[EM] Voting Theory and Populism clarification

[Raph Frank]

On Mon, Oct 20, 2008 at 7:15 AM, Kevin Venzke <stepjak at yahoo.fr> wrote:
> Consider these sincere preference orders:
>
> 40 A>B>C
> 20 B>A>C
> 40 C>B>A
>
> I can't see any reason why it couldn't be expected that only A will
> receive majority approval, but B would defeat A in a runoff.

Assume they use the 'approve the best of the top 2 + anyone you prefer
to the expected winner'.

Option 1) B + A(2nd) expected top 2

40: A
20: B
40: B + C

A: 40
B: 60
C: 40

Option 2) B+C(2nd) expected top 2

40: A+B
20: B
40: C

A: 40
B: 60
C: 40

In both cases, B is the only one with a majority.  The strategy would
likely switch between the 2, and it is unlikely to be perfectly
balanced 40-40 between A and C.