To me, this appears workable. Am I missing something?:

Do MAM, using Kevin's definition of "beat" (the one that his ICA method
uses).

Advantage:

Ranking one's entire approval-set at top would give maximum protection to
those candidates, & would always be  optimal if your goal is to elect from
your approval set.

People seem to want a rank
method.  ...& will want to rank in sincere order.

With MAM, as with any MMC complying method, if your approval-set (or at
least part of it) has a majority preferring them to the other candidates,
and your goal is to elect from your approval-set, then you can safely rank
them sincerely.

But otherwise,  if you want to elect from your approval-set, then you'd
better approval vote them, top-rank them all, & have that
effectively-counted.

Hence the desirability of IC MAM.

I like the Bucklin version that doesn't allow equal-ranking or skipping in
rankings,  but allows the option of top ranking as many candidates as you
want, and not using any other rank-position.

...if you opt to do that instead of voting a ranking.

With that method too, you'd better approval vote, if some of your approval
set isn't preferred to all others by a majority.

IC MAM brings to MAM what equal top ranking does in Bucklin.

Michael Ossipoff

Hi Mike, I'm a little rusty (having proposed ICA over 10 years ago), but I feel confident that this wouldn't completely satisfy FBC, even for the approval-style voters. The effects of path tracing (in MAM or Schulze etc.) are too arbitrary/unknowable to offer the guarantee. The method seemingly needs to be very simple. I would certainly have been eager to devise something more complicated/sophisticated than ICA (but with similar properties), but I have not been able.
Kevin

  De : Michael Ossipoff <email9648742@gmail.com>

À : election-methods@electorama.com
Envoyé le : Jeudi 1 septembre 2016 20h05
Objet : [EM] IC MAM?

To me, this appears workable. Am I missing something?:Do MAM, using Kevin's definition of "beat" (the one that his ICA method uses).Advantage:Ranking one's entire approval-set at top would give maximum protection to those candidates, & would always be  optimal if your goal is to elect from your approval set.People seem to want a rank
method.  ...& will want to rank in sincere order.With MAM, as with any MMC complying method, if your approval-set (or at least part of it) has a majority preferring them to the other candidates, and your goal is to elect from your approval-set, then you can safely rank them sincerely.
But otherwise,  if you want to elect from your approval-set, then you'd better approval vote them, top-rank them all, & have that effectively-counted.Hence the desirability of IC MAM.I like the Bucklin version that doesn't allow equal-ranking or skipping in rankings,  but allows the option of top ranking as many candidates as you want, and not using any other rank-position....if you opt to do that instead of voting a ranking.With that method too, you'd better approval vote, if some of your approval set isn't preferred to all others by a majority.IC MAM brings to MAM what equal top ranking does in Bucklin.Michael Ossipoff

Election-Methods mailing list - see http://electorama.com/em for list info