[EM] A strategy free method

[Alex Small]

Forest Simmons said:
> You're right, but like approval it does satisfy the FBC.

Say that the status quo, A, is my favorite, but he loses to candidates B
and C.  His most overwhelming defeat is at the hands of C, my least
favorite.  I can't do anything more to help A defeat B and C pairwise, but
I can at least try to give B a stronger victory over A.

Example:

40 A>B>C
 5 C>A>B
25 B>C>A
30 C>B>A

B beats A 55:45
C beats A 60:40

The issue of margins vs. winning votes is irrelevant in this example, as C
has the strongest victory over A by either measure.  C wins.

The people in the A>B>C faction have an incentive to vote B>A>C to improve
B's victory over A and hence elect B instead of C.

Now add the possibility of equal rankings, and margins vs. winning votes
matters:

If 11 people in the A>B>C faction instead rank A=B>C, we have:

B beats A 55:34
C beats A 60:40

With winning votes, B's victory is still the weaker one (55 vs. 60).  With
margins, B's victory is the stronger one (21 vs. 20).  OK, I guess it does
pass the weak FBC if we use margins.

(Note:  Since margins vs. winning votes can be a touchy issue on this
list, I am not issuing an across-the-board opinion on that matter, but
merely observing that in one particular case margins enables a particular
method to satisfy a certain property.)

Still, this satisfaction of weak FBC comes at the expense of neutrality. 
Oh, well, it's an interesting example from a theoretical standpoint.  I
guess there is more than one ranked method that satisfies weak FBC (I
thought "top 2 voting" was the only ranked method to satisfy FBC).



Alex