[EM] Meek-based maths for equal ranked PR-STV

[Dan Bishop]

Raph Frank wrote:
> Just expanding on concept from the other PR-STV thread.  There is an
> issue of how exactly to handle equal rankings.
>
> One suggestion is made by Warren in voting matters
>
> http://www.votingmatters.org.uk/ISSUE7/P5.HTM
>   
[The linked page gives an example in which a 3-seat election with the 
ballots

1: A>B
1: B>A
9: C>A>B
1: C>E>F
9: D>B>A
1: D>E>F
3: E>F
4: F
1: A=B

gives a worse outcome (C+D+F) to the A=B voter than if they had voted 
either A>B (C+D+A would win) or B>A (C+D+B would win).]

That's a rather interesting scenario.

But maybe it's not a problem with the treatment of A=B votes as much as 
a quirk of plurality loser elimination. Alternative elimination methods 
could elect A or B even with the last ballot interpreted as 0.5 A>B + 
0.5 B>A. For the third seat,

Random Candidate Elimination elects A with probability 3/8 or B with 
probability 3/8 (or F with probability 1/4).

STV-CLE ends in a tie between A and B.

Borda elimination, however, elects F.