[EM] Replies to two postings from Jameson Quinn

[MIKE OSSIPOFF]

Jameson--

You said:

There are three possible kinds of criteria:

1. Criteria which apply to ballots and do not mention preferences.

[endquote]

When I say "apply to", in connection with criteria, I'm referring to what methods can be compared with
those criteria. I don't know what "apply to" means when you use it in this posting.

I'm going to guess that "apply to" means "stipulate about in the criterion's premise".

You continue:

2. Criteria which apply to ballots and mention preferences.

[endquote]

I don't know what sort of "mention" you're referring to. I'm going to guess that, when you
say "mention", you're referring to some kind of mention other than stipulation in the
criterion's premise. But, in that case, it isn't clear what you mean by "mention".

You continued:

 for instance, as part of the justification for why the criterion is desirable

[endquote]

Such "mention" of course isn't part of the criterion, and so your #2 criteria are
essentially #1 criteria, as regards the criterion-wording itself.

You continue:

3. Criteria which apply to preferences.

[endquote]

I'm going to guess that you're referring to criteria that stipulate preferences and
a relation between preferences and votes.  And maybe sometimes directly stipulate votes too.

Remember that stipulation of preferences and a relation between preferences and votes
amounts to an indirect stipulation of votes.

You continue:

You are saying that type 1 is bad

[endquote]

No, I'm not. I'm merely saying that often type 1 criteria are ridiculous.  ...as when Plurality meets Condorcet's Criterion and Minimal Defense. Probably Majority-for-Solid-Coalitions too, unless it inelegantly stipulates against Plurality.

You continue:

so we are forced to choose type 3.

[endquote]

Some criteria that I use are votes-only criteria. Some of them aren't ridiculous.

You continue:

Everyone else is telling you that type 3 is essentially incoherent. 

[endquote]

The "everyone" who, in current discussion here, have a problem with preference criteria consists of you and one other person :-)

Can you be coherent enough to say what you mean, and then justify it?

Specifically, what particular sentence in one of my preference criteria don't you understand the meaning of? 

In what way is one of my preference criteria ambiguous in meaning. What different meanings could it have? 

Is there a reasonable or proposed method that can't be unambiguously said to pass or fail one of my criteria? 
What method and what criterion?

...Or are you just incoherently making angry noises?


>From your other posting:

You wrote:
The scenario:

> 49: C
> 27: A>B
> 24: B(>A sincere)


Options:
1. Elect B
pro: it makes sense given the ballots

[endquote]

Unsupported assertions won't do. How does it make sense?

The B voters are the smallest faction. 

Remember that, in MMPO and MDDTR, middle ratings are not approvals.

You continue:

con: the B voters are getting away with their strategy, and next time the A
voters will probably follow them.

2. Elect A
pro: It is correct for these sincere preferences
con: it is badly wrong for other possible sincere preferences that could
have given these ballots, especially if the A voters are only strategically
voting for B.

[endquote]

How is it wrong? As I said, the B voters are the smallest faction. The A voters
are the largest faction whose candidate doesn't have someone voted over hir
by a majority.

I assume you mean that maybe the A voters are really indifferent between B and C.

For one thing, true indifference is rare. For another thing, the A voters' votes are
their own, and they can use them as they wish, for whatever reason.

One flaw in the "random-fill" criticism is the rarity of genuine indifference.

You continue:

3. Elect C
pro: well, that will teach those sneaky B voters a lesson!

[endquote]

Better yet: If they know that's what defection will do, then they won't defect.

You continue:

con: What if the B voters were being sincere?

[endquote]

If they're sincerely indifferent between A and C, then they won't mind of C
wins instead of A, will they?

Or are you saying that it's unfair to the sincere B voters if B doesn't win? As
I said, B voters are the smallest faction. And, in MMPO and MDDTR, middle ratings
are not approvals. Why should they expect B to win?

You continue:

4. Allow candidate C to choose whether A or B is elected. (The SODA
solution)

[endquote]

So you're saying to let the election be decided by a faction who has a majority defeat.
No thanks.

And there is no way that the public, in any jurisdiction, will agree to let the method rules
automatically give that decision to a candidate. What are you mixing in your soda?

You continue:

My preference over these options: 4>3>1>2. Apparently, Mike Osipoff's
preferences are something close to the reverse of this.

[endquote]

Almost. My preference-ordering is: 2>3>1>4

There's nothing wrong with electing C in that example.

And maybe there's a method that can elect C, without diverging from Plurality enough to bother some people,
as do MMPO and MDDTR.

Is there a method that elects C in the above example, and meets FBC and at least one of {SFC, SFC3, 1CM, 3P}, and 
meets Mono-Add-Plump, and doesn't elect C in Kevin's MMPO bad-example that I posted?

 
Mike Ossipoff