[EM] STV+AV

[Raph Frank]

On Mon, Feb 6, 2012 at 3:57 AM, David L Wetzell <wetzelld at gmail.com> wrote:
> Moreover, if the bicameral state legislatures were selected by both LR Hare
> 3-seats and a single-winner rule (insert your favorite here), then it'd make
> it so that what helped with gerrymandering in one branch would hurt in the
> other branch.

You mean party list combined with 3 seat districts?

What do you mean opposite gerrymandering?

With 3 seats, the 2 targets are 50% and 75%.

Standard gerrymandering would aim to have have lots of districts with
75%+ support for the minority party.

The ideal for the minority party would be to make it so that when they
win a district, they have ~70%.  Similarly, when they lose a district,
they should have ~45%.

That means that their cost per seat for the PR method would be

Minority party:

2 seats for 70% = 35% per seat
1 seat for 45% = 45% per seat

Majority party

2 seats for 55% = 27.5% per seat
1 seat for 30% = 30% per seat

Thus the majority party has a lower cost per seat.

However, it would increase the accuracy needed for gerrymandering and
risks a landslide against if things shift by more than a few percent
from the estimates (as happened with the Tullymander).

> In fact, it might be a good thing to let the pretty darn proportionally
> elected state house of reps elect our US senators again!!!

The original point was that the Senate represented the States.  Also,
the original rules require that it is the legislature that picks the
Senator.

This presumably meant that the Senate was more reluctant to centralize
power, since it would be reducing the power of its electors.

I wonder if a State would have been allowed to say that its House of
Representatives is its legislature (even if there was a 2nd House
required for bills to pass).