US House Redistricting: Minnesota

[I spent the winter writing songs about getting better]

34: New district for powerful Republican figure Warren Limmer, now more safe. BTW it looks kind of ugly on a map but makes sense to anyone who's driven on I-94. Safe GOP.
35: I think this seat is actually open. It's a pretty safe Republican district. Safe GOP.
36: This is a swing seat based off the old 47 that fell in 2010. Few changes, Benjamin Kruse will have a fight on his hands. Wonder if he'll flee to the 35th? Toss Up.
37: Pretty much the same. Pam Wolf is the incumbent, should be a competitive seat. Toss Up.
38: Incumbent Roger Chamberlain lives here, and he should be thankful that it is far more Republican than his old seat. Safe GOP.
39: This is actually the closest thing to Michele Bachmann's old seat. It's actually a bit more DFL than the old seat in fact due to the pulling out of Anoka County, but not enough to be truly competitive (won by McCain but he didn't break 50%), if it weren't for that the current incumbent Ray Vandeveer is as crazy as she is. Lean GOP.
40: Chris Eaton won the special election for this this year when the old incumbent died, and it's still a very safe seat. Safe DFL.
41: Barb Goodwin, the Democrat who ousted the old corrupt incumbent in this area should have no problem winning again. Safe DFL.
42: Normally this would be a reasonably safe DFL seat. However the incumbent John Marty is one of the most liberal members of the State Senate, he ran for Governor as the most liberal candidate with a platform based on gay marriage and single-payer health care. The seat has shifted a tad to the right. He probably still wins though. He could also run in the 66th but that would open a primary battle. Lean DFL.
43: Basically the old 55, incumbnet Chuck Wiger should win easily. Safe DFL.
44: The new seat of Terri Bonoff, who survived 2010 by the skin of her teeth. But if she hung on then, she should have no problem from now on. Safe DFL.
45: Democrat Ann Rest should keep winning here fine. Safe DFL.
46: This is basically the "Jew seat", so the lean is obvious. Ron Latz is safe. Safe DFL.
47: Julianne Ortman won this seat held by Pawlenty's Lt. Gov., safe GOP seat in heavily Republican Carver County. Safe GOP.
48: This is a 52% Obama seat, but the numbers in megachurch crowd-heavy Eden Prairie are pretty misleading, where lots of rich more secular Republicans freaked out over the stock market crash and Sarah Palin. Incumbent David Hann did a joke campaign for Governor in 2010 that went nowhere and he didn't put any real effort into (think Fred Thompson) which no doubt annoyed some people, and he won by a fairly unimpressive for 2010 margin, but the district is polarized. We'll call it Lean GOP for now, might be a true Toss Up by the end of the decade.
49: This is great news for the Democrats, the district is basically the old 41 held by Republican Geoff Michel forever. It's 54.7% Obama and has got more Dem each cycle, even though the Republicans currently hold both House seats, one is an almost certain goner (more on that later.) Michel should've held it easily regardless, but he just announced his retirement. Lean DFL.
50: No incumbent here, it most resembles the old 63 represented by Ken Kelash but he lives outside the boundaries. He might want to run here regardless. Otherwise the incumbent narrowly defeated in 2010 Jon Doll might want to do a return since his old Bloomington to Burnsville seat is now far more Republican. Safe DFL no matter what.
51: Resembles the old 38, a classic swing suburban seat that narrowly fell in 2010. Ted Daley will run again but he could lose a rematch. Toss Up.
52: This is a safe seat and is home to the DFL PPT James Metzen. He'll win easily. Safe DFL.
53: This is based around Woodbury, a very affluent suburb and swing area, Republican Ted Lillie represents the closest district in resemblance but he doesn't live in it. The seat is a bit more DFL anyway since it now goes to Maplewood instead of exurbs, so he'll have a tough time no matter what. Toss Up.
54: Basically the old 57, DFLer Katie Sieben barely survived 2010 but it's a Democratic seat, not every election will be 2010. Lean DFL.
55: Safe Republican district, Claire Robling holds it, will still do so. Safe Republican.
56: This is a former swing district, now barely competitive, Obama won it but by only 0.4% in an area he greatly overran. The incumbent is pretty conservative and unpleasant though (he's known for claiming that school integration destroyed Minneapolis back in April), so it's not quite safe, call it Lean R.
57: This district is a tad more swingy than you'd expect one this distance from the city centers to be, Apple Valley for whatever reason does have quite but the incumbent Chris Gerlach has never had much trouble, with him it's Safe R.
58: Exurban seat, no trouble for incumbent Dave Thompson, Safe R.

The rest are all Twin Cities seats, except 66 which is basically evenly split between part of St. Paul and some inner suburbs, and is a safe DFL seat regardless. So 56-67 are all Safe DFL. I'll just point out that at 87.9% Obama, mine remains the strongest DFL seat in the state.

More to come.

[ilikeverin]

Wikipedia says Bonoff lives in Hopkins; the Minnesota Legislature says Minnetonka.  Are we sure what district she lives in?

[I spent the winter writing songs about getting better]

I'd trust the legislative site before Wikipedia. Mind you she's not running the 46th anyway.

The MN Sec of State site does somewhere have a list of every candidate that filed for office in various years's filing report including their address, so this stuff can be confirmed.

Edit: Oh wait, the legislative site gives her actual address. Well then yeah that's definitely more reliable. I should go fix Wikipedia.

[I spent the winter writing songs about getting better]

OK I see where the confusion comes from, her zip code is mostly Hopkins and Google Maps confuses her address for a Hopkins one. But her precinct is definitely Minnetonka. Yes I looked it up.

Never knew she was Jewish either. Probably not that surprising though considering she's not too far off from St. Louis Park.

[I spent the winter writing songs about getting better]

I'll do more tomorrow after church when I got free time, but here's something of note, I got a fundraising email from the anti-gay marriage ban amendment group that was written by Rep. John Kriesel, one of four Republicans opposed, whose speech against the amendment went viral. He notes in it that he's also not seeking re-election to "spend more time with his family".

This is notable for a few reasons:

1-He probably didn't want to risk losing and ruining all his goodwill with liberal activists (I mean he was selected as man of the year by the magazine of Minneapolis' gay community.)
2-He only served one term, and is very young. So he might just be sick of the far right in his party.
3-His seat will be near impossible to hold without him, especially as the GOP will likely nominate some teabagger nut to replace him.

The third being the most relevant here.

[I spent the winter writing songs about getting better]

OK I think I'll revive this soon, especially with my recent thoughts about assigning parliamentary seat style names (most likely Canadian-style) to all of them, which I find much easier in tracking them. For that though I think I'll create a new thread inviting others to do so for all states. This will actually help a ton since most seats have obvious predecessors but often the numbers are way different making tracking changes under the numbers a little difficult.

I will note that the DFL seems to have done a decent job with candidate recruitment in western Minnesota, and we might pick up some surprising seats, though I'll wait for the weekend to do a bit more in depth look into that.

[Torie]

Here is a map for 7 CD's in Minnesota using 2010 population figures this time, rather than projected 2020 figures, just to make it easy. If you really focus on erosity, trying to limit chops, but only doing so if it does not materially degrade erosity, or chop metro areas, my suspicion is that typically there will be but one or two maps really in the hunt in many states. Certainly, Minnesota seems to be one of them.
This map has 3 county chops (Anoka, Wright, and Hennepin), and no locality chops. The trick is to come up with the set of rules, that forces these kind of maps to be spit out of the black box.

[ilikeverin]

That seems like a less-than-insane map.  Though I wonder whether the courts, etc., might like the districts a bit more connected than that.  The Walz southern district has I-90, but the north and middle ones are a bit more random.

[muon2]

Here is a map for 7 CD's in Minnesota using 2010 population figures this time, rather than projected 2020 figures, just to make it easy. If you really focus on erosity, trying to limit chops, but only doing so if it does not materially degrade erosity, or chop metro areas, my suspicion is that typically there will be but one or two maps really in the hunt in many states. Certainly, Minnesota seems to be one of them.
This map has 3 county chops (Anoka, Wright, and Hennepin), and no locality chops. The trick is to come up with the set of rules, that forces these kind of maps to be spit out of the black box.

We seem to be generally in agreement on the map here. On my first look I got something quite similar. It would alter three of the districts to remove one chop. I'm not wild about two separate incursions into Anoka, but it seems to be better than the alternatives.



(Chops 2; Range 5901, 0.78%)

On further inspection I wanted to see if I could reduce erosity and got the following map. I think it's less erose by any measure, but it adds a larger chop into the TC metro. Technically CD 1 already chopped into it in Sibley county, which many would not consider metro at all. However, an algorithm can't tell you where the Census doesn't mesh with ground observers.



(Chops 2; Range 6866, 0.91%)

[Fuzzybigfoot]

Torie, muon, can I have PVI's for the 7 district maps please?  (If you did and I didn't spot it, I apologize)

[muon2]

Torie, muon, can I have PVI's for the 7 district maps please?  (If you did and I didn't spot it, I apologize)

I've put in the 2012 results so I can use the updated PVIs. The changed values in my second map are in parentheses. Torie's CDs 1, 2, 4 and 5 are the same as my first map, and his CD 7 should be about a point stronger D making his CD 6 more R.

CD 1: R+2.6 (R+6.1)
CD 2: R+3.6 (R+0.3)
CD 3: R+2.8
CD 4: D+9.4
CD 5: D+18.1
CD 6: R+8.4
CD 7: D+1.3

[Torie]

One chop into the twin cities metro area (say to take Dakota), in exchange for "freeing" 3 more rural counties from metro domination, and less erosity, is a fair bargain, but having two CD's chop into the metro area is not in my view, and you have Scott and Carver bit off by a separate CD.

I think my "clean" chop into Wright, taking that portion that really is part of the metro area, in exchange for considerably less erosity in the north portion of the map, is a more than fair bargain myself.

I had a couple of more map potential iterations to consider (one where MN-04 nipped off Chisago, and Washington County was bifurcated (that chop is a negative; nasty county chops need to be scored as a negative in some way, so the black box seeks other alternatives if they are reasonably out there),  with MN-02 losing its three southern rural counties), but the program crashed. They may surface later if the stars are aligned right.

Again the point is to have a total point score, so all of these sometimes inconsistent considerations can be reasonably balanced off against one another by the black box. No one factor can be the dominatrix.

[muon2]

One chop into the twin cities metro area (say to take Dakota), in exchange for "freeing" 3 more rural counties from metro domination, and less erosity, is a fair bargain, but having two CD's chop into the metro area is not in my view, and you have Scott and Carver bit off by a separate CD.

I think my "clean" chop into Wright, taking that portion that really is part of the metro area, in exchange for considerably less erosity in the north portion of the map, is a more than fair bargain myself.

I had a couple of more map potential iterations to consider (one where MN-04 nipped off Chisago, and Washington County was bifurcated (that chop is a negative; nasty county chops need to be scored as a negative in some way, so the black box seeks other alternatives if they are reasonably out there),  with MN-02 losing its three southern rural counties), but the program crashed. They may surface later if the stars are aligned right.

Again the point is to have a total point score, so all of these sometimes inconsistent considerations can be reasonably balanced off against one another by the black box. No one factor can be the dominatrix.

The point I'm trying to discern is how the black box should deal with metro areas. The only neutral definition I can count on is the one provided by the Census. The two southern CDs make exactly the same number of chops into the official metro area. But as I noted and we both did with our initial maps, some metro counties are more metro than others. A neutral procedure needs a clear set of facts to start with, and when our guts (which agree) are a mismatch to the initial facts we have a dilemma.

My solution would be to have the neutral engine produce more than one qualified map. That lets the commission select a proper map based on local sensibilities, like how truly part of a metro is any particular county. I think this was our understanding when we analyzed CA, and I think it is still a worthy goal. None of this prevents an overall score that can be used to judge plans from the qualified set, but I think that there is a useful, narrow role for local input that should not be eliminated.

On the subject of my CD 3 I would note that Isanti is also officially in the metro area, though we would agree it is there as a rural exurb like Le Sueur or Sibley. I would also contend that the relatively small increase in erosity for CD 3 is offset by a substantial decrease in erosity for CD 6 by most any measure. That plus the chop reduction make it worth consideration.

[Torie]

Maybe the way to parse this is that using the Census definition as a safe harbor, but if its chopped, no negative points, if the carve out is a low density area, and the metro area just sucks up low density areas of counties that are rural, or whole counties with low density, like Sibley, if the Census counts that. The computer should be able to have density data loaded into its data bank.

I consider our respective MN-06's to be close in erosity measures - but then I like rectangles rather than irregular shapes, and your MN-07 is more erose. Both maps should get high scores, and I agree that where they are out there for the black box to generate, it is good to have more than one map to consider, provided the vetting process is not gameable, because the pros have some say in the process. I don't like these so called non partisan commissions. They have tended to be relative flops so far.

Anyway, here are a couple of other three choppers that might be in the hunt, although the Pubs might nix the first one. The first map has the Sherburne conundrum, the county that takes a bite out of both the St. Cloud and Twin Cities metro areas, so one metro area or the other will be chopped. Most prefer that it be in the St. Cloud zone, because St. Cloud crosses the river into it with about 3 precincts, I understand. But that alone should not be a deal killer.

Muon2 and I seem to both be viewing county trichops as a negative, even if the total chop count is unchanged it appears. That is another factor to consider whether or not it should have separate weight.

 

[muon2]

Maybe the way to parse this is that using the Census definition as a safe harbor, but if its chopped, no negative points, if the carve out is a low density area, and the metro area just sucks up low density areas of counties that are rural, or whole counties with low density, like Sibley, if the Census counts that. The computer should be able to have density data loaded into its data bank.

I consider our respective MN-06's to be close in erosity measures - but then I like rectangles rather than irregular shapes, and your MN-07 is more erose. Both maps should get high scores, and I agree that where they are out there for the black box to generate, it is good to have more than one map to consider, provided the vetting process is not gameable, because the pros have some say in the process. I don't like these so called non partisan commissions. They have tended to be relative flops so far.

Most compactness measures would rate my CD 7 almost as compact as yours, and my erosity is based on a diagonal line rather than an east-west axis. I'll grant that my CD 7 is slightly more erose, but I can't see a methodology that glues itself to NSEW lines, there are too many irregular counties, even in MN. OTOH by any measure my CD 6 is much more compact than yours, that's true if it's based on circles (like MI), squares (like IA), perimeter, or counting counties that touch across the boundary.

If your preferred shape is rectangular and not square, then what prevents a long thin line of single counties from counting as an ideal district? It may not be erose, but it would not pass my sense of a compact shape. I don't find fajita strips attractive as a redistricting plan. Furthermore without a consideration of both dimensions of a shape, constructing an algorithm is challenging. I sense that you want to maximize the occurrence of straight lines on the perimeter and that the actual perimeter conforms to those ideal straight lines. That is a non-trivial test, and it only gets worse when boundaries have to conform to shapes like rivers. The Minnesota River is a good example here that's as worthy to use as a boundary as would be a metro area, with the exception at Mankato due to its metro area.

As for density, it is not commonly part of redistricting software, though it could be constructed in a general GIS framework. I want a tool that I can program on existing platforms, preferably one that could be used to get public input, too. How else are we to test whether or not a set of value-based criteria achieves the desired goal?

[Torie]

Oh, if county lines are irregular, than you can't draw perfect rectangles (and I haven't drawn perfect rectangles, just some reasonable facsimile thereof), and a long thin rectangle is not good either (and I don't do that either - well MN-01 sort of is, but that CD tends to be bit of a pain in the butt due to population and geography issues (e.g., both of us would rather lose our package then have a non metro chop in MN, and to avoid that for MN-01, the options narrow down, simply because the little counties don't have quite the right population numbers to get something less erose not using a metro county, but without a chop; we both drew maps that make MN-01 more compact, but at the cost of nipping off a metro county )).

Anyway,  the computer program if well done will "fairly" score out the balance. Back, and back again, and yet again,  we come to that irksome balancing test. You're the brain, and I'm the trouble maker - figure out a magic formula to get the balancing test right, so it meets the I know it when I see it (excessive erosity vis a vis the alternatives) test.

The density issue can be put into the software, and the software will tell you where the zone is, that can either be part of metro CD's or no, without penalty. There are not enough folks in the grey zone, to make that much of a partisan difference, and if it does, and skews the partisan profile, then one party or the other will nix that map. But most times, neither party will care that much about that particular issue, nor will the residents. So if having a grey zone, helps with compactness, while not allowing making metro areas a free fire zone, then life is beautiful.

Anyway, below is a three chop map that is just gloriously relatively compact, and keeps the metro area as tight a drum almost, but has that trichop of Hennepin (2 chops for Hennepin, and one for Anoka).  This issue needs to be resolved.

[muon2]

Your latest offering convinces me that density as a scored factor is the wrong way to go. Sibley, Le Sueur, Isanti, and Chisago are all low density counties in the TC metro. Wright is a medium density county along with Sherburne, Scott and Carver. But my classification is somewhat subjective based on relative populations.

Our various offerings tend to be loosest with the low density, and willing to give on mid-density like Wright since the TC metro has the population for about 4 1/4 of 7 districts (it may be 4 1/2 by 2020). In principle there should be no preference as to whether Wright or any of the other mid-densities split out. That leads me to more firmly believe that the metro split factor is one the final selection commission (or whomever) should consider among competing finalist plans.

On a related note, it will be helpful if you quote the population range between largest and smallest in your plans. Scoring that will be a component since the courts will want to know that population range is minimized given other scored factors. I'm working on a simple score that will allow the degree of equality to be directly compared to the chop count.

Edit: I've added chop and range measures to my two plans.

[Torie]

The chop of Wright out of the metro CD's is negative in my map. I don't pretend otherwise. However, if a metro chop generates enough positive points otherwise, it may be worth it. Such a chop is needed to get the erosity of MN-06 or MN-01, or both, down. But to ignore the factor entirely I think is a mistake. It might exclude otherwise worthy maps, or include unworthy ones, for consideration. Wright is one county by the way, where part of it per density factors, is in the metro area, and part is not. So chopping it in the right place (as my first map did), would remove it from counting as a metro chop. Severing off Isanti and Chisago should not count as metro chops at all, nor should their inclusion in metro CD's be viewed as a negative either. Those two counties swing both ways as it were. They and western Wright are in the grey zone. Ditto south Sherburne.

I don't consider something closer to exact equality superior to something less exact, provided the 1/2% rule is hewed to. Using the pad, helps to reduce erosity, keep municipalities together, have straighter lines in county chops and so forth. Having free rein to move around 6,000 folks or so between CD's is the reason why that I suspect none of our maps has any locality chops. And it will generate more reasonable maps from which to choose, without having much if any partisan effect. The computer should have that flexibility. I don't consider having something closer to exact population equality as much of a public policy goal at all. Each and every other consideration out there is more important than that to me.

And we still have the trichop issue that you did not address in your above post, which was the primary reason I put up the latest map! 

[muon2]

The Wright issue doesn't bother me since at least 1/4 of a district in the metro area must link to areas outside the metro. It could be Wright or any of the other mid-density counties, but at least one must shift outside the metro.

I didn't address the trichop since it is three chops like your first map. On chops I would treat them equally.

I also passed on the request since the range does matter. The court in Tennant made it clear that the equality relaxation only applied when other rules were tightly applied. We have chops (county and metro) and erosity as measures we must address. If I can make a map that maintains those measures compared to a given plan and has a smaller range, then it must supersede that given plan. That's why a solid erosity measure becomes vital, and eyes alone are not sufficient.

To illiustrate let me put forward the following map. It has two chops (Hennepin and Anoka), no town splits in the chops and the range is only 888 (0.12%). It keeps the metro area together except for Wright and Sherburne which is not unlike your first offering. It is consistent with what the IA compactness rules would require given the order of magnitude improvement in population equality. My road connection erosity method is designed to weigh against it, but you seem to have soured on that method.

[Torie]

Yes, your map above is a fail on the erosity measure. It's close to hideous actually (for example you took the weakest part of one of my maps (the elongated MN-02), and just stretched it out some more, and out of the metro area entirely into Pine County (not sure how you did that, since all you lost was a small chop into Ramsey, but whatever, and if that is what it takes to avoid a chop, it's just not worth it - that balancing thing again). Sure if two maps get an equal score, the one with less of a deviation in populations is better no doubt. But such equal scoring is a hypothetical that probably will not occur. If a variation will smooth out a line, or reduce erosity, I'm for it. You seem more interested in chop counts, and road connections, than erosity per se, or smooth lines for that matter. On that one, we are just not on the same page at all. You are just not into visual tests;  I am, that is what folks actually see. Sure you need an algorithm to score erosity, that is compatible with the eye test, and so far, we are not making much progress in generating one - at least one that pleases me. That just does not seem to be your focus.

In any event, a black box generating population deviations following factors that have a reasonable basis, will be just fine with the courts. On that much, I am confident.

Anyway, for what map, or maps, do you want my deviations for?  I guess you have a right to do your own thing, of course, even if I think it is largely a red herring. You don't, so that is that.

[muon2]

Yes, your map above is a fail on the erosity measure. It's close to hideous actually (for example you took the weakest part of one of my maps (the elongated MN-02), and just stretched it out some more, and out of the metro area entirely into Pine County (not sure how you did that, since all you lost was a small chop into Ramsey, but whatever, and if that is what it takes to avoid a chop, it's just not worth it - that balancing thing again). Sure if two maps get an equal score, the one with less of a deviation in populations is better no doubt. But such equal scoring is a hypothetical that probably will not occur. If a variation will smooth out a line, or reduce erosity, I'm for it. You seem more interested in chop counts, and road connections, than erosity per se, or smooth lines for that matter. On that one, we are just not on the same page at all. You are just not into visual tests;  I am, that is what folks actually see. Sure you need an algorithm to score erosity, that is compatible with the eye test, and so far, we are not making much progress in generating one - at least one that pleases me. That just does not seem to be your focus.

Let me try to sway you again why I think my method suits your needs. Note that the IA system would not, nor does MI for it's measure of chops. Both allow for the kind of erosity you do not favor.

To illustrate, I'll simplify my method by taking a larger set that's easier to visualize. We'll just count counties segments that form borders between two counties and are on the perimeter of a district. Each separate chop into a county is treated as a county. My stricter rule using roads does not change the relative results.

Let's look at your map with the Washington-Dakota district:



It has three chops and here's my simplified erosity count:
CD 1 (SE) 17 (two touches with Hennepin from chops and a bit from Watonwan-Jackson, which is why my refinement is in place)
CD 2 (Dakota-Washington) 15 (the chop in Ramsey has two touches to the other part of Ramsey and to Anoka)
CD 3 (West metro) 15
CD 4 (St Paul) 12
CD 5 (Mpls) 7
CD 6 (SW) 28
CD 7 (N) 14 (point contiguity at 4 corners does not count)

The total divided by two (since each border segment counted twice above) is 54.
BTW it looks like there are some red spots in Mille Lacs and Wadena that would throw off your populations.

Now lets look at my evil, erose, but I assume more equal map:



CD 1: 16
CD 2: 15
CD 3: 17 (including Hennepin-Ramsey along the river)
CD 4: 11 (Columbia Heights and Coon Rapids make up the Anoka chop and touch Ramsey)
CD 5: 7 (including Sherburne-Hennepin)
CD 6: 37
CD 7: 19
total/2 = 61
The ugly shape of CD 6 was especially costly here, as I think you would agree it should be. The simple erosity test allows your map to pass to the reviewers along with this more equal, fewer chop map. The reviewers can then reject the ugly map for all the reasons you cite. Isn't that what you want?

In any event, a black box generating population deviations following factors that have a reasonable basis, will be just fine with the courts. On that much, I am confident.

That might be true if the black box generated only one map. I think we concluded in CA that a single solution would often offend local sensibilities, and we needed a way out of the box. Thus my attempt to generate a set of qualifying maps. Once that happens population deviation cannot be ignored as I read the decision from WV. The plaintiffs had a set of examples that were whole county (0 chop) and within 1% range and the court agreed that population range would then select the best. WV prevailed because they had two parameters (whole county and minimum population shifts) that produced only one solution. If erosity is a simple whole number formula that can be compared to the chop count then ties in the result are going to happen. Thus one must be prepared to have the population figure as well.

I personally would go further and allow the reviewers to consider more chopped or erose plans that provided better population equality and let them decide. That's why the criteria need scoring values so that the trade off can be judged. But perhaps we disagree on that point.

[Torie]

Well we agree on a point system. If population inequalities don't generate more points, it would be hard to defend. But if it does, it's OK by me. Then the deciders based on some strike system can pick one of the high scoring maps. I would be amazed if a court rejected a high scoring map with a slightly higher population deviation than another map with less population deviation given such a process. And obviously this fact pattern has not been before the courts. The point is, is that given one approach to a map, if the population inequality is a point generator and thus has a legitimate reason, for the court to reject that in favor of a another approach picked by the deciders from an array of black box generated maps, would be court intermeddling at its worst, and I just don't see that happening. So on that aspect, we just disagree absolutely.

Surely you don't think the court would pick a plan  with less population inequality but with a higher score, than a map with lower population inequality do you? That would be insane. I guess what I really disagree with most, is giving points to a map with less population inequality, so that map gets pushed ahead even though otherwise inferior. That really offends me.

[muon2]

Well we agree on a point system. If population inequalities don't generate more points, it would be hard to defend. But if it does, it's OK by me. Then the deciders based on some strike system can pick one of the high scoring maps. I would be amazed if a court rejected a high scoring map with a slightly higher population deviation than another map with less population deviation given such a process. And obviously this fact pattern has not been before the courts. The point is, is that given one approach to a map, if the population inequality is a point generator and thus has a legitimate reason, for the court to reject that in favor of a another approach picked by the deciders from an array of black box generated maps, would be court intermeddling at its worst, and I just don't see that happening. So on that aspect, we just disagree absolutely.

Surely you don't think the court would pick a plan  with less population inequality but with a higher score, than a map with lower population inequality do you? That would be insane. I guess what I really disagree with most, is giving points to a map with less population inequality, so that map gets pushed ahead even though otherwise inferior. That really offends me.

So we agree that geographic measures for chops and erosity should be based on a point system. And I think we agree that if a map scores equal in on erosity but worse on chops or equal on erosity but worse on chops it should discarded.

I think you may also willing to see a point system for population inequality. One role it can play is to show when the population inequality in two different plans is functionally the same even if the exact values are not the same. Another role it can play is to differentiate between two plans that have the same geographic measures. I hope you would agree that in this case the plan with greater inequality should be discarded. This is the case that I think the court would be sensitive to.

Where I believe we disagree is in a case where two plans have dissimilar geographic scores. I'm willing to consider a trade off of geographic quality for population equality. I read your text as saying that you would not be willing to make that trade. I don't think that the court would force the hand in this case, since through a point scale I can show why geographic quality might trump population inequality.

On your panel I sense that population inequality would serve only as a threshold and tie breaker, where my panel sees it as something to trade, not unlike a trade of chops for reduced erosity. Is there a reason I shouldn't want to look at the population trade, or is this just a matter of taste?

[ilikeverin]

Your most recent maps make angels cry, muon.

[muon2]

Your most recent maps make angels cry, muon.

And that's of course my point. You can't establish what you want to be good without identifying what is bad. Just looking at a metro area can result in links like the Dakota-St Croix combo, and if the focus is just on the metro, look what happens to adjacent non-metro areas. Local sensibilities are going to come in eventually so one challenge is determining the point when rules stop and local control starts.

You may recall I grew up in the Twin Cities, so I can push to see what neutral rules could do that offend local sensibility.