FPP with equal rankings

[Bart Ingles]

The Equal-ranked IRO example suggests an unranked method with properties
in between FPP and Approval.  Example:

14 A
10 (A,B)
10 B
10 (B,C)
15 C
________
59 total

Under Approval, 30B > 25C > 24A; B wins.

Under Conditional Plurality, B is Dropped; 25C > 24A; C wins.

Although the (A,B) voters are unable to help the (B,C) voters, the
method may overcome FPP's equilibrium problem.  After an election with
results similar to the one above, it would be publicly known that B is a
potential winner; many B supporters who included A or C as a lesser evil
will likely refrain from doing so in the future.

Note that this method preserves one-man-on-vote and the majority winner.

Any problems found with this method will likely carry over to
Equal-ranked IRO.  Likewise, so might any improvements.

One such improvement might be:  rather than ignoring equal-ranked votes
so long as they remain equal, why not treat them as fractional votes
until only one member of the set remains?  You would give the whole vote
back to the surviving member when the others drop out.  This would also
preserve one-man-one-vote, but would retain the advantages of
conditional usage.  Example:

14 A
10 (A,B)
10 B
10 (B,C)
15 C
________
59 total

Initially:  (14 + 5)A, (5 + 10 + 5)B, (5 + 15)C
  20B = 20C > 19A;  drop A ---(note total is still 59)
You now have: (20 + 5)B, (5 + 15C)
  25B > 20C;  B wins.