Would you support the CD method if all states used it?

[minionofmidas]

You must be logged in to read this quote.

This is not the method proposed in Colorado. 

Except that it is, just using a fairly ideal (but more realistic than your comparatively contrived scenarios, except for the total absence of minor parties) example.

EDIT: No. No it isn't. Not sure how I missed that. The method Peter proposes can have some pretty effing weird results once you have more than two parties.

You must be logged in to read this quote.

Well, yes.

State-by-state 2004 (comparing Colorado method, D'Hondt, Ste Lague, Hare). (Not complete yet.)
ME 2-2 2-2 2-2 2-2
NH 2-2 2-2 2-2 2-2
VT 2-1 2-1 2-1 2-1
MA 8-4 8-4 8-4 8-4 first time we get a residual seat, although it goes to the Dems under all systems
RI 2-2 3-1 2-2 2-2 first time we get different results
CT 4-3 4-3 4-3 4-3
NY 19-12 18-13 18-13 18-13 a residual seat does something strange. Under Hare we're very close to 18-12-1.
NJ 8-7 8-7 8-7 8-7
PA 11-10 11-10 11-10 11-10
OH 10-10 10-10 10-10 10-10
IN 4-7 4-7 4-7 4-7
IL 12-9 12-9 12-9 12-9
MI 9-8 9-8 9-8 9-8
WI 5-5 5-5 5-5 5-5
MN 5-5 5-5 5-5 5-5
IA 3-4 3-4 3-4 3-4 technically another residual seat, although where it's going is so bleeding obvious on these figures that it hardly matters what you call the seventh seat
MO 5-6 5-6 5-6 5-6
ND 1-2 1-2 1-2 1-2
SD 1-2 1-2 1-2 1-2
NE 2-3 1-4 2-3 2-3 contrary to Jim's claims, Nebraska is far from locked into 3-2. A few more percentage points shift (3.39 under Hare, 3.11 under Sainte-Lague, 2.69 under Colorado) to the right gets it to 4-1 under other systems as well.
KS 2-4 2-4 2-4 2-4
DE 2-1 2-1 2-1 2-1
MD 6-4 6-4 6-4 6-4
DC 3-0 3-0 3-0 3-0
VA 6-7 6-7 6-7 6-7
WV 2-3 2-3 2-3 2-3
NC 7-8 7-8 7-8 7-8 Under D'Hondt, Bush needs just 640 votes more for an extra EV.
SC 3-5 3-5 3-5 3-5
GA 6-9 6-9 6-9 6-9
FL 13-14 13-14 13-14 13-14
KY 3-5 3-5 3-5 3-5
TN 5-6 5-6 5-6 5-6
AL 3-6 3-6 3-6 3-6
MS 2-4 2-4 2-4 2-4
AR 3-3 3-3 3-3 3-3 Yeah, a ten point lead is not enough to get an EV advantage. To get to 4-2 the Reps need, depending on system used, a 14-18 point lead. Think they might have campaigned for that EV? Nobody campaigned in Arkansas in rl.
LA 4-5 4-5 4-5 4-5
OK 2-5 2-5 2-5 2-5
TX 13-21 13-21 13-21 13-21
MT 1-2 1-2 1-2 1-2
ID 1-3 1-3 1-3 1-3
WY 1-2 1-2 1-2 1-2
CO 4-5 4-5 4-5 4-5
NM 2-3 2-3 2-3 2-3 see Iowa
AZ 4-6 4-6 4-6 4-6 another residual seat actually. Swings required to make it 5-5 vary from 0.24 (Hare) to 0.72 (D'Hondt). (St Lague 0.27, Colorado method 0.60). Still, 0.24 is almost 5000 people (to change from R to D - to change from nonvoting to D it's almost twice as much.) Outside of reasonable recount territory unless there's been Florida style ballotrigging madness (alas, can't rule that out. See Arizona 1972.), but certainly attainable by better GOTV etc.
UT 1-4 1-4 1-4 1-4
NV 2-3 2-3 2-3 2-3
WA 6-5 6-5 6-5 6-5
OR 4-3 4-3 4-3 4-3
CA 31-24 30-25 30-25 30-25 A residual seat doing funny things again. Here and in NY, a truth about the Colo. method may be emerging - it may be getting quirkier when more seats are being distributed (though probably only up to a point), unlike the other systems.
AK 1-2 1-2 1-2 1-2
HI 2-2 2-2 2-2 2-2
sum total
Bush 278
Kerry 260
As it turns out, the other three systems all give Bush 280, Kerry 258, but I would not follow that the Colo. system has a systematic pro-Dem bias.
Even the EC was pretty close to these results, at 286-252.
National PR btw is Bush 273, Kerry 260, Nader 2, Badnarik 2, Peroutka 1 (under Hare or Colorado. D'Hondt has Bush 274, Kerry 261, Nader 2, Badnarik 1. Saint-Lague has Bush 273, Kerry 259, Nader 2, Badnarik 2, Peroutka 1, Cobb 1.)

[minionofmidas]

And just for the hell of it, I'll compare the four for the last real three-party election, 1992 (Clinton-Bush-Perot):

ME 2-1-1 2-1-1 2-1-1 2-1-1
NH 2-2-0 2-1-1 2-1-1 2-1-1 fun with Colo's healing mechanism for too many seats.
VT 1-1-1 2-1-0 1-1-1 1-1-1 Clinton had more than twice Perot's vote, but under 50% (and less than three times as much).
MA 6-3-3 6-3-3 6-3-3 6-3-3
RI 2-1-1 2-1-1 2-1-1 2-1-1
CT 3-3-2 3-3-2 3-3-2 3-3-2
NY 17-11-5 17-11-5 17-11-5 17-11-5
NJ 7-6-2 7-6-2 7-6-2 7-6-2
PA 11-8-4 11-8-4 11-8-4 11-8-4
OH 9-8-4 9-8-4 9-8-4 9-8-4 Incidentally, in each of the latter four states there's a residual seat that goes the "right" way.
IN 4-6-2 5-5-2 5-5-2 5-5-2 Undeserved residual seat gain for Bush (Perot as well as Clinton had a higher residual percentage than Bush)
IL 11-8-3 11-8-3 11-7-4 11-7-4 Fun with the removal rule again, although D'Hondt gives the same result anyways
MI 8-7-3 8-7-3 8-7-3 8-7-3
WI 5-4-2 5-4-2 5-4-2 5-4-2
MN 5-3-2 5-3-2 5-3-2 4-3-3 That's a new one - Hare as the odd one out. Residual seat.
IA 3-3-1 3-3-1 3-3-1 3-3-1
MO 5-4-2 5-4-2 5-4-2 5-4-2
ND 1-1-1 1-1-1 1-1-1 1-1-1
SD 1-1-1 1-1-1 1-1-1 1-1-1
NE 1-3-1 1-3-1 2-2-1 2-2-1 Fun with a residual seat, although D'Hondt gives the same result.
KS 2-2-2 2-2-2 2-2-2 2-2-2
DE 1-1-1 2-1-0 1-1-1 1-1-1 See Vermont.
MD 5-4-1 5-4-1 5-4-1 5-4-1
DC 3-0-0 3-0-0 3-0-0 3-0-0
VA 5-6-2 5-6-2 5-6-2 5-6-2
WV 2-2-1 3-2-0 2-2-1 2-2-1 Clinton's got a little over three times what Perot got.
NC 6-6-2 6-6-2 6-6-2 6-6-2
SC 3-4-1 3-4-1 3-4-1 3-4-1
GA 6-6-1 6-6-1 6-5-2 6-5-2 See Illinois. Not that Clinton deserves more EVs than Bush in Georgia.
FL 10-10-5 10-10-5 10-10-5 10-10-5
KY 4-3-1 4-3-1 4-3-1 4-3-1
TN 5-5-1 5-5-1 5-5-1 5-5-1
AL 4-4-1 4-4-1 4-4-1 4-4-1
MS 3-3-1 3-4-0 3-3-1 3-3-1 I find it hard not to consider D'Hondt actually fairer in this case...
AR 3-2-1 4-2-0 3-2-1 3-2-1 I don't find it hard in this one.
LA 4-4-1 4-4-1 4-4-1 4-4-1
OK 3-3-2 3-3-2 3-3-2 3-3-2
TX 12-13-7 12-13-7 12-13-7 12-13-7
MT 1-1-1 1-1-1 1-1-1 1-1-1
ID 1-2-1 1-2-1 1-2-1 1-2-1
WY 1-1-1 1-1-1 1-1-1 1-1-1
CO 3-3-2 3-3-2 3-3-2 3-3-2
NM 2-2-1 2-2-1 2-2-1 2-2-1
AZ 3-3-2 3-3-2 3-3-2 3-3-2
UT 1-3-1 1-3-1 1-2-2 1-2-2 See Nebraska.
NV 2-1-1 2-1-1 2-1-1 2-1-1 Technically a residual seat, actually, that second Clinton seat.
WA 5-4-2 5-4-2 5-3-3 5-3-3 See Illinois.
OR 3-2-2 3-2-2 3-2-2 3-2-2
CA 25-18-11 25-18-11 25-18-11 25-18-11
AK 1-1-1 1-1-1 1-1-1 1-1-1
HI 2-1-1 2-2-0 2-1-1 2-1-1 Bush has more than twice Perot's share.

totals:
235-204-99 Colorado
240-199-99 D'Hondt
236-197-105 Sainte Lague
235-197-106 Hare.
231-202-102-2 Marrou-1 Gritz Hare, applied nationally.
Either way it goes to the House, as, really, it should have.

[minionofmidas]

I've done a lot of snips.

EDIT: No. No it isn't. Not sure how I missed that. The method Peter proposes can have some pretty effing weird results once you have more than two parties.
...
NE 2-3 1-4 2-3 2-3 contrary to Jim's claims, Nebraska is far from locked into 3-2. A few more percentage points shift (3.39 under Hare, 3.11 under Sainte-Lague, 2.69 under Colorado) to the right gets it to 4-1 under other systems as well.
...
AZ 4-6 4-6 4-6 4-6 another residual seat actually. Swings required to make it 5-5 vary from 0.24 (Hare) to 0.72 (D'Hondt). (St Lague 0.27, Colorado method 0.60). Still, 0.24 is almost 5000 people (to change from R to D - to change from nonvoting to D it's almost twice as much.) Outside of reasonable recount territory unless there's been Florida style ballotrigging madness (alas, can't rule that out. See Arizona 1972.), but certainly attainable by better GOTV etc.
...
Bush 278
Kerry 260
...
As it turns out, the other three systems all give Bush 280, Kerry 258, but I would not follow that the Colo. system has a systematic pro-Dem bias.

The claim is not that the Colorado-system has a systematic pro-Democratic bias

Yeah, that was a joke. More aimed at the fact that it somehow ended up being the most Dem-friendly system in 2004 than at anything else.

You must be logged in to read this quote.

This is technically true - obviously so - but the whole point of the exercise was to show how rarely this is relevant.*

You must be logged in to read this quote.

Don't you mean Hare?

You must be logged in to read this quote.

Yeah, but it still came close in 2004, didn't it?

You must be logged in to read this quote.

Why? Did it include some recount provisions or something? (And what does it do if there's two residual seats - both to the largest party, or one each to the two largest parties? This is extremely unlikely to occur in the US, but in as fractured an environment as, say, city of Frankfurt precincts, it's fairly easy to find examples actually. It's even technically possible to contrive of a scenario with two seats too many.)

You must be logged in to read this quote.

Eh... might do that another time.


* In other words:
St Lague
Hare

D'Hondt
 

Colorado









CD method with anti-gerrymandering reform





CD method without











































































































































Current System

 

[minionofmidas]

In the absence of 3rd parties, the Colorado method is the same as St.Lague.

Don't you mean Hare?

In the absence of 3rd parties, St.Lague and Hare are the same.

Yeah, I figured that out last night while trying to fall asleep.

You must be logged in to read this quote.

No single member ED's or multi-member ED's using STV?
[/quote]the list is not exhaustive. Of course there are other conceivable methods.