[EM] proportional constraints - help needed

[Raph Frank]

On Wed, Feb 6, 2013 at 9:47 AM, Peter Zbornik <pzbornik at gmail.com> wrote:
> James, Jonathan,
>
> I need that the quoted-in people are quoted-in in such a way, that the
> proportionality of the election is not significantly disturbed.

James Gilmour has the right idea.

Elect 5 seats, but don't eliminate a candidate if there aren't at
least 3 of that gender uneliminated (so you might eliminate the 3rd
lowest, if the other 2 were "protected")

That means that at the end there will be at least 2 of each gender in
the set of 5 winners.

The remaining "elections" only have 1 elimination

Elect 4 out of those 5.  When all surpluses have been transferred, the
lowest candidate of the majority gender is given rank 5

Elect 3 out of remaining 4.  When all surpluses have been transferred,
the lowest candidate is given rank 4

Elect 2 out of remaining 3.  When all surpluses have been transferred,
the lowest candidate of the majority gender is given rank 3

Do a head to head runoff between the last 2 based on the rankings, the
winner gets rank 1 and the loser gets rank 2.

You could end up with

M - W - W - M - M

It isn't guaranteed that every 2nd will be a man and a woman, but the
max imbalance is 1.

top-1: 1 man and no women
top-2: 1 man and 1 woman
top-3: 1 man and 2 women
top-4: 2 men and 2 women
top-5: 3 men and 2 women

In all cases, the imbalance is limited to 1.