[EM] Jameson: MMT, MTAOC, MJ

[Jameson Quinn]

I honestly don't understand your hostility to me. Can you explain it?

2011/11/29 MIKE OSSIPOFF <nkklrp at hotmail.com>

>
> Jameson--
>
> You had said:
>
> Majority
>  Judgment has similar advantages to MTA in this case.
>
> [endquote]
>
> I asked:
>
> Does it? Who knows?
>
> You replied:
>
> Anyone who takes the time to read the academic literature<https://sites.google.com/site/ridalaraki/majority-judgment>
> .
>
> [endquote]
>
> Translation: You yourself don't know the answer.
>
> MJ's advocates are peculiarly reticent about its specific properties,
> advantages and criterion-compliances.
>
> Alright, I'll answer it for you. MJ does not have advantages similiar to
> those of MTA.
>
>
>
> Have its proponents told what criteria it meets and specifically what
> guarantees it
> offers?
>
> How does it do in the Approval bad-example?
>
>
> You answered:
>
> Same as MTA.
>
> [endquote]
>
> Ok, in other words, MJ doesn't pass in the ABE. Thank you.
>
> You continued:
>
> That is, honest-votes will reliably give a good result, unlike unstable
> Approval; but strategic voting will lead to failure.
>
> [endquote]
>
> Nonsense. That's a remarkably naive statement. In every nonprobabilistic
> method, strategy is advantageous.
>

Sorry, you misunderstood me. I meant that strategic voting would be
advantageous for the voters, and thus would lead to the method failing the
ABE/Chicken Dilemma.

>
> So you think that you've found one for which that isn't so?  :-)
>

No. Actually, it seems to me that you think I'm an idiot and/or an enemy of
yours and so you interpret everything I say in the most uncharitable
possible way. I hope I'm wrong about that impression, because I am neither
of those things.

>
>
>
>
> (to compare it to MTAOC)
>
>
> If you're unwilling to research the published answers to your own
> questions, why do you persist in asking us to look up your alphabet soup in
> old posts? For instance, I know what you mean by MTAOC (a system with a
> strong dishonest-fill incentive
>
> [endquote]
>
> In MTAOC, your middle rating of a lesser-evil (B) can't help the
> lesser-evil beat your favorite (A). And the B-favorite voters reciprocate
> that middle rating, then it's more likely that one or the
> other of those 2 candidates will win, instead of someone (C) less-liked by
> both factions. If it looks as if that would elect A, then the B voters
> don't benefit from middle-rating A
> unless they sincerely prefer A to C.  Likewise if A and B are reversed in
> that sentence.
>

All that is true. And has nothing to do with the problem I pointed out. The
method has a strong dishonest-fill incentive. A Gore>Bush>Nader voter could
in some circumstances elect Gore only by strategically voting Gore>Nader;
but if Nader had slightly higher-than-expected support, then that dishonest
strategy could lead Nader to win even with as little as 25% honest support.

>
> You continued:
>
> , but searching past messages for that acronym just gives the written-out
> name
>
> [endquote]
>
> No, each method's definition was announced in a post whose subject line
> contained the name of that method, written out,
> and, in nearly every case, with the abbreviation in parentheses.
>
> False.

>
>
> What majority-rule guarantees does it offer? Does it meet 3P or 1CM?
>
>
> You replied:
>
> It meets 3P, which I happen to remember what it means.
>
> [endquote]
>
> The 3P complying methods that I'm aware of all make two special
> distinctions, usually a distinction between top rating vs rating below top;
> and between rating vs not
> rating. MJ makes no such distinctions. So, if it meets 3P, then what are
> the two protection-levels required for 3P compliance?
>

Yes it does. Every single rating level in MJ is a "protection level" for
3-level-protection.


>
> You continued:
>
> If you define 1CM <http://wiki.electorama.com/wiki/1CM> I'll tell you if
> it meets that.
>
> [endquote]
>
> Oh, that's ok. _I'll_ tell _you_:  It doesn't. For the reasons stated in
> my paragraph above.
>

You just told me nothing more than if I told you that it meets QWERT or
that it doesn't meet ZXCVB. I suspect that your 1CM is actually a
pernicious criterion that it's better not to pass, but since you prefer to
make unreliable assertions about what methods pass it than actually
re-posting or linking the definition, I'm not going to bother checking.


>
> As I said, I've been meaning to post my definitions in the wikipedia.
> Lately I haven't had a lot of time for computer. But I'm soon
> going to post those definitions.
>
> Ok, it doesn't pass in the ABE, and it doesn't pass 3P or 1CM. Well, if we
> defined a 2P (as compared to 3P), it would pass that, as would Approval
> and RV. Maybe there should be a 2P criterion. It probably passes WDSC too,
> and probably FBC.
>
>
>
> It probably has a strategy situation very much like that of ordinary RV.
> The method of summed scores.
>
>
> You replied:
>
> No. For most voters in real-world studies of MJ, their honest,
> not-even-normalized MJ ballot was strategically optimal. That is clearly
> far better than Range.
>
> [endquote]
>
> Oh, that's different! In MJ, some voters don't have incentive to rate
> other than sincerely :-)
>
> And the strategizing voters won't affect the outcome?
>

Remember, this discussion started when I said MJ was comparable to MTA. You
think you've refuted that by pointing out that MJ doesn't violate the
Gibbard-Satterthwaite theorem?


>
> By the way, if you're going to advocate MJ, it needs a better name.
> Majority Judgement might be ok as a promotional name, but it is not a
> descriptive name.
> How about (the obvious) "Cardinal Median" or "Median Cardinal".
>

A good name allows people to find out what a method means. It is far more
reliable to use wikipedia
<http://en.wikipedia.org/wiki/Majority_Judgment>or the
library <http://mitpress.mit.edu/catalog/item/default.asp?ttype=2&tid=12522>for
that, than just interpolating based on a descriptive name.

Jameson
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20111129/47bd2e85/attachment-0004.htm>