[EM] Russ, SFC, and wv vs margins

[MIKE OSSIPOFF]

Russ quoted my definition of SFC:

SFC:

If no one falsifies a preference, and if a majority prefer the CW to
candidate Y, and vote sincerely, then Y shouldn´t win.

[end of SFC definition]

Russ continued:

Here's my comment:

Mike considers this criterion critical and uses it as evidence that
certain Condorcet variations are less vulnerable to insincere strategy

I reply:

I´ve repeatedly said that emphasis on vulnerability to insincere strategy is 
barking up the wrong tree. The problem is the need for defensive strategy 
that conceals one´s genuine preferences. Yes I admit that offensive 
insincere strategy in Margins Condorcet can cause serious problems, but the 
problems that it causes has to do with creating a need for defensive 
strategy that falsifies one´s preferences or buries one´s favorite.

Russ continued:

...when they base defeat strength on "winning votes" rather than "margins."


Forget for now all the disputes about Mike's definition of preferences.
Even if Mike's definition of preferences is clear and unambiguous (and I
realize that's a big "if")...

I reply:

A few days ago, Russ posted a statement that SFC is well-defined, even 
though Russ claimed that he believes that SFC isn´t useful.

Russ continues:

, what is the significance of SFC?

I reply:

It´s as I´ve been explaining it all this time.

Russ continues:

Well, since Mike feels that he can write his own version of the
Condorcet criterion, I'll write my own version of SFC, and I'll call it
the margins SFC:

If no one falsifies a preference, and if the margin of the victory of
the CW over candidate Y is larger than any other margin of victory, then
Y shouldn't win.

[end of margins SFC definition]

Does it now appear that margins is less vulnerable to strategy than
winning votes?

I reply:

Is that criterion of yours supposed to be in some way a counterpart of SFC? 
Your criterion is a method-dedicated criterion, of course, mentioning 
margins.  SFC doesn´t mention winning-votes.

SFC is about a majority not not needing strategy to accomplish the familiar 
goal of ensuring that some greater-evil won´t win.

By the way, Plurality meets "Margins SFC", though Plurality is far from free 
from strategy for members of a majority who prefer the CW to Y, when the 
members of that majority have different favorites. So it´s odd that you call 
your criterion SFC. Or do you now believe that Plurality is strategy-free 
under those conditions?

SFC stands for Strategy-Free Criterion, because, with complying methods, the 
majority described in SFC´s premise doesn´t need any strategy in the 
plausible situation of that premise.

Russ continues:


When you get right down to the basics, Mike's SFC is simply an arbitrary
criterion

I reply:

It isn´t entirely clear what Russ means by an arbitrary criterion. If 
"arbitrary" means not wanting  majorities to need defensive strategy that 
conceals their preferences and surrenders their aspirations, , then call me 
arbitrary. :-)

Russ continues:

that happens to favor winning votes, but an equally arbitrary
criterion can be written to favor margins.

I reply:

Well, you certainly have proven that an arbitrary criterion can be written.

Russ continued:

Hence, Mike's SFC criterion
is completely irrelevant to the debate over winning votes vs. margins.
In fact, it's completely irrelevant, period. And so is its generalized
version, GSFC, of course. They're both really just pedantic tricks.

I reply:

And people should take your word for those things? Or is it that you´re 
saying that you´ve proven it by your explicitly margins-dedicated criterion 
that Plurality passes?

Anyone can write a criterion such as the IRV criterion that says that a 
method passes the IRV criterion if it chooses IRV´s winner.That´s 
essentionally what your "Margins SFC" is.

Russ continues:

I realize that "margins vs. winning votes" is an old topic here, but I
would just like to add my two cents worth.

I reply:

Ok, you´re going to add something new? Good, when will you start? But 
perhaps "worth" isn´t the right word for what you add.

Russ continues:

If you argue for wv, you are claiming that a 51-49 victory is "stronger"
than a 49-0 victory. Common sense tells us that's nonsense. Some of us
still have common sense.

I reply:

That´s been answered many times. You´ve shown that a pairwise defeat that is 
better by winning-votes can look not as good by margins. Amazing. It´s 
because wv and margins are different.

Russ wants us to count those 49 voters who voted against the 51-49 defeat. 
Some like to do that, saying that it´s more "symmetrical" if we count them 
as we count the winning-votes.

But, if X pairwise-beats Y, the X>Y voters and the Y>X voters aren´t 
symmetrically related. Why?

The Y voters lost.

If every candidate has a pairwise defeat, then, to elect someone, we have to 
disregard their pairwise defeat, we have to overrule the public´s statement 
that someone else would be better, if we´re to appoint that person the 
winner.

When we do that, we´re overruling the people who voted for the X>Y defeat. 
As I said, we shouldn´t do that lightly. So don´t avoidably overrule more 
voters than necessary, by using a count rule that doesn´t distinguish the 
voters who won from the voters who lost.

You see, Russ, if we keep the X>Y defeat, we aren´t overruleing the Y>X 
voters. The Y>X voters were overruled when they were defeated in the public 
vote between X and Y, when the public said that they collectively prefer X 
to Y.

Why does margins violate majority rule so readily? Because, when it 
subtracts to calculate the margins, it destroys information about 
majorities.

Aside from those reasons, another reason why I advocate wv instead of 
margins is because wv meets the defensive strategy criteria that I chose 
because they measure for the goal of getting rid of need for defensive 
strategy that conceals preferences or buries one´s favorite.

Russ continued:

Just for fun, let's frame this in terms of Mike's definitions of
preferences and sincere voting. I don't feel like searching for it now,
but Mike recently wrote something to the effect that a sincere vote is
one in which the voter does not falsify any preferences and votes every
preference that the particular method allows.

I reply:

That´s not it, but maybe you´re just trying to sound roughly like it.

Russ continues:

That means that a "sincere" vote cannot be truncated unless the voter
truly rates all the unranked candidates *exactly* equal.

I reply:

Yes, with a method that allows the ranking of all the candidates, and counts 
those rankings in a way such that ranking X over Y means voting X over Y, by 
my definition of voting X over Y, a sincere ballot would only be one that 
ranks all of the candidates in accord with the voter´s preferences.
So if you don´t rate anyone equal, then yes, you have to rank them all if 
you want to vote a sincere ballot.

Russ continues:

Well, what does
that mean? As I tried to explain previously, it requires a model of
voter preferences

I reply:

But it isn´t enough to try to explain it. You need to try to show it. In 
fact, you need to successfully show it.

I told why my criteria are well-defined no matter what "prefer" means, even 
if it has only been defined imprecisely, even if the meaning of "·prefer" is 
unknown, and even if "prefer" doesn´t mean anything.

If you want to deny that, then you need to tell which sentence you claim is 
incorrect in my discussion about that.

I refer you to my posting whose subject title was "Four approaches to the 
meaning of preference", or something like that. It was posted some days ago.


But, for those who prefer having a defintiion of preferences, I included on 
in that posting. It´s a simple definition, and so it doesn´t correspond 
perfectly to what we usually mean by preference. But that doesn´t matter:

A person prefers X to Y if, given the choice between X and Y, that person 
would choose X instead of Y. By my definition, if you would choose pepperoni 
pizza over some other kind, except that someone says that they´ll beat you 
up if you choose pepperoni, and so as a result you´d choose some other kind 
if given the choice, you then prefer some other kind to pepperoni. Well, it 
could be argued that, in a sense you do, since, taken with all the 
consequences of the choice, you prefer the results of choosing something 
other than pepperoni. Sure, that definition could be improved by stipulating 
the condition that your choice isn´t influenced by anything other than the 
intrinsic qualities of the things between which you´re choosing. That loses 
some definiteness of meaning, but that´s ok, since a precise definition of 
preference is not needed for my criteria. So, take your pick: My simple 
preference definition, or my enhanced one, both of which are stated in this 
paragraph. But remember that no defintiion of preference is needed for my 
criteria. I refer you to my posting about that.

Russ continues:

What do
I mean by a "model" of voter preferences?

I reply:

I´m sure you´ll tell us. Feel free to define it as you please. But don´t say 
that everyone must use your definition or "model".

Russ continues:

The very concept of preferences implies that the voter either explicitly
or implicitly rates the candidates on a one-dimensional scale. Suppose I
claim that the scale is continuous rather than discrete. Well, it's as
good a model as any. That means that the voter's ratings of the
candidates fall essentially randomly on a continuous real-valued scale.
It also means the the probability of two candidates being rated exactly
equal is zero.

I reply:

But not in practice. Don´t believe in your model so much that you want to 
use it to replace actual-world facts.

Russ continues:

If the probability of exactly equal ratings is zero, that means that
Mike's definition of a "sincere" vote has zero probability of being
truncated.

I reply:

So far, so good. Congratulations, Russ.

Russ continues:

It also means that no equal rankings will occur.

I reply:

No it doesn´t. It only means that a particular voter won´t vote equal 
rankings if his/her ballot is sincre, in a voting system that allows all 
candidates to be ranked, and counts ranking X over Y in such a way that 
doing so means voting X over Y, by my defintion of voting X over Y. (Most or 
all seriously-proposed rank methods do so).

Russ builds on that incorrect statement:

And what
does that all mean? Your [you're] jumping the gun! Yes, that's right! It 
means
that winning votes and margins are equivalent, given that continuous
model of voter preferences.

I reply:

Russ has confused "A particular voter won´t vote equal rankings if s/he 
votes sincerely in a rank-method" with "No one will vote equal rankings".

Indeed, if no one would equal rank or truncate, then margins would be the 
same as wv. But nothing in my definitions implies that no one will rank 
equally or truncate.

Every rank balloting election that I´ve conducted or participated in, and 
probably every rank balloting election that I´ve heard of, except those that 
explicitly require complete rankings, have truncation. In one case, the 
truncator specifically said that he was doing so as an offensive strategy, 
though he didn´t use those words.

Russ continues:

As Blake pointed out, we can think of truncated votes as more or less
equivalent to the same votes completed with random rankings.

I reply:

Certainly Russ can think of anything any way that he wants to. But when it 
comes to counting a truncated ranking, I wouldn´t add to it preferences that 
the voter didn´t vote.

Aside from demonstrating Russ´s usual confusion, this posting 
well-expemplifies what I was saying about how what Russ says has little if 
anything to do with what Russ actually believes.

A few days after I told Russ why he would stop advocating wv, Russ 
predictably stopped advocating wv. I´d asked Russ if he was going to become 
a margins advocate now. And a few days later, Russ becomes a magins advocate 
:-)

Mike Ossipoff

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