[EM] CVD wants Alt.V to be fairer but it isn't: misleading website

[MIKE OSSIPOFF]

you wrote (2 Oct 2000):
 > IRV, in examples like that, fails to elect the SCW, even when he's
 > the favorite of more people than any other candidate is. At least
 > FPP won't do that.

Marcus commented:

Of course, it is also possible to create examples where IRV chooses
a SCW and FPP doesn't.

I reply:

Sure, but what could be more typical & common than candidates'
support tapering gradually away from the voter median? And that
guarantees that IRV will do worse than FPP.


******
So say a method meets IFCC. Would it meet a more realistic version?
 > Maybe not, and so I suggest that meeting IFCC doesn't mean a whole lot.


Marcus continues:

I don't believe that you don't consider independence from clones to be
important. Otherwise you would concentrate only on PC and you wouldn't
promote three methods in your http://electionmethods.org website.

I reply:

I didn't say that there can't be an important clone criterion. I merely
said that the existing one is too unrealistic to be important.
Anyway, it doesn't apply to FPP. When a criterion can't be applied to
all methods, that tells us that it needs to be defined better.

I assume that the ICC is as I worded it, and makes no mention of
anything about voters other than their votes, as you like criteria
to be. The possible improvements that I described yesterday would
define a clone set in terms of sincere preferences rather than actual
votes. I forgot to mention that yesterday.

As for your 2nd sentence, I must admit that I don't understand why
my advocacy of Tideman(wv) & SSD instead of PC says anything about
ICC. Are yous saying that PC passes ICC and that Tideman(m) & SSD
fail ICC?

Or only that PC passes ICC, and nothing but ICC matters, and so
we should propose PC instead of better Condorcet versions?

I prefer Tideman(wv) & SSD to PC because they meet GSFC & SDSC, and
PC doesn't. Also, they meet some criteria that the academics like,
and which PC doesn't meet, such as Smith, and certain criteria that
go with Smith, such as Condorcet Loser, Mutual Majority (for what it's
worth), and Majority Loser. So of course I prefer Tideman & SSD.

But PC is pretty good too. Though it fails Smith & related criteria,
so does FPP, when those criteria are defined fairly--in terms of
sincere preferences of voters :-)  So no one can criticize PC as a
replacement for FPP. It isn't as good as the very best, but it's
still very good, and competitive with FPP.

I've been proposing Tideman & SSD, however, because many feel that it's
better to ask for the best. I find Tideman easier to demonstrate with
an example, and so that's my main proposal. But for the national
LWV study, which hasn't begun yet, but which might allow methods that
haven't been in use, there would be a good case to propose PC, because
it dominates FPP, or SD, because it dominates IRV. Nonmonotonicity
is ridiculous, but no more so with SD than with IRV. If the object is
to beat IRV, then SD is a possibility. Of course Tideman & SSD
dominate FPP & IRV, but PC & SD are simpler. Any suggestions on which
Condorcet version we should offer to national LWV as an alternative to
IRV? Anyone want to help with that effort?


Markus continues:

PC is the best Condorcet method when one concentrates only on
strategical voting and ignores strategical nomination.

I reply:

That's news to me. Did you know that PC fails GSFC & SDSC? Those
are criteria about freedom from need from strategic voting.

Marcus continues:

PC meets
Saari's positive involvement criterion and Fishburn's no-show
criterion.

I reply:

It's true that PC does better at adverse results criteria than
Smith-complying methods do. But if you think that's what's important,
then you should advocate Approval, FPP, or Borda, because they, but
not PC, comply with Participation. Or you should advocate only
Approval, because it alone complies with SARC. In addition to
Participation.

Obviously I don't consider adverse results criteria to be the
most important criteria. I'm more interested in defensive strategy
criteria. I admit that PC does better than your BeatpathWinner
by adverse results criteria.

Marcus continues:

PC guarantees that a dichotomous voter cannot be
punished for showing up and voting sincerely.

I reply:

I don't know if the better Condorcet versions guarantee that or not,
but if voters' preferences are dichotomous, then there's no reason
to use anything other than Approval. If you advocate anything other
than Approval, then you're saying that you know that preferences
aren't dichotomous. Not to be tedious about our Presidential election,
but it's obvious that everyone or nearly everyone voting for Gore
has nondichotomous preferences, with at least 3 levels:

1. Nader
2. Gore
3. Bush


Markus continues:

PC isn't vulnerable to what Steve calls "indirect strategies."
That means: If PC is used then it isn't possible to change the
winner from candidate A to candidate B by ranking two different
candidates C and D insincerely to each other.

I reply:

Does BeatpathWinner fail that one? Does Tideman(wv) pass it?

In any case, you again seem interested in what strategies are possible,
your goal being to prevent them from working, while I'm only interested
in eliminating or reducing the need for defensive strategy.

Markus continues:

Therefore when you don't consider independence from clones to
be important then why do you promote more methods than just PC?

I reply:

You know the answer to that: I prefer Tideman & SSD to PC because
they meet GSFC & SDSC. Approval has the following advantages over
PC: FBC, SARC, & Participation (by far less important to me than FBC
& SARC). PC has the following advantage over Approval: SFC, and
near-compliance with SDSC.

As for Approval vs Condorcet take a look at the chaos & disarray
among rank-balloting advocates. We're lucky if we can enact a
rank method other than IRV. To propose Condorcet, we have to
convince people that CVD, with its big money, brochures, large
membership, celebrity advisory board, is wrong. Good luck :-)

Approval the advantage that we can say "Look how much easier this
is to implement. And, unlike IRV, Approval isn't a radical change
from what we already have. It's Plurality done right." People listen
to cost arguments. Approval costs zero dollars to implement.
What change is needed? On the ballot, where it says "Vote for 1", it
would instead say "Vote for 1 or more". That's it.

With the CVD IRVies so busy, how much chance do you think we'd have
of enacting any method that meets BC?

Not that we shouldn't try that too, of course. The national LWV
study is a good place to fight for Condorcet. But in California,
Condorcet is out, and Approval is the thing to try for, as is likely
to be the case in many situations. But I'd try for Condorcet when
there's a chance.

You're right though: There may very well be a place for PC advocacy,
but the reason has nothing to do with ICC.
@planet-interkom.de

Mike Ossipoff


_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Share information about yourself, create your own public profile at 
http://profiles.msn.com.