[EM] Chris: MMT's strategy incentives

[MIKE OSSIPOFF]

Random fill:

 

 

First of all, here it isn’t random, for two reasons:

 

One: There is only one level of intermediate support. This
is a 3-slot method. So there’s no need for random choice of which

Intermediate position to give to a candidate.

 

Two: The matter of whom to middle-rate isn’t decided
randomly. If your only purpose is the maximization of your favorite’s win, then
you should middle-rate lots of small candidates, adding up to lots of
voter-support, but with none of them having many top-ratings. 

 

Of course that helps only if their supporters want to elect
your (larger) candidate as a compromise. 

 

In other words, supporters of less popular candidates choose
to support and elect a more popular compromise. 

 

If you know that your candidate is Plurlaity winner (and if
your only interest is electing hir), then you should middle-rate all of the
other candidates (at least the ones who alren’t potential Plurality rivals).,
thereby asking for membership in any and all mutual majority coalitions. Their
supporters will include you in their mutual majority coalitions (by
middle-rating you) only if they genuinely want to elect your candidate as a
compromise, more popular then their candidates, and better than the others.

 

Again:

 

In other words, supporters of less popular candidates choose
to support and elect a more popular compromise.

 

Burial:

 

Bottom-rating a candidate only has the effect of not
inviting hir to a coalition. How offensively aggressive is that?

 

If you’re only interested in electing your favorite, and you
think you can, then don’t middle-rate any candidate who might rival your
favorite’s top-ratings total. That isn’t devious burying. It’s the common-sense
of not giving to rivals.

 

Summary:

 

The expressions “random-fill” and “burial” are intended for other
methods, and don’t apply to MMT.

 

MMT doesn’t have the problems that those terms refer to.

 

Forest—

 

Could LRV, due to the bottom symmetrical completion,
sometimes have an AB E-like problem, if the numbers were somewhat different
from those of the usual ABE? Could it have a co-operation/defection problem for
that reason?

 

Mike Ossipoff