[EM] MJ SFR (preliminary). Score vs Approval, based on considerations discussed.

[Ted Stern]

On 11 Sep 2012 13:18:23 -0700, Michael Ossipoff wrote:
>
> Ted:
>
> You said:
>
>> Majority Judgment (MJ) and Continuous Majority Judgment (CMJ) are both
>> Median Ratings methods.
>
> No sh*t !   :-)    ...But wait, isn't that explicit in their definition?
>
>> As is ER-Bucklin(whole).  You're probably most familiar with the
>> latter, so let me start there.  I will put ER-Bucklin into the same
>> formulation as MJ and CMJ
>
> Ok. I've heard the claim that MJ is ER-Bucklin. Maybe it's true.

Here is where you go off track:

I don't think this has ever been claimed, and certainly not by myself.

> MJ would probably be much easier to count than ER-Bucklin.

The tabulation is essentially the same.  I don't see that this is
relevant.

> But what would their equivalence (if valid)

There is no equivalence.

>  mean, in practical terms?

Your question is meaningless, because they are not equivalent.

> There aren't many people advocating ER-Bucklin.

http://en.wikipedia.org/wiki/Argumentum_ad_populum

> So the equivalence, if valid, isn't a powerful argument for MJ.

It isn't valid.  Your statement is a 
http://en.wikipedia.org/wiki/Fallacies#Straw_man

>
>>  so you understand the terms,
>
> I assure you that it isn't hard to make terms understandable. All
> that's necessary is to define them clearly.

Apparently I have not been clear enough.

> You seem to contradict yourself, saying at one point that MJ is
> ER-Bucklin,

I believe you will find that you are contradicting yourself.  I never
stated that.

> and, at another point, that MJ will often give the same result as
> ER-Bucklin.

Voting methods can frequently agree on the same winner.  That doesn't
mean the methods are identical.  Instead, it should increase
confidence in the strength of the result.

> You've supplied additional confirmation for the conclusion that MJ's
> (and CMJ's) tiebreaking bylaws are elaborate, complicated, and
> wordy.

My goal, which I *thought* I had stated clearly, was to help you avoid
misinterpretations by being explicit.  I wasn't attempting to make a
persuasive argument about the merits of Majority Judgment.

As you should be well aware, being unambiguous and explicit can lead
to elaborate constructions.

On a playing field of good will and common understanding, it is
possible to use much simpler and more persuasive language.  I don't
think we have come to that place yet.

> MJ and CMJ are so elaborately, wordily, defined that few people
> would be willing to listen to their definitions.

I think we can conclude only that you yourself are unwilling to listen
to their definitions.

>> In many cases, MJ, CMJ and ER-Bucklin will choose the same winner.
>
> Whoa. You earlier said that MJ _is_ ER-Bucklin.

See above.  (Mis-)Proof by repetition.

> Which is it? Is MJ the same as (equivalent to) ER-Bucklin, or is it
> something different from ER-Bucklin that will, in many cases (but
> not always), choose the same answer?

The latter.  The former is clearly nonsensical, as you have asserted
repeatedly.

> But, in any case, what does it matter, since few advocate Bucklin
> anyway?

A:  There is a reasonable interpretation.
B:  You choose to ignore it, again using argumentum ad populum.
C:  Even your argumentum ad populum is incorrect.

The problem with the term "Bucklin" is that it has been applied to
several different methods over a century.  ER-Bucklin has relaxed
conditions that avoid some of the problems of other formulations.  It
is still not a perfect method.  However, since most people cite more
limited versions of Bucklin, the "popular" understanding is mostly
used for negative contrast with the citer's preferred method.

Balinski and Laracki have sought to make a mathematically rigorous
argument to support a particular form of median ratings.  They appear
to be winning a much larger audience, in part because of large scale
studies that support their conclusions.

> If you want to compare the merits of MJ to that of Score, then
> compare what MJ does to what Score does. Compare the strategy
> situation in MJ to that in Score. That's what the previous
> discussion has been about.

I was not discussing the merits of Majority Judgment.  I merely
intended to clarify the algorithm for you.  I seem to have failed.

Please continue your discussions re merits with the other MJ
advocates.

[... Here Ossipoff elides the entirety of section 2 from his first
method in this thread that my next quote refers to ...]

>> My only comment about this is that, since your quoting style is
>> non-standard,
>
> In the posting to which you were replying, I quoted in the standard
> style, using ">" and ">>" for previous text.

I think you will find that in the message I replied to, there were no
quotations at all.

>> I really wish you'd provide a glossary of abbreviations somewhere
>> in your message, either inline, using standard first-reference
>> style, or at the end of your message.
>>
>> For example, which Chris are you referring to (Benham?)
>
> Good guess! Is there another "Chris" who has been a regular poster
> here, at any time during my current duration of membership at EM?

Apparently I have waded into a deep discussion that can be apprehended
only by those with deep familiarity with the forum and its
participants.

And those complaining about the lack of connectiona to those earlier
discussions should be met with contemptuous disdain.

> Or were you referring to Kristofer, who has never, so far as I'm
> aware, been referred to here as "Chris"?
>
>>, what does ICT stand for, and where is ICT defined?
>
> For about the hundredth time, ICT stands for Improved-Condorcet-Top.
> Kevin Venzke defined ICA a long time ago, to stand for
> Improved-Condorcet-Approval. Improved-Condorcet-Top is the same,
> except that completion is by top-count instead of Approval-count.
>
> Improved Condorcet is hardly a new term here.
>
> Here is a definition of Symmetrical ICT, which I prefer to ordinary ICT:
>
> (X>Y) means the number of ballots ranking X over Y
> (Y>X) means the number of ballots ranking Y over X.
> (X=Y)T means the number of ballots ranking X and Y in 1st place.
> (X=Y)B means the number of ballots ranking X and Y at bottom
> ....(not ranking X or Y over anything)
>
> X beats Y iff (X>Y) + (X=Y)B > (Y>X) + (X=Y)T
>
> 1. If one candidate beats everyone else, then s/he wins.
>
> 2. If everyone or no one is unbeaten, then the winner is the candidate
> ranked in 1st place on the most ballots.
>
> 3. If some, but not all, candidates are unbeaten, then the winner is
> the unbeaten candidate ranked in 1st place on the most ballots.
>
> [end of Symmetrical ICT definition]
>
> I've posted, at EM, pseudocode for a Symmetrical ICT count program.
>
> I've also posted it at minguo, where it can be found among the recent
> posts, at the USA Realm.

Thank you for making your terms explicit, though by deleting the text
they were referring to, your explanations lack context.

It would have sufficed, in the original message, to add URLs instead
of restating the definitions completely.

Ted
-- 
araucaria dot araucana at gmail dot com