CA CD Wine Country Map Quest poll

[muon2]

You may be counting counties chopped, instead of the chops themselves. For example you have three CDs that go into Sacto county and also include part one one neighboring county. That's a total of six chops, but only three counties because you've tri-chopped Sacto. If you rearranged those three districts to place one CD entirely in Sacto there are still three chop fragments outside of Sacto, but only two inside. One chop disappeared because the CD is entirely within the county. One of the remaining pieces has three fragments, and in principle it can be brought to two, but it may not if you have more than two CDs with only one fragment.

I'll think about 10%, but it seems arbitrary. At least I can point to a federal decision on which to rest 5%.

Well another issue is how to count chops, but I think my chart is the proper way: how much does each CD chop?  Putting that aside, revise my map in Sacto in a way that you think loses a chop (and I will point out why it doesn't - I think ). It's easier for my tired old brain to work that way.

I have realized that any limitation of the percentage of a CD that is part of a chop would be close to a disaster for CA (the limitation on the percentage of a county that can be chopped would remain in place (unless a very small county perhaps, but maybe not - that area in Norcal is the issue there), is the protection against the real abuse anyway).  Can you figure out why? 

Then help me out with your chart, because it doesn't seem consistent in counting chops. For instance in San Bernardino there are three CDs. CD 8 you list 2 chops, which I see correspond to SanB and Kern. Fine. CD 31 has none because it's all in SanB.

But CD 35 is listed as only one chop, and I count 2: SanB and LAC. I have to count the LAC piece since it has more than the minimum one left over fragment. If a county that is cut in two generates a chop for each piece, a remaining fragment in a large county that is cut in two should also count as a chop for each piece. What am I missing?

To your second point, I'm not seeing the disaster you see. I fear perhaps you are more conscious of political sensibilities than my friends in IA and the fair map folks in OH.

[muon2]

Let me expound more on the traditional ways to count county chops. One presumes that each county has an ideal configuration. For small counties that configuration is to be entirely within one district. For large counties that configuration is to have the maximum number of whole districts within the county and the remainder in a single district.

There are two ways to count the traditional configuration. Most common is to assign it zero chops for both small and large counties. That treats the baseline with an ideal of zero. The other way is to count the total number of districts in each county so that small counties start at one and large counties start at the number of whole districts that can be placed within plus one for the remainder. This starts with a baseline ideal equal to the number of counties plus the maximum number of districts that can be placed entirely within counties.

For any district which is not ideal one counts all the districts in the county, and in the first method of counting one subtracts the districts entirely within and in the second method one does not. Note that in the first method one jumps from zero to two and it is not possible to have a county with one fragment. In the second method one can never have zero pieces in a county so there is no jump of two for the first chop.

As I noted the first method is preferred since it has an ideal of zero, and it puts a premium on avoiding the first chop. It does allow one to go below the number in my aforementioned formula by chopping more than two districts into a county. The second one has the feature of treating two counties each split in two the same as one county with no split and one with a three-way split.

Suppose there are five counties for four districts. The counties are each 80% of the population for a district. They are arranged as a central county A and four counties that wrap completely around A: B, C, D, and E in clockwise rotation. Imagine two ways of mapping: I) one where A is whole and shares into B, which shares to C, which shares to D, which ends at E; II) the other divides A into four slices and leaves the other four counties whole. The first method of chop counting scores I) as 6 chops (2n-2) and II) as 4 chops. The second method counts both as 8 chops.

[muon2]

Yes, but if a wall location ends up creating a nasty Lancaster chop, or forces stuff to go where it should not go, that is a problem.  And there may be crossings that should be "disfavored."  Having some statistical limits like you suggested might work, which can be violated under certain circumstances. Or maybe we have defined regions, in which there may be only one chop out unless the VRA demands otherwise, or to unite a city that is already mostly in one CD. And Sacto taking W. Sacto, or uniting Yuba City and Marysville, perhaps should not count as a chop for example.  

Muon's rule works better in states where you have small square counties filled with cornfields, and regions can be pretty arbitrary.   This is less true in California where there are pretty distinct regions often separated by mountain ranges.

But the problem with the mountain states is determining when it's OK to cross or not. Just look at all the debate we had about whether WA should use the Snoqualmie Pass or not. Even in CA there seems to be an inconsistency such as the lack of obstruction to use the long pass from Eureka to Redding, but no suggestion to use the mild pass from Gilroy (Santa Clara) to Los Banos (Merced).

Edit: I will note that by using that pass I can link the two section 5 counties and bring the HVAP to 51.5% with the infamous SLO wall.

[muon2]

Granted, I haven't really been following this thread all too closely, but after seeing Torie's sig, I have to ask: What did the poor people of Eureka do to get screwed over like that?

Not make the same quality wine as Napa, Sonoma, and Mendocino.

[muon2]

Muon2 I am still waiting for a Sacto area map that loses a chop. As I said, it appears that if a CD in Sacto is nested, that it creates another chop elsewhere. If somehow that reduces the overall chop count, than something is wrong with that count method. It appears that what you are saying, is that if a CD chops into a county with nested CD's, with say an extra 500,000 people, and five other CD's chop into it, picking up 100,000 people each, those 5 chops don't count as chops. Do I have that right?

They count as chops. In method one I count it as 5 chops, but if all 500K goes to one CD with the rest nested it would be zero. In your map LAC has two chops by this count, but Ventura has none.

In method two I just count all the districts in the county whether they are wholly in or not. But then all counties count at least one. So chops count there, too, it would be 5 plus the number of nested districts.

To show you a chop reduction, I need to know how you want me to count. I believe I can reduce it either way, but the plans will be different.

[muon2]

You must be logged in to read this quote.

I agree with this method of counting, having played with it, for your hypo.  It basically counts the number of chops within each county with CD's not wholly contained therein.

Here is my problem.  The map below pushes CA-06 entirely within Sacto, so now the county has two chops, losing one. However, that chopped is gained back by the chop into Sutter by CA-04.  So no lost chop.



How about the muni chop issue that  I discussed above?  

And have you figured out yet why we can't have a percentage if CD rule involved in a chop in CA? Do you just pick and choose which questions you choose to answer from your students?  

You are right. CA-35 has two chops. I was just labeling the entrance and exit routes for LA County. The entrance and exit points should each count as two chops rather than one. You are right.

I'll get to your Sacto issue now that I know how to count. In the Sacto area I count 9 chops in your original map: 3 for Sacto, 2 for Yolo, 2 for Placer, and 2 for San Joaquin. Is that right? Your map above has 8: 2 for Sacto, 2 for Sutter, 2 for Placer, and 2 for San Joaquin.

On the muni chops, certainly the commission justification route is consistent with our previous holdings. Another option for large cities is to utilize census-recognized neighborhoods. One can treat them as separate municipalities for chop rules. I've looked at this in some eastern cities, but I haven't investigated that as an option in CA.

I thought I did answer the CD percent rule, by noting that I am less connected to specific jurisdictional pairings than some others here. If there's a more technical issue that I'm missing, let me know.

[muon2]

OK, I see the count issue, but something is wrong if replacing one chop for another generates an additional chop net. I guess it gets back to the idea that if a nested CD leaves the nest to meet another CD's chop quota, that does not generate another net chop. I am having trouble getting my mind around this, but I don't think a map chopping into Yuba City is better than a map chopping into West Sacto. And I guess if I can send CD-09 to take West Sacto somehow, then I lose a chop. I don't like it. Maybe the way to count is by chops into a CD. Sacto has one chop in no matter which map it is. You count a CD as chopping in rather than out by whether a  majority of the CD is in the county or out. Nested CD's are like walled CD's, which don't really save on chops if other CD's have not yet met their chop quota. That is why I said I guess, that is only where a county has a population to support 3 whole CD's, that one must be nested unless three other CD's are using that county CD's one way or the other to meet their chop quota.

How about every time a CD crosses a county line but fails to take the whole county counts as one chop, rather not two? Does that fix it?

Here's a version that reduces chops even more and stays within your boundaries for the Sacto area CDs, except for Plumas which I moved to CD 1. I also got the Yuba City area together in one area at teh expense of the Oroville area town partitioned into separate CDs. In this plan there are only 7 chops: 3 in Butte, 2 in El Dorado, and the perennial 2 in San Joaquin.



You are really hung up on keeping West Sacto with Sacto. Is putting West Sacto with the valley that much worse than slicing Elk Grove into the Stockton CD? I'd relax and say some metro splits are just meant to be.

On the question, your suggestion would mean for instance that Ventura is worth one chop as you have drawn it. What that implies is that small counties have an ideal chop count of zero, but large counties can do no better than one. It doesn't really work, unless one goes all the way to the second method I outlined which counts all CDs in a county, so all counties have a minimum of one plus whole districts that can be contained. It's not as good of a method as the first choice, but see if that gives you what you want in Sacto, and we'll see where it takes you elsewhere.

[muon2]

The problem with the 20% or 10% or 5% CD rule, is that CA has some huge counties, and some tiny ones. Suppose that LA County has exactly 15.5 CD's in it. So LA County must chop out in three places to get each of the chopping CD's down below a 20% chop, generating one more chop, and precluding LA County from being a walled county of course, with but one chop out. And suppose that LA plus San Bernardino has 17.5 CD's. Now to meet your rules, the San Bernardino CD needs to cross Luther Pass to take a bunch of small Sierra Counties, so that its excess after swallowing whole counties gets down below 20%, creating a disgusting CD that no one could possibly praise - or tolerate.  It just won't work. We cannot have such a rule. The Midwest with a lot of counties, more equal in size, and lots of handy small ones, allows perhaps for such a rule with some constraints that are tolerable. CA does not.

In addition, a CD being bisected by two counties is fine if it is in a metro area. Nobody cares if half of CA-35 were in LA County suburbs, and half in SB suburbs. It is one urban mass.

You may have missed my suggestion to this problem. The first thing to note is that the Victorville-Tahoe district was a result of the wall selection, not the 5%/20% rule. My wall map was made by dividing the state into compact regions where there is minimal deviation in each region from a whole number of districts. This insures a maximal number of county splits that involve these minimal population shifts.



Subregions use the 5%/20% rule. When the red region is subdivided, it can't avoid the aforementioned Owens Valley CD. That was a product of the walls.



As I suggested earlier, one reasonable procedure is for the Commission to reject the initial wall proposal with instructions for what should be done to generate a different wall map, with presumably greater deviation. This would mirror the IA process where the legislature can reject up to two maps that were drawn by strict rules involving counties, population, and compactness.

So since you've seen where the walls lead, give me as your neutral drafter instructions about the how the walls are defective, and what instructions should guide my redraft.

[muon2]

Your 20% rule would force the crossing yes, if the wall is "wrong."  Without the 20% rule, it would not. So why do we need these regions based on population, just to prop up a 20% rule that does not to me seem to be that important - not nearly as important as chopping up a county. The regions of CA should be based on geography and urban versus rural and the central valley, and so forth, not based on what will cause your 20% rule to avoid creating a cf. Do you agree with me that your 20% rule (5% is just too terrible to contemplate) in my hypo will create an extra chop if LA County has 15.5 CD's of population?

My main problem with the SLO/M wall, is that it cut up the Santa Cruz area. Secondarily, it made things considerably more difficult and less clean down south. The former should require a finding, or simply be banned, if another wall would avoid it. On the second, it is a judgment call. There can be only so much micro management. And if a Ventura County wall creates a 20% "problem" for SLO, but avoids a chop of Santa Cruz metro, that is well worth it. The primary purpose of walls come to think of it beyond crossing the sierras, is to avoid ugly muni/metro chops come to think of it. And no region should force a CD to get too big or two erose, or go too far afield.

Or am I still misunderstanding you?

Yes and no. I could go into detail, but I don't think it's worth another go around. What I sense is that you want standards that can constrain the commission from doing the things they did, but those constraints may also prevent you from having a free hand. The MI standards can be pretty well twisted to partisan intent, so they aren't strong enough in MI, let alone in CA where the lack of townships really opens the field up.

When it comes to regions defined locally, the ones on this forum seem just as fluid to me as the commission's work. In IA they aren't shy about slicing the Des Moines metro any given decade even though it can easily fit in a single CD. Why? Because the constraints that forced the split pays dividends to the public in its political process.

What hard constraints can you stand? Hard in the sense that there needs to be solid justification from the commission that a constraint must be overridden.

[muon2]

I would still like to know where my criticisms of your 20% CD rule are off base. Second, is our difference on the chop count, that you don't think CA-05 going into SF should count as a chop, while I do, although obviously a permissible chop? Do you see my point, that no chop rule could sell intuitively that says that one chop into W. Sacto is OK, but another into Yuba City is not? 

As to walls, the Sierras are an impenetrable wall. In a pinch, you might be able to append Mono County to a Sierra foothills CD (but I really cannot imagine how it could ever be justified), but basically Inyo and Mono go with Socal. Alpine preferably should not, but it could. Softer walls, but walls of some sort, would be a Bakersfield CD not going into Socal, except if it is in Kern, and the Central Valley based CD's not going to the coast.  Lake and San Benito and Napa are flex counties. 

What I am opposed to is regional walls based on population solely for the purpose of meeting your 20% CD (or 5% CD rule).  I don't think I can be persuaded on that, or that it would ever sell in CA. But you can try. If  a wall constraint for a regional approach has some other purpose, maybe that has merit.

In general, I think, what is crossed should be left up to the Commission if there is a road etc.  What I am more concerned about is if your rules force maps that would render any recommendations made on a fix for how CA does it, subject to the kind of criticism, that would cause folks not to take the work product produced seriously, and now we are looking for dams to mitigate that potential for derisive criticism. Nor do I really see a public policy reason why a CD as opposed to a county, cannot be bifurcated, that has much that is compelling about it. It seems to me more like rounding error issue. Maybe what drives you is simply to limit the amount of Commission discretion, to make the game or how to game more difficult. I doubt we need to go quite that far, if we have the right procedural fixes.

My frustration stems from the fact if you want a procedure where all choices are discretionary or the end map justifies the means, I don't think my skills can help.

I offered two common models for counting chops, and you picked the one you thought made the most sense. But when I applied it uniformly, you didn't like the result. We can pursue your suggestion, though I think it's a posteriori thinking, but I suspect as we get to some other part of CA, you'll find it has flaws. I know that it does when applied in other states. Any count by CD can be morphed to a county method and vice versa. That's a mathematical theorem.

You referenced minimizing chops and their size. But every serious constraint is rejected. Your proposed standard puts virtually no constraint on a state with as many populous counties as CA. Write it in detail, and I'll draw a map to drive everyone here crazy. I'm left feeling that the next commission will draw whatever it feels meets CofIs it defines, and any loose constraints will be fit to their needs. If they want to tilt like they did this decade, they'll be free to do so. Even worse look at what happened in AZ on the second go around for a commission.

What's a scientist to do?

[muon2]

My philosophy for neutral mapping is extremely simple. These are my primary points.

1. Districts should be internally connected according to consistently applied rules.

2. There should be as few county splits as possible, and when a county is split in two one fragment should be as small as possible.

3. Counting splits should maintain parity between small and large counties.

Without these I find little way to avoid gaming the system. These are my next tier philosophical points.

4. The VRA provides an out to any rules otherwise created. I feel strongly about section 2 and I hope SCOTUS will finally provide uniformity to how it should be measured. I am not enamored of section 5 as I find the jurisdictions to be anachronistic.

5. Splits small enough to allow a state to avoid exact equality to minimize county splits under SCOTUS rules are best. This allows a state to create two maps, one with county integrity and one with the split as AR did in 2001. My sense is that the map with county integrity survives a challenge.

6. Splits larger than that, but small enough to meet state legislative district deviation under SCOTUS rules are next best. This is because I'd like a set of rules that can apply to mapmaking at multiple levels.

Beyond that is implementation.

[muon2]

My philosophy for neutral mapping is extremely simple. These are my primary points.

2. There should be as few county splits as possible, and when a county is split in two one fragment should be as small as possible. Yes, if we can get a count procedure that does not favor one split over another, due to some nesting preference.

3. Counting splits should maintain parity between small and large counties. No, not if it distorts the count procedure, to force a line crossing one place versus another.

I bolded my comments.

I have dozens of states I've looked at and I've presented detailed rules. Give me your detailed counting rules and your procedure to deal specifically with point 2 given your exact counting rules. I'll test them, though I can say with confidence that any specific count rules will in fact force line crossings in one place versus another.

[muon2]

5. Splits small enough to allow a state to avoid exact equality to minimize county splits under SCOTUS rules are best. This allows a state to create two maps, one with county integrity and one with the split as AR did in 2001. When it is clear that one can vary population, then sure you would to avoid chops within the variance constraints. Until then, a chop is a chop. I think when it comes to the size of the chop, for a chop rule, the percentage of a county taken is sufficient. Otherwise focusing on chop size by the size of a CD, you might force an extra chop or crazy CD's without a further overlay of rules. We need to try to minimize the rules where possible.

I bolded my comments.

It's clear to me that you can vary from exact equality up to a range of 1% if the state can show a compelling interest. County and municipal integrity have withstood SCOTUS scrutiny on this point.

[muon2]

I would still like to know where my criticisms of your 20% CD rule are off base. Second, is our difference on the chop count, that you don't think CA-05 going into SF should count as a chop, while I do, although obviously a permissible chop? Do you see my point, that no chop rule could sell intuitively that says that one chop into W. Sacto is OK, but another into Yuba City is not? 

As to walls, the Sierras are an impenetrable wall. In a pinch, you might be able to append Mono County to a Sierra foothills CD (but I really cannot imagine how it could ever be justified), but basically Inyo and Mono go with Socal. Alpine preferably should not, but it could. Softer walls, but walls of some sort, would be a Bakersfield CD not going into Socal, except if it is in Kern, and the Central Valley based CD's not going to the coast.  Lake and San Benito and Napa are flex counties. 

What I am opposed to is regional walls based on population solely for the purpose of meeting your 20% CD (or 5% CD rule).  I don't think I can be persuaded on that, or that it would ever sell in CA. But you can try. If  a wall constraint for a regional approach has some other purpose, maybe that has merit.

In general, I think, what is crossed should be left up to the Commission if there is a road etc.  What I am more concerned about is if your rules force maps that would render any recommendations made on a fix for how CA does it, subject to the kind of criticism, that would cause folks not to take the work product produced seriously, and now we are looking for dams to mitigate that potential for derisive criticism. Nor do I really see a public policy reason why a CD as opposed to a county, cannot be bifurcated, that has much that is compelling about it. It seems to me more like rounding error issue. Maybe what drives you is simply to limit the amount of Commission discretion, to make the game or how to game more difficult. I doubt we need to go quite that far, if we have the right procedural fixes.

My frustration stems from the fact if you want a procedure where all choices are discretionary or the end map justifies the means, I don't think my skills can help.

I offered two common models for counting chops, and you picked the one you thought made the most sense. But when I applied it uniformly, you didn't like the result. We can pursue your suggestion, though I think it's a posteriori thinking, but I suspect as we get to some other part of CA, you'll find it has flaws. I know that it does when applied in other states. Any count by CD can be morphed to a county method and vice versa. That's a mathematical theorem.

You referenced minimizing chops and their size. But every serious constraint is rejected. Your proposed standard puts virtually no constraint on a state with as many populous counties as CA. Write it in detail, and I'll draw a map to drive everyone here crazy. I'm left feeling that the next commission will draw whatever it feels meets CofIs it defines, and any loose constraints will be fit to their needs. If they want to tilt like they did this decade, they'll be free to do so. Even worse look at what happened in AZ on the second go around for a commission.

What's a scientist to do?

What a scientist is to do, is persuade an informed lay person, that you don't have an intuitive unacceptable flaw, or that any rule has such flaws. I don't think either of us have defined the nub of the problem here adequately. I think my quota approach gets it, but maybe not. Is there any count rule, where switching out one chop for another does not change the total number of chops?  If all count rules still cause a count change depending on where the chop is, that makes the quest to avoid that Quixotic, let me know.

Anyway, maybe we can't agree on everything, but it would be nice to understand better why where we disagree. And we agree on some things, to constrain the Commission further, and some things are better than no things, no?  Be happy.

In any event, the chop count is a technical issue, albeit important because the chop count I agree should be a constraining factor. The percentage of CD is a substantive issue, not in my opinion worth the candle of the potential consequences (if you limit the percentage of county taken absent a finding), maybe requiring more rules to avoid the rule causing nutter results. And I still have not got an answer that if LA County as 15.5 CD's, does that force another chop under a 20% percentage of CD rule?  

Lawyers I guess in their deposition like specific answers to specific questions.

Sorry if I am frustrating you.

As do scientists so they can test the hypothesis.

[muon2]

5. Splits small enough to allow a state to avoid exact equality to minimize county splits under SCOTUS rules are best. This allows a state to create two maps, one with county integrity and one with the split as AR did in 2001. When it is clear that one can vary population, then sure you would to avoid chops within the variance constraints. Until then, a chop is a chop. I think when it comes to the size of the chop, for a chop rule, the percentage of a county taken is sufficient. Otherwise focusing on chop size by the size of a CD, you might force an extra chop or crazy CD's without a further overlay of rules. We need to try to minimize the rules where possible.

I bolded my comments.

It's clear to me that you can vary from exact equality up to a range of 1% if the state can show a compelling interest. County and municipal integrity have withstood SCOTUS scrutiny on this point.

OK, we can suggest doing that then as an option. I have absolutely no objection to doing that, although again I am worried that if we have an absolute minimum chop rule, that might create map distortions. On that one though, maybe if staff can come up with maps that minimize chops using the 1% rule, a finding could say thanks but no thanks, the maps all suck, versus ignoring the 1% flex rule. I am worried however, that it introduces another level of complexity. So it should probably be in an addendum.

The problem is that if it is optional the courts will disallow it. It must be consistently applied to be a valid departure from exact equality. However, one can exchange tighter tolerance on the deviations for more regions. That was the point of the table I compiled and displayed earlier in the thread. The commission could say we don't like the 4-region plan, can you produce one with more regions (and hence fewer splits) but with greater deviation up to the 0.5% SCOTUS limit? That I can do, but the commission will have to abide by one of the regional plans' strictures or the rule has no validity.

[muon2]

My philosophy for neutral mapping is extremely simple. These are my primary points.

2. There should be as few county splits as possible, and when a county is split in two one fragment should be as small as possible. Yes, if we can get a count procedure that does not favor one split over another, due to some nesting preference.

3. Counting splits should maintain parity between small and large counties. No, not if it distorts the count procedure, to force a line crossing one place versus another.

I bolded my comments.

I have dozens of states I've looked at and I've presented detailed rules. Give me your detailed counting rules and your procedure to deal specifically with point 2 given your exact counting rules. I'll test them, though I can say with confidence that any specific count rules will in fact force line crossings in one place versus another.

Can you give me a grand unified theory as to why you are so confident?  This is important. You "win" the argument if you are right, at least up to the point that the chop rules need to take cognizance of the one chop location versus another conundrum, maybe to allow enough flex to do one chop versus another.

I don't think either of us fully understand this yet, sad to say. It's that tough. If one of us did, the grand unified theory could be pounded out on the keyboard.

In the meantime, you want me to come up with chop count procedures that I think might avoid the conundrum, so you can knock them down, leaving me wondering if there is one out there that solves the conundrum, that I lack the imagination to come up with, because there is no grand unified theory readily available to moot all of that. I feel so alone.

I have looked at many models over the years, and I gave you two that have wide acceptance because they embody some degree of parity. As you could see in my example, the two models I described gave two different interpretations of a swap of split locations. In my experience, that is true for every model I've seen, which is why political scientists look to parity as a way to judge instead.

I'm sincere that if you have a counting method that I can't knock down I'd be overjoyed, but my experience in this subject makes me very skeptical of such claims. I've looked at models that are probably like what you describe. What I suspect yours will do is to favor pie-slicing into large counties to keep small counties intact, but I can't say that with certainty without a model to test.

[muon2]

5. Splits small enough to allow a state to avoid exact equality to minimize county splits under SCOTUS rules are best. This allows a state to create two maps, one with county integrity and one with the split as AR did in 2001. When it is clear that one can vary population, then sure you would to avoid chops within the variance constraints. Until then, a chop is a chop. I think when it comes to the size of the chop, for a chop rule, the percentage of a county taken is sufficient. Otherwise focusing on chop size by the size of a CD, you might force an extra chop or crazy CD's without a further overlay of rules. We need to try to minimize the rules where possible.

I bolded my comments.

It's clear to me that you can vary from exact equality up to a range of 1% if the state can show a compelling interest. County and municipal integrity have withstood SCOTUS scrutiny on this point.

OK, we can suggest doing that then as an option. I have absolutely no objection to doing that, although again I am worried that if we have an absolute minimum chop rule, that might create map distortions. On that one though, maybe if staff can come up with maps that minimize chops using the 1% rule, a finding could say thanks but no thanks, the maps all suck, versus ignoring the 1% flex rule. I am worried however, that it introduces another level of complexity. So it should probably be in an addendum.

The problem is that if it is optional the courts will disallow it. It must be consistently applied to be a valid departure from exact equality. However, one can exchange tighter tolerance on the deviations for more regions. That was the point of the table I compiled and displayed earlier in the thread. The commission could say we don't like the 4-region plan, can you produce one with more regions (and hence fewer splits) but with greater deviation up to the 0.5% SCOTUS limit? That I can do, but the commission will have to abide by one of the regional plans' strictures or the rule has no validity.

What is wrong with saying that a 0.5% population variance does not count as a chop, so any excess chop within the 0.5% does not count as exceeding the minimum?  And maybe the absolute minimum chop number should be preferred absent a finding that the forced map or maps suck. Bear in mind here, that the Commission is not made of political pros, so these procedural rules should have more bite than if the Commission were made of partisan pros who know how to game to the max - as opposed to being gamed.

Anyway, isn't this 0.5% thing a side issue to what we are discussing?

Nothing is wrong with that at all. It would be my preference, recognizing that one could generally remove them and still be within the required range in federal court.

As part of a counting method it is not a side issue. If it is part of the counting method then it favors creating regions of whole counties that can take advantage of that rule. A good example of how that comes into play is when you expand to 5% then look at states like OH or KY that have that as part of their mapping rules for the legislature.

I'm still reluctant to say we have these rules, but if you don't like the result you can chuck them. There can be a veto provision, but the veto shouldn't lead to wide open mapping. That just invites trouble. A veto should be tightly constrained, with a backup set of rules (such as my suggestion above) that kick in during a veto.

[muon2]

OK, I went the uber nesting route for Sacto, and using the county based chop count method, I come up with the same number of chops: 3 in Sacto, and 2 in Sutter and El Dorado, in lieu of 2 in Yolo and Placer, so I guess I need your help Mike seeing a map using the county chop count method that loses a chop. I did my best and failed. Thanks.

I would count Sacto here as 0. It's as ideal as it gets, so it counts like a county entirely in a district. I am assuming you aren't counting that way.

I take it you are counting all nested and fractional pieces of Sacto the same. That means Sacto can never have less than three, but small counties can have zero then jump to two with the first split? Is that the method of counting you propose?

If you say that small counties count as one as well since they are served by at least one CD, then you have my second counting method.

In order to take advantage of an uberpacked county, one sometimes needs to shift some other pieces to make the extra chop go away. That's what I was trying to demonstrate in this map. I shifted your Sutter chop into Butte, which already had one and that's when the extra from Sacto vanished.

[muon2]

Thanks for the map Mike (horrible as it is ). Does the chop count change with your second count method? You have now tri-chopped Butte, so that counts as three.

With the second method your signature, your two packings, and my version would all count the same. Consider the counties with changes: Sacto+Yolo+Placer+Sutter+Butte+El Dorado. The total for these is 11 for all four maps. Counting CD 1 in Butte, there are six counties and 6 CDs, and 11 is the optimal number for for this arrangement.

In general the optimal number equals the total number of counties plus the number of districts minus one. You can test your map to see if it passes with this method. If not, we can look for a defect. In either case you may want to review my example for both methods and see if you are comfortable with the second counting method, since your initial reaction was to select the first counting method.

If you are comfortable with counting method two and if you are still agreed to the suggestion that chops equal to less than 0.5% of a district don't count, I can see what optimization that might lead to. There's still the question about encouraging smaller chops for chops over 0.5% of a CD, and if there is a workable threshold.

[muon2]

My concern about a .5% chop not counting as a chop is that it will force twisted sister maps. I suppose staff could come up with the options using that escape hatch to get a minimum chop count, and the Commission could say that they all suck - in other words the Commission could veto the maps with a supra majority vote, just like is required to pass a map. Then you go back to the minimum chop rules.  That plus the supra majority, with a majority from each party, should preclude anything approaching a Michigan style gerrymander.

I still strongly reject however banning the percentage of CD that chops limitation rule.  The 20% limitation of a county rule that is chopped is sufficient, and the Commission should have the power to override that with a finding as well (e.g., it is OK to have a more than 20% chop of SLO in my signature map on the grounds or protecting the Santa Cruz metro area, with a wall on the Ventura County line, because wall placement elsewhere does lead to a SC chop or a crossing of the Tehachapi 's, or over Luther Pass).

I am pleased that counting the number of chops per county method does not bias where the chops are apparently.  Whew!  

So let me summarize the proposed rules.

1. A plan shall minimize county CD pieces. All pieces count towards the total, including one for the case where a county is entirely in a CD. (BTW this works against minimizing the number of counties chopped, since a tri chop counts the same as two dual shops, but that may be OK in CA).

2. When a county is split, the plan shall minimize the number of census places split within the county (ie cities). Census divisions that span a place boundary can be counted on either side of the line.

3. Pieces of counties less than 1 CD, or pieces of places between 20K and 1 CD, shall not be split leaving less than 80% of the county or place in one piece. (I don't think we settled on a number here, so I'm guessing, but you suggested that some small communities may need a larger split.)

4. Pieces that are less than 0.5% of a CD shall not count in assessing rules 1 through 3.

5. Violations of the aforementioned rules require a finding based in federal or state law approved by a supermajority of the mapping body.

Are we ready? Remember, I don't want to draw and then change the rules to suit a change. That's what the biased groups do, and commissions can have biases that aren't partisan.

As an aside, I'm not sure how we want to count the VRA in some areas. For instance, I don't think your Merced-Fresno district would meet standards, since there was a finding that a majority-minority district was required for Fresno, and since a 50% HCVAP district would have been possible by including part of Madera. I estimate that requires 61.5% HVAP or more.

[muon2]

As an aside, I'm not sure how we want to count the VRA in some areas. For instance, I don't think your Merced-Fresno district would meet standards, since there was a finding that a majority-minority district was required for Fresno, and since a 50% HCVAP district would have been possible by including part of Madera. I estimate that requires 61.5% HVAP or more.

That Fresno/Merced CD that I drew is more Hispanic than the one the Commission drew (mine is 58%; theirs is 52.8% HVAP). So that unleashed me, to try to draw a cleaner map. I wonder if there is any commentary on that issue in the Commission's written material, or in the transcripts, all of which will need to be read carefully?  Do you have any idea?  Was it because it was not deemed one community of interest?  Was it because it was impossible to get up to two 50% HCVAP CD's in the area?
[/quote]

Here's what they say in their report:

You must be logged in to read this quote.

Then they go on to say what section 2 districts they created, but they don't list one for Fresno. Instead they put part of Fresno with Kings and Bakersfield, call it a section 5 district and leave it at 49% LCVAP. Merced is then left with a lower 41% LCVAP, linked to Madera and another part of Fresno, and is again called section 5. I think there is a section 2 vulnerability, and it is implied in the MALDEF submittals. Two full section 2 districts could be drawn, but then the commission wouldn't get its cute Fresno-Visalia district at 30% LCVAP.

[muon2]

So let me summarize the proposed rules.

1. A plan shall minimize county CD pieces. All pieces count towards the total, including one for the case where a county is entirely in a CD. (BTW this works against minimizing the number of counties chopped, since a tri chop counts the same as two dual shops, but that may be OK in CA).

2. When a county is split, the plan shall minimize the number of census places split within the county (ie cities). Census divisions that span a place boundary can be counted on either side of the line.

3. Pieces of counties less than 1 CD, or pieces of places between 20K and 1 CD, shall not be split leaving less than 80% of the county or place in one piece. (I don't think we settled on a number here, so I'm guessing, but you suggested that some small communities may need a larger split.)

4. Pieces that are less than 0.5% of a CD shall not count in assessing rules 1 through 3.

5. Violations of the aforementioned rules require a finding based in federal or state law approved by a supermajority of the mapping body.

This is probably buried back in your lengthy discussion, but I don't see any rationale for rule 3.  You're saying that if a county has 600,000 people it can't be split 450,000-150,000 but it can be split 500,000-100,000 (or 500,000-50,000-50,000)?  What is the distinction? 

There are a couple of goals here in search of a rule. One is to encourage smaller chops rather than large chops. The other is to limit the urge to split counties and cities right down the middle. That tends to either dilute or magnify their vote depending on what fraction of a CD they make.

I hope this is the rule Torie massages. It can go a lot of ways. One is to simply say that when there is a choice of cuts, one chooses the one that keeps a bigger fraction of the split entity intact.

[muon2]

You must be logged in to read this quote.

That Fresno/Merced CD that I drew is more Hispanic than the one the Commission drew (mine is 58%; theirs is 52.8% HVAP). So that unleashed me, to try to draw a cleaner map. I wonder if there is any commentary on that issue in the Commission's written material, or in the transcripts, all of which will need to be read carefully?  Do you have any idea?  Was it because it was not deemed one community of interest?  Was it because it was impossible to get up to two 50% HCVAP CD's in the area?

Here's what they say in their report:

You must be logged in to read this quote.

Then they go on to say what section 2 districts they created, but they don't list one for Fresno. Instead they put part of Fresno with Kings and Bakersfield, call it a section 5 district and leave it at 49% LCVAP. Merced is then left with a lower 41% LCVAP, linked to Madera and another part of Fresno, and is again called section 5. I think there is a section 2 vulnerability, and it is implied in the MALDEF submittals. Two full section 2 districts could be drawn, but then the commission wouldn't get its cute Fresno-Visalia district at 30% LCVAP.

Is there pending litigation, or a resolution of litigation, on the Section 2 issue? If you draw a map with two 50% HCVAP CD's in the area, could you put it up again? Thanks. It is still unclear to me, whether inter county Hispanic communities are considered a community of interest, but perhaps you have an opinion on that. The text you quote is a bit loose.

I can only go with what's in the report, vague as it is. In the CA suits the litigants did not challenge the CV districts, so the question did not come up. I presume the logic is that section 5 districts can include coalition and crossover voting, which is not allowed in section 2. This seems like a case where the two sections are in conflict and the commission used the easier standard presented by section 5. We'll see what happens in TX this cycle, but I have a hard time seeing SCOTUS say a weak coalition district will suffice just because it's a section 5 area when a performing district for the single minority is possible under section 2.

[muon2]

So let me summarize the proposed rules.

1. A plan shall minimize county CD pieces. All pieces count towards the total, including one for the case where a county is entirely in a CD. (BTW this works against minimizing the number of counties chopped, since a tri chop counts the same as two dual shops, but that may be OK in CA).

2. When a county is split, the plan shall minimize the number of census places split within the county (ie cities). Census divisions that span a place boundary can be counted on either side of the line.

3. Pieces of counties less than 1 CD, or pieces of places between 20K and 1 CD, shall not be split leaving less than 80% of the county or place in one piece. (I don't think we settled on a number here, so I'm guessing, but you suggested that some small communities may need a larger split.)

4. Pieces that are less than 0.5% of a CD shall not count in assessing rules 1 through 3.

5. Violations of the aforementioned rules require a finding based in federal or state law approved by a supermajority of the mapping body.

This is probably buried back in your lengthy discussion, but I don't see any rationale for rule 3.  You're saying that if a county has 600,000 people it can't be split 450,000-150,000 but it can be split 500,000-100,000 (or 500,000-50,000-50,000)?  What is the distinction? 

There are a couple of goals here in search of a rule. One is to encourage smaller chops rather than large chops. The other is to limit the urge to split counties and cities right down the middle. That tends to either dilute or magnify their vote depending on what fraction of a CD they make.

I hope this is the rule Torie massages. It can go a lot of ways. One is to simply say that when there is a choice of cuts, one chooses the one that keeps a bigger fraction of the split entity intact.

Even if I grant that one should prefer 80-20 chops to 50-50 chops (which isn't at all obvious to me), your rule seems pretty weird to me.  The 80% cutoff is quite arbitrary - you're saying that you'd prefer a plan that splits one county 81-19 to a plan that splits a different county 78-22, even if the first plan requires all sorts of weird contortions (still keeping counties whole) and violations of CoI.  That's a ridiculous standard. 

Splitting a jurisdiction into two relatively even parts is used to dilute their vote if the split jurisdiction can no longer compete against the other parts of the district.

It can also be used when the jurisdiction is large enough or nearly so for its own district but by making a relatively even split the jurisdiction can effectively control two districts.

Either way it is a classic gerrymandering technique. I think it is desirable to limit it by attempting to keep most of a jurisdiction intact. I agree that the 20% number is arbitrary, but that's what we had been batting around. As I just suggested it can be modified to remove the reliance on a specific number yet still having a meaningful impact on a map.

[muon2]

Maybe maps need to be drawn both ways, given the ambiguity in the VRA law - assuming two 50% HCVAP CD's can in fact be drawn, without looking ludicrous.

Here are the CDs I drew for the CV during our Jan '12 exercise. The Merced-Fresno CD is 62.0% HVAP which should be about 50.5% HCVAP. The Bakersfield-Kings-Tulare CD is 65.3% HVAP which is equivalent to 50.1% HCVAP for that area.