[EM] NPV vs Condorcet

[Raph Frank]

Another option would be something proportional completion of the ballots/

If the votes are

A: 50
B: 30
C: 20

Assume, B is preferred to C by 3/5 (30/(20+30)) of the A voters

30: A>B>C
20: A>C>B

Assume, A preferred to C by 5/7 (50/(50+20))of B supporters

21: B>A>C
9: B>C>A

Assume, B preferred to A by 3/8 (30/(50+30)) of C supporters

12: C>A>B
8: C>B>A

This gives the right proportions of first choice votes.  Ofc, it
assumes that second choices follow the first choices, and I have no
idea what the strategic incentives are.