Gerrymandering again

[Worldtraveller]

This time, in NC. The new, for some reason, approved, plan has been released.

DRA Analysis: 2024 North Carolina voting districts

www.commoncause.org

It's basically going to all but shut the Dems out of the state representation. From the link above, it gives one example of a district broken down by the criteria used to grade the fairness of the district. Note that in proportionality, it gets a zero.
from the site:

Proportionality is one measure of how fair a map is. In a proportional map, if 50% of North Carolinians voted for Republicans to represent them in Congress and 50% voted for Democrats, the map’s number of leaning Republican or Democratic seats should be about even.

When are we going to get some kind of national referendum to get rid of this most undemocratic process?

 

[Jimmy Higgins]

Their gerrymandering in local state legislatures is where the real power is. They turn purple states into Alabama.

 

[Tigers!]

All the more reason to note that the only criteria to be used in determining electoral boundaries should be the number of electors.
Any other criteria used esp. previous or future voting intentions etc. will lead to gerrymandering

 

[Worldtraveller]

Yeah. I wonder what congress would look like if every state just had an open ballot, and people voted for the number of representatives they have for that state. I realize a lot of states would probably lose "local" representation, but that's something that went out the window a long time ago.

 

[lpetrich]

There is a very simple solution to gerrymandering.

 Proportional representation

I've seen a little advocacy of it, but not much.

PR is successfully used by several countries, and users of PR often have more than two parties, often with an effective number well over two. When elected, the parties then negotiate governing coalitions.

 

[Jimmy Higgins]

All the more reason to note that the only criteria to be used in determining electoral boundaries should be the number of electors.

As long as we ignore 200 or so years of history and what people in power have done, that will work out just fine.

Also, just looking at a state like Ohio, using number of electors doesn't particularly help. You can go from 3 D / 12 R to 6 D / 9 R +/-1. There is geography, economics, etc... that help make up reasonable districting.

 

[lpetrich]

How to do proportional representation? Let us consider how to do elections in general. First single-seat or single-winner, then multiseat or multi-winner.

 Comparison of electoral systems and Electowiki, the election methods wiki

First-past-the-post or plurality voting: a single-seat method that is simple, dumb, and vulnerable to the spoiler effect for more than two candidates. One votes for one candidate out of all of them.

For doing FPTP with multiseat bodies, one may divide the body's region into electoral districts, each of which elects one candidate. One may also vote for slates of candidates in single-winner fashion - general ticket - or else vote for only one candidate: single non-transferable vote.

Voting for more than one candidate is approval voting, either single-seat or multiseat. For multiseat elections, this is often restricted by having at most votes as there are seats: bloc voting.

Approval voting can be generalized to partial-sized votes, thus giving us range, rated, or score voting. For multiseat elections, this is often restricted by having a maximum sum of ratings: cumulative voting.

In a partisan vote, bloc vote reduces to FPTP general ticket, approval voting to approval-vote GT, and rated voting to rated-vote GT.

In ranked voting, one orders the candidates in order of preference: first choice, second choice, ...

There are several algorithms have have been devised for counting ranked votes, each of which has its pluses and minuses.

The Borda count is turning rankings into ratings, then counting those up. In a partisan multiseat vote, this method reduces to Borda-count GT.

Instant runoff voting is often presented as ranked-choice voting, even though it's the ballots that are ranked. If a candidate has a majority of the top-rank vote, and not just a plurality, that candidate wins. Plurality is having the most of all the candidates, without necessarily having a majority. Otherwise, drop from the count the candidate with the fewest top-rank votes and repeat the count until one candidate gets a majority. For multiseat elections, one can stop the count when the number of candidates is equal to the number of seats, but in a partisan vote, this reduces to IRV GT.

Condorcet methods feature a matrix of candidate vs. candidate performance, a virtual round robin. There are various ways of disambiguating cyclic preferences. In a partisan multiseat vote, this reduces to that method of GT.

 

[lpetrich]

How can avoid general ticket in partisan multiseat elections and get proportional rrepresentation?

With these methods, there is a way. Count in rounds, use the winner in each round, and then remove that candidate from the next rounds. Without the next step, one gets something that reduces to general ticket, just like selecting all the top candidates at once. That step is to downweight the ballots that helped elect that winner. With suitable downweighting, one gets proportional representation. An example is  Sequential proportional approval voting though its downweighting method is completely general.

There is an alternative way that's typically used in multiseat instant runoff voting: single transferable vote. One establishes a victory quota or count:

V = (total votes) / ( (number of seats) + 1 )

and then accepts the candidate that has the most top-rank votes if that number of votes is at least V. The downweighting part is either random removal of V winner-vote ballots from the count or else multiplies those ballots' weights by

(1 - V / (total top-rank votes for the winner) )

(all the ballots initially had weight 1)

If no candidate is above threshold, then drop losers as in IRV until a candidate is above threshold. One stops when all the seats are filled, or if some are unfilled, one uses the most recently dropped losers to fill those ones.

A bit complicated, but it works. It also does not work well with districts with large numbers of seats. So one will have to divide up a multiseat body into districts with small numbers of seats, like 3 to 5 each.

 

[Tigers!]

All the more reason to note that the only criteria to be used in determining electoral boundaries should be the number of electors.

As long as we ignore 200 or so years of history and what people in power have done, that will work out just fine.

Also, just looking at a state like Ohio, using number of electors doesn't particularly help. You can go from 3 D / 12 R to 6 D / 9 R +/-1. There is geography, economics, etc... that help make up reasonable districting.

If you are worried about things such as 3D/12R TO 6D/9R then you are not looking at just the no. of electors. I am not particularly up to scratch as to that symbolism actually means but presumably it is the purported change in voting intentions.
When you have past voting intentions, future voting intentions, skin colour, ethnicity, religion, economics, <insert non-numeric criteria here> influencing/affecting, no matter how big or small, the location of boundaries then you are gerrymandering.
Boundaries should be based SOLELY upon the no of electors +/-%, with one exception. Geography does affect boundaries. In Australia state electorates cannot cross state boundaries. Victorian state electoral boundaries - 2014
In Australia we do not care how you voted last time, how you will vote next time, your skin colour etc. when drawing boundaries. Only number of noses count. And we do NOT allow pollies to be involved.
It seems in the US that pollies, parties have far too much say. And spare me the whataboutery that R is worse than D. Both seem to do it and need to be kicked out of the process.

Add in compulsory marking of the electoral roll and you can do well to eliminate gerrymandering.

 

[Tigers!]

There is a very simple solution to gerrymandering.

 Proportional representation

I've seen a little advocacy of it, but not much.

PR is successfully used by several countries, and users of PR often have more than two parties, often with an effective number well over two. When elected, the parties then negotiate governing coalitions.

So long as it is acknowledged that PR is not the panacea as it is so often portrayed. PR will make it harder for a tyrant, autocrat to be elected but not impossible. Vigilance is still key.

The problem will PR is that what you voted for may not be what you get.

 

[Tigers!]

How can avoid general ticket in partisan multi-seat elections and get proportional representation?

With these methods, there is a way. Count in rounds, use the winner in each round, and then remove that candidate from the next rounds.

The disadvantage of multiple rounds is always the tiresome nature of it. One round is always best, if possible.

 

[bilby]

All the more reason to note that the only criteria to be used in determining electoral boundaries should be the number of electors.

As long as we ignore 200 or so years of history and what people in power have done, that will work out just fine.

Also, just looking at a state like Ohio, using number of electors doesn't particularly help. You can go from 3 D / 12 R to 6 D / 9 R +/-1. There is geography, economics, etc... that help make up reasonable districting.

I suspect that Tigers was using "electors" as a synonym for "voters", as he lives in the civilised world where "electors" are not a minuscule elite who wield power "on behalf of" the voters, just in case the voters give the wrong answer when asked their opinions.

 

[lpetrich]

Single transferable vote is not used very much, and the most common route to proportional representation is to vote for parties, and then each party gets seats in proportion to how many votes it received. That is often called the party-list method after the tradition of parties publishing lists of candidates that they want seated.

Since the number of candidates seated must be a nonnegative integer, there are various algorithms for going from numbers of votes to numbers of seats. Such algorithms have also been used for similar kinds of allocation, like allocating the seats of the US House of Representatives between the US states, where the "parties" are the states and the "voters" are the states' total populations.

Party list has several variations. In closed-list systems, the parties choose all the candidates, while in open-list systems, voters can choose which candidates they want seated first. In mixed-member systems, the body is divided between district seats, with one member for each district, and list seats, who are selected to make the overall count proportional. Parallel systems are similar, but with the list seats only proportional among themselves.

 

[lpetrich]

The problem will PR is that what you voted for may not be what you get.

Like what?

 

[lpetrich]

I suspect that Tigers was using "electors" as a synonym for "voters", as he lives in the civilised world where "electors" are not a minuscule elite who wield power "on behalf of" the voters, just in case the voters give the wrong answer when asked their opinions.

Are you talking about the US Electoral College? That system is a half-baked aggregated and weighted quasi-popular vote, and not quite what the Founders had in mind: The Avalon Project : Federalist No 68 They didn't anticipate it becoming a rubber-stamp body that uses winner-take-all, though with two states being partial exceptions.

No other nation in the world uses anything like it, and none of the US's subdivisions do so. Some nations have indirect elections of their heads of state, but those are all elections by legislature and sometimes also legislature delegates.

 

[Loren Pechtel]

All the more reason to note that the only criteria to be used in determining electoral boundaries should be the number of electors.

As long as we ignore 200 or so years of history and what people in power have done, that will work out just fine.

Also, just looking at a state like Ohio, using number of electors doesn't particularly help. You can go from 3 D / 12 R to 6 D / 9 R +/-1. There is geography, economics, etc... that help make up reasonable districting.

Which is why I favor a mathematical approach.

Chop the state up into small blocks. Perhaps one address (for multi-family locations) or one physical block (every intersection divides them up) of houses. The boundaries of every block carry a weight that is defined by how closely associated it is with the next block. The boundary weight is based on how readily people move across it. Land that precludes the ready flow (protected land, rivers, unsafe land etc) has a boundary weight of zero. The weight to the continuation of our street is high (you can take one step to the next person's property), the weight to the neighbors behind us is somewhat lower (it would require going about a block and a half to reach them. Fences are concrete block and theirs is raised for a pool--unless somebody is on a ladder it's pretty much impossible to talk over the back fence.)

The proper redistricting is the arrangement that minimizes boundary weight while maintaining a required equality in population. Anyone can challenge the boundaries by showing a better solution. Obviously, it can be gamed somewhat by choosing weights--but require that no change can go into effect until after the election at least a year from now makes it very hard to do so.

 

[Tigers!]

The problem will PR is that what you voted for may not be what you get.

Like what?

Quite a number of years ago when the UK had a brief flirtation with introducing PR a joke when around about the possible unforeseen side effects of PR

There was an PR election with 4 candidates - A, B, C, D. But only 3 would be elected.
(note it does not matter which candidates won for this exercise)
Each candidate promised each elector 1 week paid holiday if they were elected.
Candidate A promised a holiday at Brighton
Candidate B promised a holiday in the Lakes District
Candidate C promised a holiday at Blackpool
Candidate D promised a holiday on the Isle of Wight
The election was duly held - Candidates A, C, D were elected.
The voters waited to hear when they would get their holidays and where (either Brighton, Blackpool, Isle of Wight).
Finally it was announced that the electors would get 5 days in Cardiff.

PR involves much haggling and horse trading behind the scenes as often contradictory promises must be kept. Plus dividing the spoils of office is hard work. You may not get nor like what you elected.

Plus let us always remember that Hitler became Chancellor of Germany under the Weimar republic's PR system. (Before anyone jumps on me I am well aware that Hitler was selected rather elected). But his party was numerically the largest in the lower house. Horse trading can have unforeseen consequences.

PR all to often produces unstable coalitions
Italy had 48 governments in the first 51 years after elections were introduced after WW2.(I think that is accurate)
Netherlands government falls
Israel had 5 general elections between 2019-2022
Belgium managed 589 days without a government

PR is useful but not the be all to end all concerning elections.

 

[Jimmy Higgins]

All the more reason to note that the only criteria to be used in determining electoral boundaries should be the number of electors.

As long as we ignore 200 or so years of history and what people in power have done, that will work out just fine.

Also, just looking at a state like Ohio, using number of electors doesn't particularly help. You can go from 3 D / 12 R to 6 D / 9 R +/-1. There is geography, economics, etc... that help make up reasonable districting.

Which is why I favor a mathematical approach.

Chop the state up into small blocks. Perhaps one address (for multi-family locations) or one physical block (every intersection divides them up) of houses. The boundaries of every block carry a weight that is defined by how closely associated it is with the next block. The boundary weight is based on how readily people move across it. Land that precludes the ready flow (protected land, rivers, unsafe land etc) has a boundary weight of zero. The weight to the continuation of our street is high (you can take one step to the next person's property), the weight to the neighbors behind us is somewhat lower (it would require going about a block and a half to reach them. Fences are concrete block and theirs is raised for a pool--unless somebody is on a ladder it's pretty much impossible to talk over the back fence.)

The proper redistricting is the arrangement that minimizes boundary weight while maintaining a required equality in population. Anyone can challenge the boundaries by showing a better solution. Obviously, it can be gamed somewhat by choosing weights--but require that no change can go into effect until after the election at least a year from now makes it very hard to do so.

We don't need a new system for redistricting. Computers can do it already, they can be programmed to uber-gerrymander to fair map.

What we need are adults to do the redistricting, not partisan politicians. Courts throw a map out there (at until that entire process gets perverted), the maps appear fair. It is too bad, at this point, that California isn't gerrymandered to the point where it is 10 to 1 D's in the House from that state. Florida, North Carolina, Texas, Ohio (until they got too greedy) are all gerrymandered like crazy. Other states employ it as well, but we are talking a seat or two. Where as when PA had to change the map, that was 3 or 4 seats from one state.

 

[Tigers!]

All the more reason to note that the only criteria to be used in determining electoral boundaries should be the number of electors.

As long as we ignore 200 or so years of history and what people in power have done, that will work out just fine.

Also, just looking at a state like Ohio, using number of electors doesn't particularly help. You can go from 3 D / 12 R to 6 D / 9 R +/-1. There is geography, economics, etc... that help make up reasonable districting.

Which is why I favor a mathematical approach.

Chop the state up into small blocks. Perhaps one address (for multi-family locations) or one physical block (every intersection divides them up) of houses. The boundaries of every block carry a weight that is defined by how closely associated it is with the next block. The boundary weight is based on how readily people move across it. Land that precludes the ready flow (protected land, rivers, unsafe land etc) has a boundary weight of zero. The weight to the continuation of our street is high (you can take one step to the next person's property), the weight to the neighbors behind us is somewhat lower (it would require going about a block and a half to reach them. Fences are concrete block and theirs is raised for a pool--unless somebody is on a ladder it's pretty much impossible to talk over the back fence.)

The proper redistricting is the arrangement that minimizes boundary weight while maintaining a required equality in population. Anyone can challenge the boundaries by showing a better solution. Obviously, it can be gamed somewhat by choosing weights--but require that no change can go into effect until after the election at least a year from now makes it very hard to do so.

Why complicate the exercise? You admit that weighting can be manipulated. Why then add any weighting at all?
Purely numeric. Each electorate has N +/- %N persons within.

 

[lpetrich]

Quite a number of years ago when the UK had a brief flirtation with introducing PR a joke when around about the possible unforeseen side effects of PR

There was an PR election with 4 candidates - A, B, C, D. But only 3 would be elected.
(note it does not matter which candidates won for this exercise)
Each candidate promised each elector 1 week paid holiday if they were elected.
Candidate A promised a holiday at Brighton
Candidate B promised a holiday in the Lakes District
Candidate C promised a holiday at Blackpool
Candidate D promised a holiday on the Isle of Wight
The election was duly held - Candidates A, C, D were elected.
The voters waited to hear when they would get their holidays and where (either Brighton, Blackpool, Isle of Wight).
Finally it was announced that the electors would get 5 days in Cardiff.

I don't get that joke.

PR involves much haggling and horse trading behind the scenes as often contradictory promises must be kept. Plus dividing the spoils of office is hard work. You may not get nor like what you elected.

So what? This sort of haggling can also happen inside of parties. Consider in the US House, back in 2021, Joe Manchin and Kyrsten Sinema vs. Build Back Better, and last year, not one but two nasty fights over who will be Speaker. A two-party system is supposed to be immune to such nastiness, right?

Plus let us always remember that Hitler became Chancellor of Germany under the Weimar republic's PR system. (Before anyone jumps on me I am well aware that Hitler was selected rather elected). But his party was numerically the largest in the lower house. Horse trading can have unforeseen consequences.

So what? I've seen that as an argument against democracy in general.

PR all to often produces unstable coalitions ... (Belgium, Netherlands, Italy, Israel)

As if that is typical of PR systems. I think that PR is good because it's a good way of sidelining extremists.