[EM] Juho , 5/21/12, roughly 0800 UT

[Michael Ossipoff]

I'd said:

Tricky isn’t the word for it. Try “unknown”, for most typical situations in
Condorcet.

But I told the expectation-maximizing strategy for a u/a election in which
Compromise is the only acceptable perceived to be able to beat Worse. No,
I’m not going to repeat it for you.

Juho says:


Since you don't want to point out any such strategy

[endquote]

I've only described it many times already, in replies to you. I'm not going
to waste any more time repeating it for you.

Juho says:

 I assume there is no such single strategy that you want to recommend for
all regular voters.

[endquote]

To maximize expection with Condocet in a u/a election where Compromise is
the only acceptable able to beat the unacceptables, rank Compromise alone
in 1st place.

I don't know what you mean by "all regular voters". The above strategy is
for voters who perceive the above-desecribed conditions.
If you want a general strategy for Condorcet, none is known.

Juho says:

 Actually any working general strategy that regular voters or parties can
use will do, either for Approval or Condorcet, either expectation-maximizing
or other strategy that is expected to improve the outcome. But so far I
have thus not seen any (lots of u/a discussion though but no general
strategy for all situations).

[endquote]

No general strategy is known for Condordet. The general
expectation-maximizing strategy of Approval is the better-than-expectation
strategy.

Juho says:

I got the two numeric examples though. Thanks for them. At least one of
them didn't however seem to work well enough, so I assume that the theory
and strategy behind those examples is not quite well thought yet.

[endquote]

As usual you're vague. I have no idea what examples you're referring to. If
you want to say that one of them isn't good enough, then you need to
clearly specify it, and then tell what's wrong with it, and why you think
so.



But I fully admit that expectation-maximizing strategy isn’t known for
Condorcet in most typical situations. It isn’t that Condorcet doesn’t need
strategy. It’s just that you don’t know what the expectation-maximizing
strategy is. But don’t feel bad—no one else does either.




In theory there are cases where one could cheat the system. But in practice
sincerity is by far the best strategy that voters have in large elections
where voters make independent decisions. The challenge is to find practical
situations where regular voters, after hearing some poll results (and
possibly some poll based situation specific strategic advices by the
media), would have good reason to vote otherwise (in a way that they can
master an that is likely to improve the outcome).

[endquote]

I described one to you, and no, I’m not going to repeat it. I don’t have
time to keep repeating things for you.


Sorry, I don't recall, or maybe didn't identify the the general solution
when I read it. You might point that out to help my poor memory. :-)

[endquote]

No, sorry, I don't have time for that. I've repeated it enough already.  I
refer you to my previous replies.



I’ve abundantly and sufficiently discussed that. It’s time to just agree to
disagree.

Ok. My statement is pretty well covered above. My assumption is that you
wll stick to the claim that there are often working strategies in Condorcet

[endquote]

…unknown ones, yes. But I described a strategy for a particular situation,
showing that sometimes favorite-burial is optimal in Condorcet.


If you describe a theoretical vulnerability or a practical strategy for
some specific situation, and the strategy is intended for regular people or
parties, then it can be made a general and usable strategy by adding some
criteria that tell the voters when that strategy should be applied.

[endquote]

Fine. Then do so. Write a general strategy for Condorcet :-)


Juho says:

That would make the strategy a working strategy (although not necessarily a
strategy that would work often).

[endquote]

Then it wouldn't be a general strategy, would it.





, and voters would be foolish toassume that they can generally vote
sincerely.

[endquote]

Correct.







Two: In Approval, if you like strategy, I’ve given simple instructions for
determining the way of voting that maximizes your expectation. I’ve
described it for u/a elections,
………..and for non-u/a elections.

I'd be interested in the one (or ones) that the regular voters are supposed
to follow in real life Approvan elections.

[endquote]

Any one of them that they like, or any one of them that makes use of the
information, perception or feel possessed by the voter. Yes, in real life
Approval elections.

That was not an answer. A concrete strategy please.

[endquote]

No, actually that is an answer. I’ll repeat it again for you: Look at the
last part of my Approval article.  If you feel that some part of the
strategy suggestions there are insufficiently “concrete”, then tell me what
it is that you need more details about. Ask a specific question.


Ok, but which Approval article?

[endquote]

The one that I posted to EM. The one that is at Democracy Chronicles.






You said:

You mentioned also sincere approval of "approvable" canddates as a strategy
that could be recommended to the voters. Do you think Approval can handle
well situations where some voters or voter groups are strategic while some
are sincere?

[endquote]

It’s easy to show strategizers taking advantage of sincere suckers in Range
or Majority Judgment. Maybe you’re saying that you fear that if you approve
the candidates you like, then the supporters of one of them will take
advantage of you by approving the candidates that they perceive as
acceptable, above-mean, or better-than-expectation. Sorry, but I don’t see
it. If you think that there’s a problem there, then you need to explain
what and why.

Simple example:
2: A>B>>C
1: B>>A>C

If the first group of voters approves sincerely A and B, B will win. Ins't
this a good enough reason for the first group to vote strategically and
place the approval cutoff between the two potential winners? That is, if
they want to win and not just elect the most approved candidate.

[endquote]

If you’ve approved some candidates whom you like, and one of them wins,
then I guess that I’m not quite understanding what your problem is.

If, on the other hand, you prefer to vote strategically, then of course you
should do so.

Juho says:

The problem is that we usually talk about competitive political elections
where all the players want to win.

[endquote]

You need to clarify, with yourself, what you mean, what you want. Do you
mean that you only want your favorite to win? Then, in Approval, approve
hir only.

If you want to maximize your expection, I've told Approval strategy for
that purpose.

Juho says:

Few lines above you assumed that people would use their burying
possibilities to the maximum.

[endquote]

Available evidence indicates that many voters would do whatever it takes to
maximally help Democrat against Republican.

Juho says:

I don't understand why in this case voters would be indifferent with
respect to the outcome of the election.

[endquote]

If you're asking about favorite-burial need in Approval, there is none.
Those same voters wouldn't favorite-bury in Approval because it is
transparently obvious that there can be no reason to do so.

But if you're questioning the assumption that people wouldn't strategize in
Approval, I merely suggest voting for all whom you like. If you want to,
you can strategize. Suit yourself.

Juho says:

Maybe you recommend Approval as a good method for non-competitive elections.

[endquote]

...and for competitive elections.







As for the defection problem, we’ve discussed it before, and the fact that
Approval has ways of dealing with it, and the fact that Condorcet fully has
that  problem too.

Disagreed. I don't know how Approval can handle it.

[endquote]

I posted some solutions some months ago.  I’ll find that posting and
re-post it.


Thanks.


But my suggestions included Forest’s solution in which the A voters give to
B only enough approvals such that if C’s favoriteness-percentage is as
estimated, then the larger of {A,B} will win. In Approval that would be
done probabilistically. The A voters tell the B voters that they should do
the same, if they don’t want C to win, and if the A faction might be bigger
than the B faction. No, not perfect, due to imperfect predictive
information, but still helpful.


Without exactly knowing what the strategy is, I note that in practical
elections there may be some problems with making the voters vote in line
with the strategy.

[endquote]

The strategy is as described above. Any difficulty in following it would be
likewise encountered when it is needed in Condorcet.


One more thing in my mind. Regular voters may be interested only in the
outcome of this election and never mind if their opponents get angry.

[endquote]

Fine, with Tit-For-Tat in use, the B voters will keep defecting, and so
will the A voters. But the B voters will know that as soon as the
co-operate, so will the A voters.

Anyway, Forest's suggested solution has its effect in the current election.
The non-secretness of a faction's voting intentions is also relevent to the
current election.


This claim was just a reflection of your idea that Approval could "maximize
expectation".

[endquote]

Nonsense. We haven’t been speaking of some methods that maximize
expectation. We’ve been speaking of expectation-maximizing strategies.

It’s been established and agreed on EM that the strategies that I’ve
described do indeed maximize expectation in Approval.


"agreed on EM" :-)

[endquote]

Correct. And many special cases of, implementations of,  the
better-than-expectation strategy have been well-known for a long time in
the broader voting system discussion.

Juho says:

Are there multiple such strategies?

[endquote]

As I've already repeated for you many time, there is one
expectation-maximizing strategy for Approval: Vote for the candidates who
are better than expectation. That strategy can be implemented directly, or
via various implementations that are special cases of it.

This time, pay attention, because I'm not going to repeat that for you
again.


Juho says:

 Does that mean that each voter can pick his favourite, and they work well
in sync that way?

[endquote]

Yes to both questions. Pick whichever implementation you like or have the
information or intuitive feel for. Remember that they're just special cases
of the same strategy.






Maybe better to focus on concrerte practcal strategic vulnerabilities
(unless there is something more in this).

[endquote]

So focus on it then, instead of just making a vague reference to it.


What should I provide? I'm willing to be more concrete if you tell me what
you want.

[endquote]

In general, what you should provide is the specifics of what you mean. You
never do that, and no doubt you never will. That's why talking to you is a
waste of time.

In particular, in this instance, you speak of focusing on concrete
practical strategic vulnerabilities. I suggested that you specify and focus
on one.

Juho says:

I'd like you to point out concrete and practical strategies for real life
elections for both or either of the methods.

[endquote]

I've written this sentence many times for you: I've pointed out a practical
expectation-maximizing strategy for certain conditions in Condorcet
elections. I've pointed out that no general strategy is known for Condorcet.

I've suggested concrete and practical strategies for real life elections
with Approval. I've told you where to find them. I won't repeat it for you
again. I refer you to my previous replies, and to my Approval article. The
one that I posted to EM fairly recently. The one that is at Democracy
Chronicles.





And in addition you say that any (your listed) strategy is ok. Why so? Is
there no good working strategy that all could use?

[endquote]

Yes there is. The better-than-expectation strategy. As I’ve said here many
times, all of the Approval expectation-maximizing strategies for Approval
are special cases of better-than-expectation.


I don't have a full description of that strategy yet.

[endquote]

I've repeated it for you many times, and told you where to find it. I've
helped you all I can. You're wasting my time.




The direct implementation of better-than-expectation could just consist of
approving the candidates who are better than (or maybe exactly as good as)
the result-merit that you expect from the election.

Juho says:

Is this the definition of the better-than-expectation strategy for regular
voters?

[endquote]

It's for any voter, regular or otherwise, who wants to vote strategically
in Approval.

Juho says:

Does "or maybe" mean random selection or selection based on what the voter
feels like that day?

[endquote]

I said "(or maybe exactly as good as)" because approving a candidate who is
exactly as good as your expectation doesn't affect your expectatation. It
doesn't matter whether or not you approve a candidate who is exactly as
good as your expectation. You can flip a coin, or yes, go by how you feel
that day.

Juho says:

Does "result-merit that you expect" mean the value of the (single) guessed
winner or maybe the weighted average of potential winners?

[endquote]

Answer to both questions: Yes, if that's what you feel that you know, or
have a perception or feel about.

People are not going to determine their expectation in the election by
multiplying the win probability of each candidate by hir utility, and
summing the products. But you can do that if you want to.


But you know how good a result you expect from the election.


Juho says:

Does the expectation refer to the sincere opinions or does it include the
expected strategic voting too (much more complicated and cyclic)?

[endquote]

...if you want it to.

Do as complicated and elaborate an analysis as you want to, to try to
calculate your expectation. It isn't for me to tell you if or how to
calculate your expectation, if you want to calculate it. That's your
business.

But, as I said above, you know how good a result you expect from the
election.

Or, if you really don't, then just use the 0-info Approval strategy of
approving the above-mean candidates.

Mike Ossipoff
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20120521/7cf3c3ae/attachment-0003.htm>