 choices, 
throwing a die, or drawing one of 3 marked pieces of paper from a bag might 
be more convenient.

  But, typically, the fact you don't know anything about winnabilities 
doesn't mean that "*no* formula can help". It it's a 0-info election, then 
vote for the above-mean candidates.

Russ continues:

Incidentally, Approval strategy seems to be more sensitive to polling
results more than most other methods. In plurality, pre-election polling
results might affect whether you bother to vote at all, but they are not
likely to swing your vote one way or the other.

I reply:

It often will, if you believe the poll. If you use the poll to determine who 
the frontrunners will be, then you'll  vote for whichever of those two you 
prefer to the other. That uses the polling information, and the polling 
information influences your vote, if you believe the poll.

\But, if you have no information that some candidates are more winnable than 
others, or that some candidates are more likely to be in ties or near-ties, 
or that particular candidates will be the 2 fdrontrunners, then your best 
expectation-maximizing strategy is to just vote for your favorite.

Russ continues:

The same is not true of
Approval. In Approval, your strategy is likely to depend strongly on
pre-election polling results, particularly if several candidates are
competitive.

I reply:

Russ has done it again! Will his astounding new discoveries never end?

Russ continues:

So what, you ask? Well, has it occurred to anyone that
pre-election polls can easily be rigged?

I reply:

:-) Actually yes, it's been discussed here.

  But, if polls can be easily rigged, why did you say that your vote in 
Approval would be determined by polls? I mean, why base your vote on them 
when they're so easily rigged? (Rhetorical question)

Summary:

Russ, this is what I meant when I said that you're getting your elementary 
education from EM. But I was mistaken to say that, because it's  becoming 
that you haven't learned anything. . You won't gain anything by asking 
questions unless you listen to the answers before continuing to expound. So 
far, you've basically been merely continuing to repeat what you were saying 
when you first began to expound on Approval strategy.

Mike Ossipoff

_________________________________________________________________
Express yourself instantly with MSN Messenger! Download today - it's FREE! 
http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/