Fair redistricting: Texas

[True Federalist (진정한 연방 주의자)]

Considering how convoluted and arbitrary our county borders are, giving preeminence to avoiding county chops might work in a mathematical sense, but not when it comes to putting reasonable districts together. Of the ten SC finalists, muon2 A was the worst in my opinion, not one of the two best. That second district was particularly heinous even tho it resulted from having the whole county districts 5, 6, and 7.

[Torie]

Considering how convoluted and arbitrary our county borders are, giving preeminence to avoiding county chops might work in a mathematical sense, but not when it comes to putting reasonable districts together. Of the ten SC finalists, muon2 A was the worst in my opinion, not one of the two best. That second district was particularly heinous even tho it resulted from having the whole county districts 5, 6, and 7.

What metrics would you use in lieu of county chops? Muon2 and I concluded long ago, that focusing on communities of interest was so subjective and being subject to being gamed, that it was just not a functional metric to use. All it did was encourage hacks to spew out GIGO in their testimony. If you ignore counties, the only more objective metric left is erosity (and UCC areas if you care about those). One practical consideration, is that it is a lot easier to prepare ballots, if you have districts that track political jurisdictions. The more chops you have, the more "chopped" ballots you have. That is why, going down to the micro level, when I chop towns here in Columbia County, I try to track voting district (precinct) lines.

[cvparty]

Considering how convoluted and arbitrary our county borders are, giving preeminence to avoiding county chops might work in a mathematical sense, but not when it comes to putting reasonable districts together. Of the ten SC finalists, muon2 A was the worst in my opinion, not one of the two best. That second district was particularly heinous even tho it resulted from having the whole county districts 5, 6, and 7.

i feel you :/

[True Federalist (진정한 연방 주의자)]

Considering how convoluted and arbitrary our county borders are, giving preeminence to avoiding county chops might work in a mathematical sense, but not when it comes to putting reasonable districts together. Of the ten SC finalists, muon2 A was the worst in my opinion, not one of the two best. That second district was particularly heinous even tho it resulted from having the whole county districts 5, 6, and 7.

What metrics would you use in lieu of county chops?

In general avoiding county chops are a good idea, but South Carolina is a State that helps show that generalities don't always work.  A number of our smaller counties were carved out specifically as gerrymanders back in the days when our General Assembly was apportioned by county. McCormick is the most egregious case of that, but far from the only one.  We've also got several non-contiguous counties as far as road networks are concerned. (Colleton, Georgetown, and Lexington all have bits you can only get to by car if you drive through another county. I'm not even going to count that small smidgen of Charleston County that's like that because I believe no one actually lives in the cut-off section.

I would say that metrics can help weed out bad districts, but they aren't that useful when it comes to deciding which district is the best.  For instance, I'm certainly not upset by my B-plan not even making the finals.  It was purposefully intended to be a reasonable-looking gerrymander rather than a "fair" plan.

[muon2]

Considering how convoluted and arbitrary our county borders are, giving preeminence to avoiding county chops might work in a mathematical sense, but not when it comes to putting reasonable districts together. Of the ten SC finalists, muon2 A was the worst in my opinion, not one of the two best. That second district was particularly heinous even tho it resulted from having the whole county districts 5, 6, and 7.

What metrics would you use in lieu of county chops?

In general avoiding county chops are a good idea, but South Carolina is a State that helps show that generalities don't always work.  A number of our smaller counties were carved out specifically as gerrymanders back in the days when our General Assembly was apportioned by county. McCormick is the most egregious case of that, but far from the only one.  We've also got several non-contiguous counties as far as road networks are concerned. (Colleton, Georgetown, and Lexington all have bits you can only get to by car if you drive through another county. I'm not even going to count that small smidgen of Charleston County that's like that because I believe no one actually lives in the cut-off section.

I would say that metrics can help weed out bad districts, but they aren't that useful when it comes to deciding which district is the best.  For instance, I'm certainly not upset by my B-plan not even making the finals.  It was purposefully intended to be a reasonable-looking gerrymander rather than a "fair" plan.

This is one of the reasons I spent a lot of time scoring the SC maps - it does provide some interesting tests of the metrics. One observation I would make is that the problem isn't about the chops, but the size of the chops. The recommended factors for the exercise include "County and especially town splits are kept small and reasonable". This matches the muon rule distinguishing ordinary chops from deep macrochops into a county (I use 5% as a threshold). Yet none of the plans except jimrtex's and mine strove to make chops small and reasonable, and the scores reflect that.

A scoring element that used to exist was a microchop, a chop less than 0.5% of the quota, and when it existed it added an extra reward for keeping shops so small. It seems like this would be useful in cases like Edisto. The danger was that microchops could potentially be gamed, so it would take some investigation to see if there were situations where they were warranted.

[cvparty]

1: R+16
2: D+5
3: R+21
4: D+29
5: D+30
6: R+18
7: D+1
8: R+14
9: R+31
10: R+19
11: EVEN
12: D+2
13: D+6
14: R+28

[muon2]

Here's my entry A for GA. It strives for low erosity CDs that chop the minimum number of counties, but did not consider racial or political data. The range is 0.996%, just under a 1% limit. Though the VRA was not considered here, at least 2 CDs would likely elect the candidate of choice for the black population, and three other CDs provide crossover opportunities for the minority.

Only Fulton and Gwinnett are chopped. In Fulton, Atlanta, Sandy Springs, Johns Creek, Roswell, Mountain Park, East Point and Hapeville make up one CD with no chops. A small chop into unincorporated south Fulton is used to equalize population in CD 6. The Gwinnett chop keeps cities whole to the extent permitted by the voting districts. The Atlanta UCC is 7.08 times the quota, and this plan covers it with the minimum of 8 CDs and nests 6 CDs entirely within the UCC. No other UCC is chopped.



CD 1: +2025; R+5.9; BVAP 30.4%
CD 2: -3520; R+1.4; BVAP 41.8%
CD 3: +1547; R+7.6; BVAP 35.3%
CD 4: -82; D+29; BVAP 51.5%
CD 5: -216; D+16; BVAP 38.7%
CD 6: -383; R+3.9; BVAP 23.5%
CD 7: +2403; D+0.5; BVAP 23.1%
CD 8: +3373; R+20; BVAP 26.4%
CD 9: -19; R+31
CD 10: -223; R+4.9; BVAP 27.7%
CD 11: -804; R+26
CD 12: -1366; R+15; BVAP 24.6%
CD 13: +500; D+11; BVAP 48.3%
CD 14: -3232; R+18; BVAP 20.2%

[muon2]

Here is my plan B. It is designed to provide the black minority 4 CDs where they would be likely to elec the candidate of choice. The range is within 1%. Only three counties are chopped, and Fulton is chopped by three CDs (including small parts from CD 10 and 13) for a total of 5 county chops. In addition there are two cover penalties for the Atlanta UCC.



CD 1: -1299; R+8.3; BVAP 29.3%
CD 2: +314; D+6.8; BVAP 47.3%
CD 3: +1218; R+23
CD 4: -82; D+29; BVAP 51.5%
CD 5: -1358; D+24; BVAP 51.2%
CD 6: -3897; R+4.1; BVAP 23.1%
CD 7: -2703; R+0.3; BVAP 21.3%
CD 8: -803; R+18; BVAP 25.8%
CD 9: +2543; R+26
CD 10: +962; R+15; BVAP 27.5%
CD 11: +2351; R+21
CD 12: +1510; R+2.3; BVAP 33.8%
CD 13: -699; D+14; BVAP 49.7%
CD 14: +1946; R+31

[True Federalist (진정한 연방 주의자)]

Considering how convoluted and arbitrary our county borders are, giving preeminence to avoiding county chops might work in a mathematical sense, but not when it comes to putting reasonable districts together. Of the ten SC finalists, muon2 A was the worst in my opinion, not one of the two best. That second district was particularly heinous even tho it resulted from having the whole county districts 5, 6, and 7.

What metrics would you use in lieu of county chops?

In general avoiding county chops are a good idea, but South Carolina is a State that helps show that generalities don't always work.  A number of our smaller counties were carved out specifically as gerrymanders back in the days when our General Assembly was apportioned by county. McCormick is the most egregious case of that, but far from the only one.  We've also got several non-contiguous counties as far as road networks are concerned. (Colleton, Georgetown, and Lexington all have bits you can only get to by car if you drive through another county. I'm not even going to count that small smidgen of Charleston County that's like that because I believe no one actually lives in the cut-off section.

I would say that metrics can help weed out bad districts, but they aren't that useful when it comes to deciding which district is the best.  For instance, I'm certainly not upset by my B-plan not even making the finals.  It was purposefully intended to be a reasonable-looking gerrymander rather than a "fair" plan.

This is one of the reasons I spent a lot of time scoring the SC maps - it does provide some interesting tests of the metrics. One observation I would make is that the problem isn't about the chops, but the size of the chops. The recommended factors for the exercise include "County and especially town splits are kept small and reasonable". This matches the muon rule distinguishing ordinary chops from deep macrochops into a county (I use 5% as a threshold). Yet none of the plans except jimrtex's and mine strove to make chops small and reasonable, and the scores reflect that.

A scoring element that used to exist was a microchop, a chop less than 0.5% of the quota, and when it existed it added an extra reward for keeping shops so small. It seems like this would be useful in cases like Edisto. The danger was that microchops could potentially be gamed, so it would take some investigation to see if there were situations where they were warranted.

Yet following those rules to mathematical precision is precisely what led your map and jimrtex's maps to be so ugly and led to them breaking apart communities of common interest because of the math.  It also does sort of skew the selection process in favor of those who are familiar with your arcane calculations. I readily admit to not being familiar with them, nor do I care to learn. I did this because I was interested in redistricting South Carolina, which is why I probably won't be doing any other States, nor had I planned to do so.

Also, when it comes to South Carolina municipal boundaries, the less attention paid to them, the better, as they typically make those of the State of West Virginia seem compact by comparison.

[muon2]

Also, when it comes to South Carolina municipal boundaries, the less attention paid to them, the better, as they typically make those of the State of West Virginia seem compact by comparison.

I thought that school district boundaries made more sense than municipalities in SC, so I used them.

[True Federalist (진정한 연방 주의자)]

Also, when it comes to South Carolina municipal boundaries, the less attention paid to them, the better, as they typically make those of the State of West Virginia seem compact by comparison.

I thought that school district boundaries made more sense than municipalities in SC, so I used them.

Agreed. Tho in some cases, you'd get better results from older boundaries such as in Orangeburg County which consolidated from eight to three districts about two decades ago.

[jimrtex]

Considering how convoluted and arbitrary our county borders are, giving preeminence to avoiding county chops might work in a mathematical sense, but not when it comes to putting reasonable districts together. Of the ten SC finalists, muon2 A was the worst in my opinion, not one of the two best. That second district was particularly heinous even tho it resulted from having the whole county districts 5, 6, and 7.

What metrics would you use in lieu of county chops?

In general avoiding county chops are a good idea, but South Carolina is a State that helps show that generalities don't always work.  A number of our smaller counties were carved out specifically as gerrymanders back in the days when our General Assembly was apportioned by county. McCormick is the most egregious case of that, but far from the only one.  We've also got several non-contiguous counties as far as road networks are concerned. (Colleton, Georgetown, and Lexington all have bits you can only get to by car if you drive through another county. I'm not even going to count that small smidgen of Charleston County that's like that because I believe no one actually lives in the cut-off section.

I would say that metrics can help weed out bad districts, but they aren't that useful when it comes to deciding which district is the best.  For instance, I'm certainly not upset by my B-plan not even making the finals.  It was purposefully intended to be a reasonable-looking gerrymander rather than a "fair" plan.

This is one of the reasons I spent a lot of time scoring the SC maps - it does provide some interesting tests of the metrics. One observation I would make is that the problem isn't about the chops, but the size of the chops. The recommended factors for the exercise include "County and especially town splits are kept small and reasonable". This matches the muon rule distinguishing ordinary chops from deep macrochops into a county (I use 5% as a threshold). Yet none of the plans except jimrtex's and mine strove to make chops small and reasonable, and the scores reflect that.

A scoring element that used to exist was a microchop, a chop less than 0.5% of the quota, and when it existed it added an extra reward for keeping shops so small. It seems like this would be useful in cases like Edisto. The danger was that microchops could potentially be gamed, so it would take some investigation to see if there were situations where they were warranted.

There is nothing that says that a state could not adjust county boundaries to better reflect communities of interest. But that should be done now, rather than while a redistricting map is being drawn. In the latter case, this would simply result in a lot of post hoc rationalization.

One of the things that happened in Florida is that a map would be drawn. Comments from public hearings would then be studied to provide a rationale for districts. In some cases, public comments were made up or assigned to a voter without their knowledge. If someone said that they lived in St. Petersburg, but attended an AME Church in Tampa, this could be used to provide a rational for drawing a cross-bay district.

So for example, Grand Isle, LA could be treated as separate "parish" from the remainder of Jefferson. Cities in Ohio could be treated as being in the county with the major share of their population. If you are concerned about splitting communities of interest, you should first avoid splitting houses, then blocks, wards, cities, and finally counties. It is better to split a county around a city, Discontiguous political entities can be treated as whole. The portion of Chicago that is in DuPage county could be treated as part of Cook County. The whole of Edisto Island could be treated as being part of Charleston.

[True Federalist (진정한 연방 주의자)]

I sincerely hope you mean that for redistricting purposes the States could define "municipalities" and "counties" rather than actual changes in boundaries.  Those won't happen. (Or rather if they do, it won't be for anything as esoteric as redistricting, but rather for purposes of gaming tax rates and revenue as is already the case.)

[KingSweden]

1: R+16
2: D+5
3: R+21
4: D+29
5: D+30
6: R+18
7: D+1
8: R+14
9: R+31
10: R+19
11: EVEN
12: D+2
13: D+6
14: R+28

I LOVE this map!

[cvparty]

-snip-

I LOVE this map!

fanks!!

[jimrtex]

I sincerely hope you mean that for redistricting purposes the States could define "municipalities" and "counties" rather than actual changes in boundaries.  Those won't happen. (Or rather if they do, it won't be for anything as esoteric as redistricting, but rather for purposes of gaming tax rates and revenue as is already the case.)

Yes, I should have made that clearer.

In Texas, school districts were originally organized within a county. These "dependent" school districts did not have their own school board, and would not operate a high school. Since there weren't buses or cars, students had to be able to walk or ride a horse to school, which might be a one or two-room classroom. All a school district needed was a schoolhouse, a teacher or two, and a few dozen students. There might be a high school in a town. Exceptional students might board in town so that they could go to high school.

As motorized transportation became more available, school districts consolidated. The consolidation often followed transportation patterns. If farm kids could attend school five miles away across a county line it made more sense to go there rather than to the county seat fifteen miles away, particularly if the highways had developed in a different pattern than county based. In most of the West, counties were created before there was much settlement. Settlement would be based on a source of water, or proximity to an oil field, etc.

So in Texas, while you can generally associate school districts with a county, they often include territory from adjacent counties. If you were going to use school districts as secondary building blocks, you would want to nest them in "counties*" where the asterisk indicates that they weren't identical with the legal counties.

The same thing appears to have happened in South Carolina. While Greenville has one school district, in laps over into adjacent counties, and districts from adjacent counties lap over into Greenville.

[cvparty]

EDIT: HERE'S A BETTER MAP

1: R+16
2: R+33
3: D+1
4: R+27
5: R+19
6: D+7
7: R+16

[muon2]

Why the chop of 2 counties between CD 4 and 5. You should never need more than one chop.

[muon2]

The VRA may not be in play,, but I thought this might be useful to identify the black community of interest, particularly in rural counties.




Yellow 25.0 - 33.3% BVAP
Lime 33.4 - 39.9% BVAP
Green 40.0 - 49.9% BVAP
Dark green over 50.0% BVAP

[Since I'm the mad scientist proclaimed by myself]

It's may be a little too similar to CVpartys map. IDK. EDIT: Why did I put that?

 I don't have much time so I'll keep this brief. The first was a VRA district at D+7.76. The third unites SW AL and the Montgomory suburbs. It has to have that hook/horseshoe shape to get a VRA district that doesn't tendril out into Birmingham. I did the extra splits in the 4th and 5th districts for the sake of appearances. The 6th gets most of the area around Huntsville.

[President Punxsutawney Phil]

Why the chop of 2 counties between CD 4 and 5. You should never need more than one chop.

I think he was trying to pack AL-05 with Democrats.
But there are better ways of doing that.

[President Punxsutawney Phil]

Colbert+Franklin+Lawrence+the AL counties along the TN border: one CD.

[Dr. MB]

AL-01: R+16
AL-02: R+18
AL-03: R+26
AL-04: R+34
AL-05: R+16
AL-06: EVEN
AL-07: D+7

Colbert+Franklin+Lawrence+the AL counties along the TN border: one CD.

I basically did that, but to get one with low deviation you need to add a couple precincts to a neighboring CD.

[muon2]

This was drawn in 2013 on a thread that featured a number of maps trying to balance the need to recognize the black population cluster I posted earlier along with the preserving UCCs. AL has the following UCCs: Birmingham (Jefferson, Shelby), Montgomery (Montgomery, Elmore, Autauga), Florence (Lauderdale, Colbert); Birmingham must be covered by at least 2 CDs.



AL divided into 7 districts using whole counties and urban clusters to the extent possible. No UCCs are chopped and the only county chop is Jefferson. The black belt MCC is only chopped the one time that is required. The county erosity is 54. CD 6 is 43.6% BVAP and D+4 and would provide opportunity for the black minority to elect the representative of their choice by controlling the primary and then winning with a crossover coalition.

Deviation and demographics:
CD 1: +2428
CD 2: +877
CD 3: +909
CD 4: -2955
CD 5: +1181
CD 6: +935; BVAP 43.6%; 2008 D+4
CD 7: -3372; BVAP 46.0%; 2008 R+1

I don't have access to DRA until Mar 28, so if someone wants to run this plan to get the new PVIs I'll add them.

[GM Team Member and Senator WB]

I didn't include the voting age population, so numbers are the general population.

2 black plurality districts.
01 (Blue): (83.9% white. 9.1% black) +1959, R+28.02
02 (Green): (75.0% white, 14.7% black) -234, R+19.47
03 (Purple): (47.1% black, 46.9% white) -3069, D+8.68
04 (Red): (82.0% white, 11.0% black) +1519, R+30.01
05 (Yellow): (66.3% white, 27.3% black) -236, R+16.21
06 (Teal): (47.6% black, 47.4% white) -615, D+1.53
07 (Grey): (68.0% white, 25.6% black) +679 R+17.04

8 county chops.