[EM] Enhanced MMPO fails FBC but conditionally meets expectation FBC.

[MIKE OSSIPOFF]

Your sincere ranking starts with:

1. Favorite
2. Favorite's worst  (that's the candidate with the most votes against 
Favorite)

You and one other person are concerned that you might need Favorite's Worst 
as a compromise, and so you rank him in 1st place, and bury Favorite to 2nd 
place.

Favorite loses to a candidate named "Winner1".

Then someone walks up who says, "Say, don't you two know that MMPO meets 
FBC? You don't need to bury your favorite."

You say, "Thanks for the information. Let's both move Favorite to 1st place, 
and have another count!"

So the two of you move Favorite to 1st place, and it turns out that Favorite 
was only one vote below Winner1, and so now Favorite wins.

But this is MMPO with AERLO. There are four people who have Winner1 above 
their AERLO line and Favorite below their AERLO line. Their rankings  
(identical to eachother) look like this:

1. Your last choice's worst
2. Your last choice
3. Winner1
4. Winner1's worst
AERLO line
5. Favorite

Because you've activated AERLO for those four voters, their top four 
candidates are all promoted to 1st place. Winner is already ranked over 
Winner1's Worst, and so moving them all to 1st place has no effect on 
Winner1`s worst vote against him. But Your Last Choice is ranked below Your 
Last Choice's Worst, and so, when they're all promoted to 1st place, Your 
Last Choice loses a vote-against for each of those 4 voters. And it happens 
that Your Last Choice was only 1 vote behind Favorite's initial score. So 
now Your Last Choice wins.

By moving your favorite to 1st place, you made your last choice win, where 
your 2nd choice would have otherwise won.

But what if AERLO results situation with your 1st and 2nd choice were 
reversed. Then, it would by by _burying_ your favorite that you'd make your 
last choice win.

And how likely would it be that you'd have information that one of those 
situations would be more likely than the other?

Because you have no information about that, that AERLO result has no 
contribution to your expectation. And ranking your favorite in 1st place has 
an obvious positive contribution to your expectation.

So, conditionally meeting expectation FBC is nearly as good as actually 
meeting outcome FBC .

What this means is that, without AERLO, MMPO strictly meets outcome FBC. 
But, with AERLO,  MMPO meets FBC in the same way that it meets SPCA, ASPCA & 
CSPCA: It conditionally meets   expectation FBC.

As I was saying, plain MMPO, with its super-brief definition, is the best 
initial public proposal. It's also good, with that first reform proposal, to 
be able to assure people of strict compliance with FBC.  Later, when AERLO, 
or AERLO, ATLO and power truncation are proposed, it would be explained that 
then there would still be no reason to bury a favorite, even though FBC 
compliance is no longer strict. But that, with those enhancements, other 
conditional compliances with valuable expectation criteria would be gained.

How does plain MMPO compare with Approval?

On the downside, MMPO doesn't let you vote against your unacceptable set as 
strongly as Approval does. On the positive side, MMPO meets SFC and WDSC.

The value of MMPO meeting SFC is in question, though, because, with 
strategic voters, MMPO doesn't guarantee safety for the sincerity stipulated 
by SFC. As I said, voters might know that they're in the majority referred 
to in SFC, but if they had that kind of informtion, they'd do fine with 
Approval. So FBC probably can't be counted for plain MMPO. But WDSC can.

So plain MMPO has FBC and WDSC, as does Approval. But plain MMPO doesn't let 
voters vote against their unacceptable set as strongly as Approval does.

So MMPO isn't quite as good as Approval. But, for one thing, many people 
would be more interested in a rank-balloting method than in Approval. 
Additionally, people tend to have the fallacious concern about 
1-person-1-vote with Approval, but wouldn't have that problem with MMPO. 
Further, once MMPO is enacted, the enhancements could be proposed beginning 
in the next election.

Adding power truncation would make MMPO as good as Approval. Adding only 
AERLO would bring conditional SPCA compliance. Adding power truncation to 
that would get ASPCA. Then, adding ATLO would get CSPCA.

What about MMPO with only AERLO vs Approval? The top-end advantages brought 
to MMPO by AERLO have strategic importance, assuring people about not 
needing preference-concealment, and that seems more important than 
Approval's stronger vote against the unacceptable candidates.  After all, 
the opposition also has less strong vote against its unacceptable 
candidates.

Mike Ossipoff

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