[EM] James A. didn't read SDSC before expounding on it

[MIKE OSSIPOFF]

James A. said:

Dear election methods fans,

	I suggest that ordinary winning votes methods (beatpath, ranked pairs,
river, etc.) fails Mike Ossipoff's "strong defensive strategy criterion",
according to what I think is the most reasonable interpretation of that
criterion, whereas cardinal pairwise passes the criterion.
	Here is the definition, from electionmethods.org

Mike's strong defensive strategy criterion (SDSC):
"If a majority prefers one particular candidate to another, then they
should have a way of voting that will ensure that the other cannot win,
without any member of that majority reversing a preference for one
candidate over another or falsely voting two candidates equal."

I reply:

  James A. was in too much of a hurry, and didn't notice that our SDSC 
definition is accompanied, immediately preceded, by a definition of voting 
two candidates equal.

  It goes something like this:

A voter votes X and Y equal if s/he votes X over someone, and votes Y over 
somone, and doesn't vote X over Y, and doesn't vote Y over X.

[end of definition of voting two candidates equal]

  James A: Look at all the time you spent writing that posting, and the 
others that you subsequently wrote, in the same discussion-thread. If you 
were willing to spend all that time, it's too bad you didn't take the time 
to read the criterion before wasting all that time on your long postings 
based on your fail to read it.

James A. said:

	I suggest that if my sincere preferences are A>B>C>D>E, and I truncate my
ballot after candidate C, I am essentially voting A>B>C>D=E, that is,
"falsely voting two candidates equal."

I reply:

Suggest whatever you want to, but, accompanying SDSC's definition, we define 
voting 2 candidates equal, and our definition is different from yours.


James A continued:

Hence, although Mike may not have
intended it this way

I reply:

Not only did I not intend it that way, but I said it otherwise.

James A. continued:

, I suggest that the majority of voters in his
definition should not have to use either order-reversal, OR truncation, in
order to absolutely prevent the other candidate from winning.

I reply:

Then you're saying it the method should pass if they have to rank 2 
candidates explicitly at the same rank position, as long as they don't have 
to truncarte? :-)  You aren't quite saying what you mean, are you.

Anyway, of course what you need to do then, is define your own criterion. 
What's the point of saying that an already-existing criterion, in your 
opinion, should be different from how it is? It would make more sense for 
you to define and name your own criterion that says that those voters 
shouldn't have to reverse a preference or truncate. And did you mean that 
they also shouldn't have to insincerely vote candidates equal, with both 
candidates in their ranking? Anyway, so write it and name it and post it.

James A. continued:

	The point of this e-mail, the positive point, is that cardinal pairwise
actually does pass this criterion, even with the strict definition that I
have given it.

Because, in cardinal pairwise, if a majority ranks X above
Y, and rates X at 100 and Y at 0, there is no way that Y can win. In doing
this, the majority does not need to make a reversal or a false
equalization.

I reply:

No, they just have to give maximum points to X, rating X equal to more 
preferred candidtates, and give minimum points to Y, rating Y equal to 
candidates to whom they prefer Y.

Of course, if you rate X & X1 equal, but rank X1 over X, and if we count 
only your ballot, with all the candidates but X & X1 deleted from it, most 
likely X1 would be the unique winner. Is that true? If so then, by Richard's 
definition, which I use, you've voted X1 over X, and therefore haven't voted 
X & X1 equal.

If so, are you sure that by a majority giving 100 points to X, and 0 points 
to Y, and ranking X over Y, that ensures that Y can't win? Can you 
demonstrate it?

IIf so, then your method indeed meets your stronger version of SDSC.

The usefulness of that depends on how much influence they're having on X1 vs 
X, when they rank X1 over X, but rate X equal to X1. Even if technically 
they're still voting X1 over X, to what extent are they actually helping X1 
win against X?  If not much, then your method's compliance with that 
stronger version of SDSC is a hollow victory. Though you're voting X1 over 
X, you're doing so in a way that's surely less effective than if you'd 
ranked X1 over X in ordinary wv Condorcet.

So you're gaining for that majority the freedom to truncate, in return for 
less effectively voting X1 over X. And the question then is how much less 
effectively are they voting X1 over X, as compared to if they ranked X1 over 
X in  PC, SSD, CSSD, BeatpathWinner or RP?

Sure, even with those wv methods, the guaranee of ordinary SDSC only is that 
they can make Y lose. Just as with your method and your SDSC, they can't 
simultaneously make X lose because they constitute a majority for X1>X. SDSC 
only guarantees the defeat of one candidate. So I'm of course not saying 
that your method should be able to guarantee that the X1>X majority should 
be able to ensure X's defeat while also ensuring Y's defeat. I"m merely 
asking just how ineffective their vote for X1 over X is, given that they're 
rating X & X1 equal.

Mike Ossipoff

_________________________________________________________________
Is your PC infected? Get a FREE online computer virus scan from McAfee® 
Security. http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963