[EM] Strategy details, cont'd.

[Janet Robinson]

First a few clarifying comments about MMPO:

In the standard chicken dilemma example, I'm not saying that the other  CD methods are wrong to choose C.  I'm just saying that there's nothing unqualified about A.

Yes, MMPO doesn't strictly meet MMC.

But failure needs a top-cycle. Not just any top-cycle. One in which every mutual majority preferred set (MMPS) member has big majority pairwise opposition.

In political polls at CIVS, I'm not aware of any top-cycle, ever.

Engineering a strategic one would be prohibitively difficult to organize.

MMPO MMC failure isn't a realistic problem.

I call MMPO the best unlimited-rankings method.

It meets CD. 

Its strategy in ordinary (non chicken dilemma) situations is MAM-like, and that's saying a lot.

It effectively meets MMC.

It meets FBC, meaning that it offers the option if fully effective approval votling.

Regarding MAM:

As I mentioned, as in Bucklin, if you aren't majority favored, you depend on people not ranking past the CWse.

But their doing so is harmless unless burial is attempted.

At worst if would be like CWs misjudgment in Bucklin.

But what if there's a faction who are nearly indifferent between the CWs and their next-favorite? They aren't deterred from burial. Well, then it's no worse than Bucklin when people mistakenly vote past the CWs. 

What if, by mistake or by not caring, a faction bury when people are correctly not voting past the CWse?

Then they elect someone they like less than the CWs.

Will that draw criticism? Well, the perps can't complain--They did it.

The people opposite the CWs will like that result. Only the CWs preferrers would object. There's no objecting majority.

In any case, these failures to elect the CWs, due to misjudgment, ark no worse than misjudgment results in Bucklin or other methods.

MAM is the best unlimited-rankings method after MMPO.

Michael Ossipoff