[EM] 3 or more choices-Condorcet

[Michael Ossipoff]

Regarding the Margins vs wv question:

First, I don't like either, for the reasons that I've given. But, if
you want to try to ignore the chicken dilemma, then under some
conditions, TUC(margins) could be argued for.

I spoke of a time when the public have just elected a Green
government. I pointed out that, if they could do that, then they
obviously can make good use of FBC-failing methods. For that reason,
IRV's FBC failure becomes forgivable. Likewise, then, the FBC failure
of TUC. But I've said that there's no excuse for a rank method to fail
CD, and I stand by that.

But suppose you want to disregard that. Then, in that future Green
government scenario, there'd be a case for TUC(margins).

I should qualify my forgiveness of FBC-failure, with a
Greens-electing, Plurality-competent, electorate: Though FBC wouldn't
be _needed_ under those conditions, it would still always be
desirable. That's why, in my better-electorate ranking of voting
systems, I ranked IRV below the other methods that I approved.

Anyway, what would be the advantage of TUC(margins)? The combination
of the Condorcet Criterion and 0-info Sincerity. I could be mistaken,
but it seems to me that TUC meets 0-info Sincerity.

If we're supposing an electorate that has just elected a Green
government, using Plurality, then the Democrat-disinformation is no
more.

We've often argued at length about whether people will rank sincerely
in TUC, or whether they'll favorite bury. That argument was
unnecessary. I'm not trying to re-start that argument, but I'll just
say that the mass media, particularly the tv, have nearly everyone
convinced that the Democrat is acceptable, and that the winner must be
a Democrat or a Republican. It's a simple, obvious and undeniable
conclusion that, given those assumptions, with TUC, optimal strategy
consists of ranking the Democrat alone in 1st place. That's why I
reject TUC, for the existing electorate.

But now I'm talking about an electorate that no longer believes that
Democrat-disinformation, and has just elected a Green government. So
the assumptions listed in the previous paragraph are no longer
believed. Maybe then we'd have the situation in which the detailed
information needed to rank other than 0-info isn't available. That's
when TUC(margins)'s 0-info Sincerity (if it really has that property)
would come into importance.

If, under those conditions, TUC(margins) would encourage sincere
ranking, and because TUC meets the Condorcet Criterion, then the goal
of encouraging the election of the CW could become attainable.

...except for the chicken dilemma. Aye, that's the rub.

I wouldn't want any TUC method, for any electorate, but in an
electorate free of the Democrat-disinformation, one could argue for
TUC(margins), if one wanted to try to ignore the chicken dilemma.

The Plurality criterion isn't my reason for not wanting TUC(margins).
Plurality doesn't have the obvious importance and relevance of
Participation, or Mono-Add-Solo-Top (the latter passed by IRV, but
failed by TUC). I've never seen what others seem to see in the
Plurality Criterion.

If you're going to disregard Plurality-failure, then you might want to
consider the fact that MMPO meets FBC, CD, Later-No-Harm,
Mono-Add-Solo-Top, Participation, and Mono-Add-Top.

MMPO has a much more insincere bottom-end strategy need than the other
methods, but I've never considered bottom-end strategy need to be very
important. I proposed Symmetrical ICT to gest rid of bottom-end
strategy-need, because bottom-end strategy-need is a nuisance--but not
a real problem.

MMPO can be said to have great "burial incentive". Sure, its bottom
end u/a strategy is to rank the unacceptables in reverse order of
winnabililty. Others will be doing that to your favorite. But so what?
You'll be doing the same to theirs.

By the way if (and I realize this isn't likely) Mono-Add-Solo-Top
hasn't been named yet, then that would mean that I get to name it. I
don't like "Weak-Participation", because that implies unimportance,
even though a "weak" criterion is more important to not fail.

So maybe:

Mono-Add-Solo-Top
Mono-Add-Only-Top
Mono-Add-Unique-Top

But I prefer the following two names:

Defeat-Your-Favorite
Make-Your-Favorite-Lose

Of course it's met by every method that meets Participation, such as
Plurality and Score.

IRV passes. ICT, TUC and Majority-Judgment fail.


Juho said:

I guess Condorcet methods can't ever be successful since they fail
such "terrible" criteria as "favourite betrayal" and "burial". ;-) I
mean that whatever method you want to promote, there are some nasty
negative marketing pitches that you must be ready to answer.

[endquote]

As I said above, given what the current electorate believe, the
Democrat-disinformation, FBC-failing methods have the optimal strategy
of ranking the Democrat over everyone else.

Right now that's a problem. In the future in which the voters have
overcome that disinformation, it would be a different story. Then, FBC
wouldn't be _necessary_, though it would always remain desirable. (My
rankiing of methods, for those future conditions reflects FBC's
continuing desirability).

CD is important too, and will remain important even as the need for
FBC is reduced by the demise of the Democrat-disinformation.

Juho said:

Most ranked methods lose a considerable part of their benefits if
people generally truncate.

[endquote]

Well, information needed to elect a CW is lost. But, then, it's
usually lost anyway, for other reasons, and so that doesn't really
matter.

As I said, without the Democrat-disinformation, TUC(margins) would
have a very good (probably unequalled) chance of electing CWs, if it
weren't for the chicken dilemma.

Symmetrical ICT encourages truncation, when it frees you from need to
rank unacceptables. That's a good thing, unless you want to elect a
CW. But IC methods encourage top-ranking all of the acceptabless, and
that won't help elect a CW. Your best chance for electing a CW is to
use TUC(margins), when people no longer believe the
Democrat-disinformation, and when you hope that there won't be a
chicken dilemma.

Juho said:

In the worst case the behaviour of the methods starts to resemble
plurality. It is quite common that the "opposite side" will win, and
then one's vote is "wasted" unless one tells which one of the
"opposite side" candidates is best. I believe that many enough people
will learn not to truncate, to make the results of the election
meaningful and fair.

[endquote]

Most rank methods have a bottom-end strategy that calls for ranking
all of the unacceptables. MMPO calls for ranking them in reverse order
of winnability. IRV, TUC and ICT call for ranking them sincerely.
Symmetrical ICT frees you of need to rank them.


Btw, I think that in some sense margins are easy to advocate. The
reason is that there is one simple answer that, if you want, can be
used to answer to most attacks against Minmax(margins). The idea is
roughly that you thwart all attacks on details by repeating that the
method will elect the candidate who needs least additional votes to
beat all others.

[endquote]

No it doesn't. Dodgson does. But we've agreed to disagree on that.

Juho said:

That is a simple and quite understandable rule that makes sense, not a
complex algorithm, and any deviation from that simple rule can be
presented as violation of the basic principles of the election.

[endquote]

Whoa. Juho is now making the rules of the election.




Juho said:

We may thus accept that there are some theoretical cases of burial etc

[endquote]

Certainly with MMPO. See above. But maybe not with other methods, if,
as I've suggested, the detailed information needed to do other than
0-info ranking won't be available.


Juho said:

., but we can say that we should still stick to the idea of electing
the best candidate as defined by the "least additional votes" rule,
since doing otherwise would not elect the best winner (maybe more
often).

[endquote]

There's no "best winner". We've been over that. But, if you really
want a best winner, then look at the significant social optimizations
of Approval and Score.

Juho asked:

Isn't that a valid marketing approach?

[endquote]

Yes. It's a valid marketing approach for Approval and Score.




>
>> I think the correct message to voters
>> in most ranked method based elections is to encourage them not to truncate but
>> rank all relevant (good and less good) candidates sincerely.

Well, it;s a good message to send if you want to elect the CW. If you
want to elect the CW, then sincere ranking is the _necessary_ message.
TUC(margins) may be the unique method that meets the Condorcet
Criterion and 0-info Sincerity. I've spoken of how FBC failure and CD
failure could spoil that. Say we've just elected a Green government,
by voting for the Green, even though s/he isn't your favorite. Are you
going to have the courage (or foolhardiness) to not rank the Green in
1st place, or alone in 1st place, then, if we're using TUC(margins),
when that could allow the return of the Republocrats?

So I doubt that TUC(margins)'s potential CW-electing ability would
really materialize. Maybe, then, people should reluctantly give up the
elusive goal of electing the CW. That's my take. Just work on reducing
strategy needs, eliminating the worst strategy needs.

Juho said:

Maybe the question of truncation will pop up in various places. I
wonder who would give a general advice "yes, truncation is what you
should do".

[endquote]

I would, if the method were Symmetrical ICT. A lot of people won't
want to have to rank all of the candidates. In Symmetrical ICT they
don't need to. In IRV, in a u/a election, they don't need to, even
though they might benefit some from doing so. Ideally they should in
IRV, but ranking unacceptables doesn't matter with regard to making an
acceptable win.

The only method I know of in which truncation is generally best, in a
0-info u/a election is Symmetrical ICT.

But, what if there are 20 or 30 candidates? Wouldn't you prefer a
method that doesn't make you need to rank the unacceptables?


Juho said:

That doesn't make much sense. I guess it would be more popular to say
"you can rank the candidates in the new method, so be thankful about
that and rank them".

[endquote]

Most methods encourage you to rank them all. Some methods, including
ICT, TUC, and iRV encourage you to rank them sincerely. But, if there
are lots of them, or if ranking unacceptables is distasteful to you
(as it is to me), then you might prefer to not need to rank them. You
don't need to in Symmetrical ICT or IRV. You probably shouldn't if the
method is Symmetrical ICT. You definitely shouldn't if it's 0-info and
Symmetrical ICT.

Juho said:

It is probably easier for the politicians to say "at least I will not
even rank their candidates". That is a message saying that this
party/wing is so bad that I don't even want to touch them.

[endquote]

Exactly my sentiment. I prefer a method that doesn't make me have to
rank the Democrat, to help protect against the Republican. In
Symmetrical ICT there's no need to. In IRV there's no need to, to
protect acceptables against unacceptables, though it's ideally optimal
to rank all.

Juho said:

But I guess people are used to hearing this kind of negative talk from
the politicians. They might react by ignoring such recommendations, or
alternatively the polarization would raise to the levels of plurality
style voting between the two major wings.

[endquote]

With a genuinely better method, there won't remain "the two major
wings". With Approval, Score, and Symmetrical ICT, it will be a matter
of approving, top-rating, or top-ranking all the acceptables, and
ignoring or bottom-rating the unacceptables. That isn't a bad thing.


Juho said:

I hope and guess that the first approach will become more popular in time.

[endquote]

It depends on the method. There's nothing wrong with the Approval,
Score, Symmetrical ICT strategy described above. Not everyone wants to
rank unacceptables.

TUC(margins) can, under ideal conditions, encourage sincerely ranking
everyone, but it's questionable whether those conditions will ever
really obtain.

Also, it isn't obvious that that goal, or the corresponding goal of
electing the CW, is necessary.

Mike Ossipoff