[EM] dissatisfying Strong FBC methods

[Alex Small]

Kevin Venzke said:
> I don't agree that this method should be considered to meet strong FBC.
> I think we can safely declare that it fails it, because even when the
> ballot A>B>C is the sole ballot in the election, the result is not an A
> victory but an AB tie.
>
> We could modify the method to break such ties by counting first
> preferences, but then of course there would be order-reversal incentive.

An easier way to modify it is to break ties with a pairwise comparison. 
In that case there's no incentive to betray your favorite.  Say there's a
tie between your favorite and second favorite:  If the pairwise comparison
gives the victory to your second favorite, you can always break the tie by
insincerely ranking your second favorite in last place.  Your favorite
then wins, and we don't need the pairwise comparison.  Or, if the tie is
between your favorite and least favorite, there's no incentive to betray
your favorite because doing so won't break the tie or change the outcome
of the pairwise comparison.