Voting system and reform thread

Discussions worth keeping around later.
User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Voting system and reform thread

Post by mèþru »

A general thread about discussing democratic electoral systems, hypothetical or not.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Voting system and reform thread

Post by mèþru »

I wrote:I'm making yet another system(s) to vote for the President of the United States. What would be a disqualifying percentage of voters selecting the "Don't know enough about the candidate" option on the ballot if such an option existed?
Frislander wrote:
mèþru wrote:I'm making yet another system(s) to vote for the President of the United States. What would be a disqualifying percentage of voters selecting the "Don't know enough about the candidate" option on the ballot if such an option existed?
Certainly no higher than 50% imho, and I'd be tempted to make it 40%.
I was thinking of making it somewhere 15-40%. (Yes that's a broad range.)
Last edited by mèþru on Sun Mar 05, 2017 7:29 pm, edited 1 time in total.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

zompist
Boardlord
Boardlord
Posts: 3368
Joined: Thu Sep 12, 2002 8:26 pm
Location: In the den
Contact:

Re: Voting system and reform thread

Post by zompist »

Can you step back a bit and answer a few questions?

- What problem are you trying to solve?
- Are you trying to disqualify voters, or votes, or candidates?

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Voting system and reform thread

Post by mèþru »

The criteria, ranked by importance:
  1. People who don't want the office shouldn't be considered in the voting process.
  2. The winning candidate should be the one most people are okay with.
  3. Get rid of the whole ballot access system and let people vote for whomever they want.
  4. There should be a consensus among states (possibly plus the territories), in which any candidate vehemently rejected by one area should not be elected.
The criteria don't necessarily have to be filled, but I want to at least address them.

I'm trying to eliminate candidates with the "Don't know enough about the candidate". It will be a system based off of my favourite system for single-winner elections, majority judgement.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
Salmoneus
Sanno
Sanno
Posts: 3197
Joined: Thu Jan 15, 2004 5:00 pm
Location: One of the dark places of the world

Re: Voting system and reform thread

Post by Salmoneus »

mèþru wrote:The criteria, ranked by importance:
  1. People who don't want the office shouldn't be considered in the voting process.
  2. The winning candidate should be the one most people are okay with.
  3. Get rid of the whole ballot access system and let people vote for whomever they want.
  4. There should be a consensus among states (possibly plus the territories), in which any candidate vehemently rejected by one area should not be elected.
The criteria don't necessarily have to be filled, but I want to at least address them.

I'm trying to eliminate candidates with the "Don't know enough about the candidate". It will be a system based off of my favourite system for single-winner elections, majority judgement.
1. This is obvious
2. This is vague - both 'most' and 'okay with' are highly imprecise
3. This is grotesquely undemocratic. You would be asking people to meaningfully cast ballots on a paper a hundred pages thick - participation would plummet*, and the results would have a high degree of randomness (because when you ask people to vote like that, they become erratic - the order of names, for instance, becomes very important, as they refuse to think seriously about candidates after a certain point)
4. That seems a terrible idea. For one thing, you could have an election where nobody won. For another, it would enable winners whom everybody disliked to beat those whom almost everybody loved. And, for another, there would instantly be an armed uprising if you tried it. "Clinton won the vote by at least a 70% margin in 49 states, but people in Wyoming really hated her because of something she said about the state last year, so we're appointing Trump as president!" - that's not a practical thing to try to tell the electorate.

*participation is particularly an issue under Majority Judgement, of course, because MJ actively discourages participation. Voters will often be better served by staying at home, and candidates will very often be better served by most voters staying home. So political ads will be "just stay home this year, we'll sort it out for you!" - hardly democratic. [consider: Candidate A gets ratings of (just taking a 10-point scale for sake of argument and not worrying about words) a 4, a 4, a 6, a 10 and a 10, giving a median of 6.8. But if they can persuade the 6 voter, theoretically a supporter, to instead stay at home, then their median rating actually improves, to 7! Similarly, if your voters are giving your rival anything less than a damning rating, it's more important to persuade them to stay at home than it is to persuade them to come out and support you. Parties will endeavour to ensure that only their most fanatical voters actually turn out to vote.]

5: remove candidates with too many "don't knows". This seems a bad idea! Let's say: Clinton wins the popular vote in every single state. But Trump gets about 1/3 of his supporters to vote "don't know" on Clinton. Bingo, that's 15% of the electorate, so Clinton gets disqualified. More generally, if you attach such power to a "don't know" vote, you force voters to vote "don't know" strategically, because it's more likely to give them the result they want than a sincere vote is.

6: have Majority Judgement. This is incompatible with having winning candidates most people are OK with! Indeed, MJ even fails the condorcet loser criterion. That is, consider: you have candidates A, B, C, D, and E. Most people prefer A to E. Most people prefer B to E. Most people prefer C to E. Most people prefer D to E. Most people, in other words, would prefer any other candidate to E. But under Majority Judgement, E will sometimes win the election! Now, try explaining that to voters! Under real world conditions, where people sometimes vote strategically, Majority Judgement will sometimes even fail the majority criterion - more than 50% of the population can want A to win, yet A ends up losing. And MJ also fails the consistency criterion. In other words, you could have a situation where the newsreader has to say "Clinton beat Trump in every single one of the 50 states of the Union, a comprehensive whitewash, with Trump not having the upper hand in even just one single state, in a result Democrats are calling a historic repudiation of the Republican agenda. In unrelated news, Donald Trump has been elected President." Again, the electorate is unlikely to understand or appreciate such a result, no matter what the mathematical advantages of the system may be.

In practice, I strongly suspect MJ would devolve into SP pretty quickly. If you want Clinton to win the election, the rational strategy is to give the highest rating to clinton and the lowest rating (or a "don't know") to everybody else. Not only does giving your true opinion of Clinton make her more likely to lose to a candidate you like even less, but giving even a lukewarm opinion about a third party candidate could make a Clinton defeat more likely. The first rule parties would drum into their voters is "don't give anyone else any shred of support" - MJ would work, inasmuch as MJ can ever work, for an election or maybe two at most, before people worked out why the results were so completely out of line with popular opinion, and would revert to effective SP.

I suspect MJ works great when a) you have a small electorate where everybody knows all of the candidates well, and b) nobody is particularly passionate or partisan and doesn't really care who wins. A literature award, for instance, it might be fine for. But in real politics, I think it would be at first disastrous, and soon after irrelevent.



Why do you think MJ is better than, say, Borda, AV, cloneproof Schwartz sequential dropping, or the cumulative vote?
Blog: [url]http://vacuouswastrel.wordpress.com/[/url]

But the river tripped on her by and by, lapping
as though her heart was brook: Why, why, why! Weh, O weh
I'se so silly to be flowing but I no canna stay!

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Voting system and reform thread

Post by mèþru »

Clarification:
2. I mean that the winner should be someone that as many people in the nation can accept (not bad, not necessarily good) as president as possible. Even if that person doesn't have a majority.
3. I am not sure how I will actually do this, but I am already aware of the problem of too many candidates. I'm trying to cut down on that. One method I made shortly before posting made all candidates write-in, with all states having at least four spaces to write in candidates. States that currently have more than four registered parties must have as many spaces as registered parties to write-in candidates. States may also pass legislation to increase or decrease the amount of spaces, as long as four spaces remain.
4.
I wrote:The criteria don't necessarily have to be filled, but I want to at least address them.
I was mainly thinking of this when writing it. But, yeah. If Donald Trump fulfills 1, 2, and 4, while Hillary Clinton fulfills only 1 and 2, he wins. If Hillary Clinton fulfills only 1 and 2 while Donald Trump fulfills 1 and 4, then I guess that Donald Trump shouldn't win. I need to make some changes.
Salmoneus wrote:*participation is particularly an issue under Majority Judgement, of course, because MJ actively discourages participation. Voters will often be better served by staying at home, and candidates will very often be better served by most voters staying home. So political ads will be "just stay home this year, we'll sort it out for you!" - hardly democratic. [consider: Candidate A gets ratings of (just taking a 10-point scale for sake of argument and not worrying about words) a 4, a 4, a 6, a 10 and a 10, giving a median of 6.8. But if they can persuade the 6 voter, theoretically a supporter, to instead stay at home, then their median rating actually improves, to 7! Similarly, if your voters are giving your rival anything less than a damning rating, it's more important to persuade them to stay at home than it is to persuade them to come out and support you. Parties will endeavour to ensure that only their most fanatical voters actually turn out to vote.]
Your example actually doesn't work. Majority judgement doesn't take the average of values in the case of an even number of votes. It takes the lower one. If the voter votes, the median is 6. If they don't, it's a 4. Also, trying to make only fanatics vote makes it more likely that an extremist party would win, so I think all non-extremist parties will encourage their voters to vote.

So I guess the don't knows are a bad idea. I won't use that, unless if you have an idea for how to make it work better.
Salmoneus wrote:This is incompatible with having winning candidates most people are OK with! Indeed, MJ even fails the condorcet loser criterion. That is, consider: you have candidates A, B, C, D, and E. Most people prefer A to E. Most people prefer B to E. Most people prefer C to E. Most people prefer D to E. Most people, in other words, would prefer any other candidate to E. But under Majority Judgement, E will sometimes win the election!
This is a feature, not a bug. Let's say the voters of A hate B and vice versa. Let's say that a large minority of A voters like C, and a much smaller minority like D. Let's say that a large minority of B voters like D, and a much smaller minority like A. The voters of C and D hate A and B, but a majority of the entire electorate doesn't like C and D. Barely anyone likes E, but E has few haters. E wins, because it is the compromise candidate everyone can at least accept.

What is SP? This is why I'm avoiding abbreviations.

Majority judgement is better than ranked systems because you can put multiple candidates on the same ranking and can jump between rankings to express the distance of support. In short, in expresses the level of support a voter has for each candidate. I suspect, that under such a system, most voters will. Such a system would have to have rank limit, however, or else there is the problem of people invoking 1-1000 scale voting which messes up calculation and makes any tactical voting very dangerous. A small number of given grades, in my point of view, helps simplify the system mathematically. Giving only two grades, however, eliminates much of the advantage over ranked systems. Cumulative voting restricts how many candidates you can vote for (which I guess my system does as well, but on a lesser scale).
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Voting system and reform thread

Post by mèþru »

Since writing the above post, I got an idea for an alternative idea to majority judgement. You give each candidate an integer value. The highest absolute value is set as a denominator for all the values. Find the mean score of all candidates combined, and add that value to the candidates until all have the same amount of votes. Finally, find the median value for each candidate as per the rules majority judgement.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

Axiem
Avisaru
Avisaru
Posts: 260
Joined: Tue Oct 22, 2013 8:15 pm

Re: Voting system and reform thread

Post by Axiem »

How much research have you done into the research and work other people have done towards voting systems and their problems?

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Voting system and reform thread

Post by mèþru »

Not much, but I'm hoping to learn.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

zompist
Boardlord
Boardlord
Posts: 3368
Joined: Thu Sep 12, 2002 8:26 pm
Location: In the den
Contact:

Re: Voting system and reform thread

Post by zompist »

mèþru wrote:Since writing the above post, I got an idea for an alternative idea to majority judgement. You give each candidate an integer value. The highest absolute value is set as a denominator for all the values.
You have to think about these things like a petty-minded QA analyst determined to break the system. So: if you allow people to pick their own scale, some people will immediately go to a scale of 1 to 1,000,000,000.

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Voting system and reform thread

Post by mèþru »

I wrote:The highest absolute value is set as a denominator for all the values
I meant in each individual's ballot. Not for everyone's ballot. This is the key thing that prevents people from breaking the system.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
Salmoneus
Sanno
Sanno
Posts: 3197
Joined: Thu Jan 15, 2004 5:00 pm
Location: One of the dark places of the world

Re: Voting system and reform thread

Post by Salmoneus »

mèþru wrote:Clarification:
2. I mean that the winner should be someone that as many people in the nation can accept (not bad, not necessarily good) as president as possible. Even if that person doesn't have a majority.
With respect, that's no clearer than before. What does "not accepting" a President mean? So far, America has never had a President that the nation has not accepted, with the exception of Abraham Lincoln. So on the one hand, if Lincoln is the only President you're aiming to prevent, are you really solving a real problem? Every other election has been fine, after all. But on the other hand, do you really want a system that prevents the election of Lincoln?

I know there's unhappiness with the US system, but I don't think "our system is terrible, because it elected Lincoln!" is really the first objection most people have...

More broadly, "the one most people are OK with" suggests that what you really want is President Transformers: a candidate that pleases nobody and does nothing well. I don't think it's particularly obvious that "I recognise his name but have no clear opinion about him, so I guess I don't hate him!" is the qualification people really want in a President!

3. I am not sure how I will actually do this, but I am already aware of the problem of too many candidates. I'm trying to cut down on that. One method I made shortly before posting made all candidates write-in, with all states having at least four spaces to write in candidates. States that currently have more than four registered parties must have as many spaces as registered parties to write-in candidates. States may also pass legislation to increase or decrease the amount of spaces, as long as four spaces remain.
What is gained by this, other than disenfranchising poorer (and disproportionately non-White) voters, who are the ones who a) are more likely to fail to correctly remember and spell their candidate's name, and b) are more likely to be put off voting by the introduction of such a punitively difficult voting method? Except, I suppose, that it would also benefit the two major parties (because you're putting more burden on name-recognition - you wouldn't be able to, say, turn up and vote for the Constitutional Party, you'd have to have learnt who their candidate was first), and would also benefit advertising executives (because drilling the names (and spellings) in would be more important that debating policy).
Your example actually doesn't work. Majority judgement doesn't take the average of values in the case of an even number of votes. It takes the lower one. If the voter votes, the median is 6. If they don't, it's a 4. Also, trying to make only fanatics vote makes it more likely that an extremist party would win, so I think all non-extremist parties will encourage their voters to vote.
Sorry, I was thinking of means, not medians. However, although the specific numbers don't work out, the underlying point remains: MJ fails the participation criterion (voters are generally better off not voting (unless they vote hyper-tactically)). And parties are better off having some people not vote for them.

[New numbers: 4, 4, 8, 8, 10, 10, 10, 10. Median (rounding down): 8. But take out that 8 vote (theoretically a really strong supporter), and the median goes up to 10.]

This also illustrates, incidentally, another huge problem with your system: it counts the votes of extremists as more valuable than those of centrists. If Democrats, say, give their candidate all 7s and their rival all 4s, they will lose every time to Republicans giving their candidate all 10s and their rivals all 1s. [Three of each voter: Trump gets 4, 4, 4, 10, 10, 10, median rounds down to 4. Clinton gets 1, 1, 1, 7, 7, 7, median rounds down to 1. Trump wins. Add one Democrat and one Third Party (who gives both of them a 3): 3444-4XXX, median 4, vs 1113-7777, median 3 - Trump now has the votes of only 3/8ths of the population, but still wins the election, not because of any actual advantage in popularity but purely because his voters have been primed to hate the opposition more. Do you really want to be handing out elections to reward the party who most succesfully polarises the electorate and fills them full of hatred and conspiracy theories? I'm not sure that's what America really needs right now.

Then again, maybe that's the best case scenario. Because if the two parties fail to spend enough money filling the electorate with hatred toward every single third party, there's a risk that Joe the Dentist from Spokane gets a couple of write-in '5' ballots, his median beats both Trump and Clinton, and then you've got some random guy most people haven't heard of and who has sod-all political experience and no allies in Congress being given the Presidency. Is that a good thing? And do you really think the electorate would stand idly by and tolerate this outcome, while you patiently explain "but our mathematics prooves that he deserves to win, for reasons you probably won't understand!" ? I don't think that's a viable process in the real world!]
So I guess the don't knows are a bad idea.
Oh, but you have to. No seriously, you really do (well, it doesn't need to be on the ballot - in your write-in system people just wouldn't write them in). It's a terrible idea, but it's better than the alternative, where five drunk guys in Spokane are the only people who have heard of Joe Wobblewazzard, are the only people who write his name in, and manage to give Mr Wobblewazzard (whose entire platform is the reintroduction of slavery and the prohibition of asparagus) a perfect median 10 score, comfortably beating all the other candidates.

The only viable way to do this is to get rid of the write-in idea and force everybody to express an opinion on every candidate. But then you're not really measuring their reasoned opinions, you're measuring how they react when asked to give lots of opinions on people they have no opinion on, which is hardly a fair way to decide the Presidency.

This is why this system may work fine for book clubs, but isn't really suited to governing the country.
Salmoneus wrote:This is incompatible with having winning candidates most people are OK with! Indeed, MJ even fails the condorcet loser criterion. That is, consider: you have candidates A, B, C, D, and E. Most people prefer A to E. Most people prefer B to E. Most people prefer C to E. Most people prefer D to E. Most people, in other words, would prefer any other candidate to E. But under Majority Judgement, E will sometimes win the election!
This is a feature, not a bug. Let's say the voters of A hate B and vice versa. Let's say that a large minority of A voters like C, and a much smaller minority like D. Let's say that a large minority of B voters like D, and a much smaller minority like A. The voters of C and D hate A and B, but a majority of the entire electorate doesn't like C and D. Barely anyone likes E, but E has few haters. E wins, because it is the compromise candidate everyone can at least accept.
And then there are riots in the streets. Because when you go and tell people "so, the majority of voters prefered Clinton over Trump. So we're making Trump the President,", they will not be happy. They weren't very happy this time, and they'll be even less happy when this is advertises as a "feature" not a "bug".
[Imagine, for instance, 3 Dems, 3 Repubs, 4 right-leaning Inds, and 4 left-leaning Inds. All Dems go 9-3 for Clinton, all Reps go 9-1 for Trump, all LLIs go 5-4 for Clinton, 3 RLIs go 3-2 for Clinton, and 1 RLI goes 4-3 for Trump. Clinton gets 1113333-5555999, median 3, while Trump gets 2222224-4444999, median 4. In this scenario, 4 out of 14 voters prefer Trump to Clinton and 10 out of 14 voters prefer Clinton to Trump - this would be a historic 71% for Clinton (the largest victory in US history so far was 61%). It would make victories like Reagan's re-election (only 58% of the vote!) look like narrow, close-run things. It would be a politics-shatteringly conclusive victory for the Democrats.
And your electoral system would hand it to the Republicans. I don't think the electorate would like being told that 70% of them had prefered Clinton, but that you were going to hand the Presidency to Trump anyway. And they'd like it even less when you told them that ignoring the popular vote in this way was "a feature, not a bug" of the new system.]

[Oh, and that example also illustrates how MJ is incompatible with the "winner most people are OK with" criterion. Because in that example, only 3 out of 14 voters even gave Trump a higher-than-50% approval, and he still won. Whereas 7 out of 14 gave clinton at least 50% approval, and she lost]
What is SP? This is why I'm avoiding abbreviations.
Oh, sorry. Simple Plurality. It's the system that you currently use.
Majority judgement is better than ranked systems because you can put multiple candidates on the same ranking and can jump between rankings to express the distance of support. In short, in expresses the level of support a voter has for each candidate.
Why is that a good thing?[/quote] I suspect, that under such a system, most voters will.
They may, but if they do they'll throw away the election.
Cumulative voting restricts how many candidates you can vote for (which I guess my system does as well, but on a lesser scale).
Again, why is that a bad thing? That's just a fact about the voting process, not an assessment of its effects.

[Not that I'm supporting CV either. Just curious what analysis you've done to make MJ seem like the best system.]
Blog: [url]http://vacuouswastrel.wordpress.com/[/url]

But the river tripped on her by and by, lapping
as though her heart was brook: Why, why, why! Weh, O weh
I'se so silly to be flowing but I no canna stay!

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Voting system and reform thread

Post by mèþru »

Here's a draft from before I started the thread:
More: show
Notes:
  • states meaning either the areas with separate representation in the Electoral College or meaning all states as well as the inhabited territories of the US. I prefer the first with the condition that the US Virgin Islands and Puerto Rico is are admitted as a single state (with the state government being a loose confederation of Puerto Rico and the Virgin Islands’ current governments so that they don’t interfere with each other) and that the Constitution starts being applied to all territory currently under American administration (they can vote in the presidential election).
  • The method used to elect the Vice President is the same unless otherwise noted.
Method 1

Ballot:
A minimum of four lines with spaces to write candidate names, with a series of four boxes following each line. Individual states may have more lines as per state laws. States with more than four registered parties at the time of the reform will have as many lines as registered parties. Above each box, there is a label and number. The label says, in whichever language(s) the states choose, excellent above the first box, good above the second, okay above the third, and bad above the fourth. The numbers are 2, 1, 0, and -1 respectively.
Voting system:
  1. Eliminate all candidates who either don’t want to be president or are ineligible to do so.
  2. Eliminate all candidates who aren’t on a majority of the ballots in each state. If less than three candidates pass, lower the percentage individually for each state until at least three candidates do.
  3. Eliminate all candidates who aren’t on a majority of the ballots nationwide. If less than three candidates pass, lower the percentage until at least three candidates do.
  4. Find the mean rating of all candidates combined. Don’t round it. Add enough ratings of this value to candidates with less ratings than the candidate with the most ratings so that all candidates have the same amount of ratings.
  5. Eliminate all candidates with an overall negative rating in at least one state. If no one succeeds, raise it by one state until one person succeeds. If only one person succeeds, they become president.
  6. Find the median grade for each candidate. If the middle falls between two different grades, the average is used. If several candidates share the highest median grade, all other candidates are eliminated. Then, one copy of that grade is removed from each remaining candidate's list of grades and the new median is found until no more candidates can be excluded. If, at the end of this step, there is only one candidate left, they become president.
  7. The House of Representatives votes (for, against, abstain) on whether to directly elect among the remaining candidates. At least 60% of the House of Representatives must vote for it. If they chose not to, they must select a date between 2 and 8 days after their decision to hold a Schulze method popular vote. If they chose to, the House of Representatives votes for the President and the Senate votes for Vice President
I've changed some things and am currently changing it again.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Voting system and reform thread

Post by mèþru »

I assume most (not all) people won't vote tactically if they can still get an acceptable candidate and win.
Salmoneus wrote:With respect, that's no clearer than before. What does "not accepting" a President mean?
Acceptable means that they are definitely not bad, but it doesn't necessarily mean that they are good. In the recent election, an unusually large amount of people found one or more candidates to be unacceptable.
Salmoneus wrote:But on the other hand, do you really want a system that prevents the election of Lincoln?
Yes. I'm not pro-slavery, but I would be against any candidate that is so unacceptable to a part of the country that they secede violently (I'm all for peaceful secession if the majority of people in the seceding territory are in favour).
Salmoneus wrote:I know there's unhappiness with the US system, but I don't think "our system is terrible, because it elected Lincoln!" is really the first objection most people have...
Nor is it mine. I didn't really even think of it until you mentioned it.
Salmoneus wrote:More broadly, "the one most people are OK with" suggests that what you really want is President Transformers: a candidate that pleases nobody and does nothing well. I don't think it's particularly obvious that "I recognise his name but have no clear opinion about him, so I guess I don't hate him!" is the qualification people really want in a President!
The qualification I want is "I recognise their name and policies, and whether I like them or not, they are acceptable." Also, I imagine that the winner of the first election under this system will either be a Green or Libertarian, but eventually it will shift to a system of mainly big party winners with occasional smaller party winners. It will also feature political parties unofficially fielding more than one final candidate while pretending to be united. The more candidates your party has, the more likely one will win. I view all of this stuff except for Green or Libertarian winning the first election as desirable.
Salmoneus wrote:What is gained by this, other than disenfranchising poorer (and disproportionately non-White) voters, who are the ones who a) are more likely to fail to correctly remember and spell their candidate's name, and b) are more likely to be put off voting by the introduction of such a punitively difficult voting method? Except, I suppose, that it would also benefit the two major parties (because you're putting more burden on name-recognition - you wouldn't be able to, say, turn up and vote for the Constitutional Party, you'd have to have learnt who their candidate was first), and would also benefit advertising executives (because drilling the names (and spellings) in would be more important that debating policy).
Good point. I'm keeping it though unless if you have a better alternative that addresses my goals.
Salmoneus wrote:(well, it doesn't need to be on the ballot - in your write-in system people just wouldn't write them in)
I already made a system before starting the thread using a method of exclusion not on the ballot. Look at the post above.
Salmoneous wrote:Oh, and that example also illustrates how MJ is incompatible with the "winner most people are OK with" criterion. Because in that example, only 3 out of 14 voters even gave Trump a higher-than-50% approval, and he still won. Whereas 7 out of 14 gave clinton at least 50% approval, and she lost
The chance of something like this happening with millions of votes are unlikely, but I see your point. I'll add some additional rules to address this. I'm going to use the fraction system I mentioned earlier instead of majority judgement.
Salmoneus wrote:Again, why is that a bad thing? That's just a fact about the voting process, not an assessment of its effects.
I think people should be free to vote for whomever they want. Ballot access laws give unfair advantages by allowing the top parties to restrict minor and independent candidates. People psychologically seek out names which are already written. When there are none, they are freed. Cumulative vote forces people to think about tactical voting to block out others' tactical voting. With majority judgement, if most people don't vote strategically, those who do will have little impact on the vote (in a vote on the scale of the presidential election, not in a 20 voters election).
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
Chuma
Avisaru
Avisaru
Posts: 387
Joined: Sat Oct 28, 2006 9:01 pm
Location: Hyperborea

Re: Voting system and reform thread

Post by Chuma »

mèþru wrote:Not much, but I'm hoping to learn.
Here are a couple of points from basic voting theory.

Voting theory mostly deals with single-winner elections, even though many political systems don't work like that. Since you seem to mostly be talking about the US presidential election, let's focus on that.

It is common in voting theory to talk about different principles. One obvious principle is the majority principle, which says "if one candidate gets a majority of the votes, that candidate wins".

Among the various principles, these three seem like a good combination:
• The democratic principle; all votes (or voters) are equal.
• The deterministic principle; the outcome is completely determined by the votes, there is no randomness or other effects.
• The honesty principle; you never benefit from voting for a candidate you like less - voting for A before B will never result in A losing. In other words, there are no voting tactics.

Interestingly, and unfortunately, it turns out that these three principles are impossible to combine. You have to skip at least one of them. Skipping the first, for example by letting one person choose, is certainly convenient in its own way, but not quite what we want here. Skipping the second can be surprisingly useful. The obvious example is the Random Ballot method, which is pretty much what it says - pick one ballot at random, and only count that one. Works fine for some situations, but maybe not for national politics. Virtually any (democratic) political system will thus compromise on the last.

There are various ways of organising your ballot. The most primitive is to pick only one candidate. Another method is to divide your vote; say, everyone is given 10 points, which they can divide between any number of candidates. Alternatively, you can say that you have give one point each to 10 candidates.

A problem with those two methods is that they tend to invite a lot of voter tactics. Suppose A and B are the two main candidates; most people agree that one of those should win. You prefer A. The honest thing might be to give the most points to A, and a few to B (or in the second case, include both A and B in your 10 votes). But since you're fairly sure either A or B will win, you're actually better of not giving B anything, thus promoting A over B as much as possible. So in effect, weighted votes won't make a difference, since any rational voter will give all their points to one candidate (and in the second case, "waste" the other 9 votes on bottom candidates).

A different approach is to just rank the candidates, without specifying values. One method that does that is Instant Runoff Voting, or IRV. It goes like this: After the election, put all ballots in piles based on their first votes. Pick the candidate with the lowest number of votes; cross out that candidate's name and move those ballots to the appropriate pile. Repeat with whichever candidate has the lowest number of votes now. Continue until a candidate has the majority of the votes. This method does lead to a little bit of voter tactics, but it's not as bad as some other methods.

In a ranking system, we (the system designers) need to ask ourselves what we want in a winner. Suppose for example the results look like this:
45% vote for A first, B second
15% vote for B first, C second
40% vote for C first, B second
It's a fairly realistic result, although a little simplified - A and C might be the two big parties, left/right, conservative/liberal, or whatever. B is a smaller compromise candidates. So, given this result, who should win?

In a plurality system, like the US one, A would be the winner. This is good in situations where you want a winner to have strong support, no matter what the opposition thinks. For example: A yoghurt company is releasing a new flavour of yoghurt, and letting their customers vote to pick a flavour. They will prefer this method, because they want strong support - those who vote against it won't buy the yoghurt anyway. In a political context, this kind of system encourages the formation of two parties, thus keeping away extremists; this has traditionally been seen as a benefit of the US system. On the other hand, many argue that it too strongly shuts out diversity, and it also leads to a strong divide between the two parties.

In an IRV system, B would win. This is useful in situations where you specifically need the support of the majority. It's effectively what happens in a proportional system, when a publicly elected parliament chooses a prime minister; this way, the prime minister will have the support of the parliament majority. A system like this allows plenty of diversity; many say that a proportional system can lead to too much diversity, although it's less of a problem in a single-winner system like IRV.

For the third option, let's take a look at another principle: the Condorcet principle. It says:
- If at least half the voters put A higher than B, then B is not a Condorcet winner.
- If there is a Condorcet winner, then that should be the winner of the election.

We're assuming that each voter lists all the candidates, but it's not that hard to generalise by saying that any candidate not mentioned goes at the bottom. It's easy to see that there can't be more than one Condorcet winner. In the example above, B would be the winner. Generally, a Condorcet method picks a compromise - someone who is at least accepted by the majority, as opposed to strongly liked by a few. That seems to be more or less what you want, right?

This way, there will never be a majority of the people who can come after the election and say "but we would have preferred candidate X over the one that won!", which seems to be more rule than exception in recent US elections. Some see this as an extension of the majority principle, and thus a vital extension of democracy.

In this kind of system, we might get a candidate which doesn't appeal that strongly to anyone. But it does mean that we can have all the diversity we want, and still get the best compromise as the winner. We avoid the division that seems to plague the US, as well as flip-flopping, that is, policies changing back and forth between presidential terms. What's more, we can fulfil all three of the principles listed above - there is no need for voting tactics, so everyone can vote exactly as they feel.

The problem is that this only works if there is a Condorcet winner. In some cases, no candidate is a Condorcet winner. For that case, there are various different rules, generally trying to find who was the closest to being a Condorcet winner, by some definition. When that happens, the system isn't free from voter tactics. But in many real cases, there is one main political axis, so people on one end will choose someone closer to the middle higher than those on the other end, and we get a Condorcet winner in the middle.

If you want a presidential system, while avoiding widely hated candidates and the two-party division, this might be a good idea.

A minor detail: If the voter is forced to list all candidates, many voters misunderstand the system, and put the main opposition candidate at the bottom of the list. That way, you might end up unwittingly electing a relatively unknown candidate. I would suggest a system where you choose how many candidates you want to list, and the rest are counted as going on the bottom.

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Voting system and reform thread

Post by mèþru »

I knew most of the stuff you talked about. I'm only partial going with the 1st principle (I've listed my criteria in a previous post, criterion #4 is incompatible with principle #1, so I'm trying to compromise between them), but I agree with the 2nd. I'm not sure about how I can even do the 3rd. As I've already stated, the Condorcet criteria is incompatible with the goals of whatever system I choose.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
Torco
Smeric
Smeric
Posts: 2372
Joined: Thu Aug 30, 2007 10:45 pm
Location: Santiago de Chile

Re: Voting system and reform thread

Post by Torco »

3. This is grotesquely undemocratic. You would be asking people to meaningfully cast ballots on a paper a hundred pages thick -
this can be solved trivially through having a write-in only ballot: fair enough, there's a problem to be solved regarding what you do when juan soto gets 40% of the vote, 30% goes to pedro kuzcinsky and 25% to pedro kuchinski, but this can be solved in a variety of ways.

elections where no one wins are not a bad thing, especially if you think no winner is a preferable outcome to the wrong winner.

the voting for 'don't know enough' is, indeed, a terrible idea.

now, as for OP's criteria:

1 is trivial to fulfill
2 is, as has been pointed out, vague: but it can be operationalized into something useful by saying "the winning candidate must be the person whom the greatest number of voters find acceptable"
3 is interesting, i find the notion of a write-in only ballot quite cool, but i suspect its impossible to compatibilize 3 with 2 within a single round of voting, since in order to procure information about which candidate a voter finds acceptable you have to ask them directly, present the name and have them mark it as acceptable or unacceptable [or, i suppose, have them cast a number of yeses, a number of noes, and whoever gets neither a yes nor a no from the voter is assumed to be meh which is to say acceptable]
4 is kind of an impossible ask: i get that the point is consensus rule as opposed to majority rule [which to me is not even clear that its a good idea so this degree, I mean, some protection for minorities are okay, but it would be trivial for a political bloc to veto a candidate by winning the "we vehemently don't want trump" vote in, say, montana or something]. considering this one's the least important, let's summarily dismiss it.

so the question is how do we compatibilize write-in ballots with meh criterion. [we're engaging in a bit of a conworlding effort here, since it would take aliens to actually go for a voting system built upon these principles, but what the hell, let's do it anyway]. the solution seems apparent: two rounds of vote. it's not so weird, we do it here! and, may i add, you guys do it too, with primaries and whatever. voting rounds give the system more information, and different kinds of information, so they're good.

the nope voting system

so in this system, what we do is we first have a write-in referendum. everyone gets to write let's say five names: we could have it that if you only write one name, your vote counts as five points for that name but since what we want is some political pluralism <why else would we want a write-in system anyway> let's stipulate that if you write one name your guy gets three points, if you write two names each gets two point and a half, three names get two point each, four names get one and a half points and five names gets one point each.

okay, so after everyone sends their write-in vote, you tab the results. bob nobody gets one point, kim kardashian gets a thousand points, and dylan clump gets a million points or whatever. you take the ten people with the most mention-points who want to be elected for the office, and they go to the second round.

in the second round, what you do is you only cast nope votes: a nope vote for X means you've said that you vehemently do not wish for X to be elected: the ballot has ten names and you can only nope for up to three people. or four, whatever. After this happens, the nopes are tabulated and whoever gets the least nopes wins.

maybe we can sneak in criteria 4 after all, though! let's say that if in any state or territory a majority of people [or perhaps 66% of people] nope for a certain candidate, then the rules say that that territory, as a whole, has noped said candidate. he is instantly disqualified. so the algorythm goes something like this:

round 1: write-in ballots, candidates get mention points for being mentioned according to the rule, and the 10 people who don't decline and have the most mention points go to round two
round 2a: have people cast negative votes, or nopes. and disqualify any candidates that get 66% or more nopes in a given state or territory
round 2b: if there's more than one non-disqualified candidate, compare the amount of nopes each got: the non-disqualified candidate with the least nopes wins.


** the reason why i think this would take aliens is that our moral intiutions about voting tell us, I think, that what we want is for the voting process to reflect the preferences of the voters, instead of the dispreferences of the voters: I wouldn't say I'd like to live under this system, but it fulfills the criteria of OP so... yeah,

User avatar
KathTheDragon
Smeric
Smeric
Posts: 2139
Joined: Thu Apr 25, 2013 4:48 am
Location: Brittania

Re: Voting system and reform thread

Post by KathTheDragon »

Torco wrote:whoever gets the least nopes wins.
It's extremely unlikely, but what happens in the event of a tie?

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Voting system and reform thread

Post by mèþru »

Your system is a heavily simplified version of what I created, essentially. It's specifically for the US presidential elections. No other country intended and no other public positions than president and vice president (voted for separately, so it is possible that they are from two different parties). It wouldn't work for a small country like Israel.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
alynnidalar
Avisaru
Avisaru
Posts: 491
Joined: Fri Aug 15, 2014 9:35 pm
Location: Michigan, USA

Re: Voting system and reform thread

Post by alynnidalar »

Having a president and vice president from different parties does not sound like a great idea. There's a reason the US stopped voting separately for those two positions.
I generally forget to say, so if it's relevant and I don't mention it--I'm from Southern Michigan and speak Inland North American English. Yes, I have the Northern Cities Vowel Shift; no, I don't have the cot-caught merger; and it is called pop.

zompist
Boardlord
Boardlord
Posts: 3368
Joined: Thu Sep 12, 2002 8:26 pm
Location: In the den
Contact:

Re: Voting system and reform thread

Post by zompist »

I don't want to detail the election system wonkery— it's fun topic with no clear solution, so discussion can continue forever— but it does seem to me that you have a very unusual analysis of what went wrong in 2016. It sounds like you think the problem was that Trump was "unacceptable" to most voters, or that Hillary was, or both of them.

Now, there are systems that try to address acceptability— mostly regional acceptability. E.g. Nigeria requires presidents to win all the major regions of the country. (Or did, I don't know if they still have this.) The intent is to avoid an ethnic candidate that everyone else hates, but the effect is also to reduce the candidate pool: only a few possible candidates will have sufficient nationwide recognition to run.

Personally, I'd find it undemocratic and outrageous if the election had been handed to Jill Stein or Gary Johnson on the grounds that neither major-party candidate was "acceptable" enough. The response to the polarization in the US is not to ignore the wishes of 94% of the voters.

I also would oppose any system which threw out the major candidates and replaced them with also-rans selected by some more arcane process, or by a smaller electorate (e.g. Congress). I would not be happier if the system threw out Trump for some reason and then gave us Ted Cruz, Mike Pence, or Marco Rubio. I think Bernie Sanders is a great senator but would be an embattled and unsuccessful president.

One particular point from your proposal illustrates the problem:
mèþru wrote:[*]Eliminate all candidates with an overall negative rating in at least one state. If no one succeeds, raise it by one state until one person succeeds. If only one person succeeds, they become president.
This is absurd; no one except George Washington has ever won all states. If this was your criterion, no one can ever be elected president. The only people who are "acceptable" to all voters are, basically, the people who aren't running. The very process of running turns on the polarization process and eliminates the general favorability. Or if you could magically find some person no one hates— Will Smith? Betty White?— the moment they entered office they would become A Politician and their acceptability would plummet.

(Consider: what does your Magic Outsider do about Obamacare? Are they pro or anti-abortion? Do they take in more or fewer or no Syrian refugees? What's their stance on climate change? Whatever they do, including nothing, will outrage some faction.)

Polarization is a tough situation, and it's easy to be nostalgic for (say) 1976, when the parties were each motley ideological coalitions and successful candidates were centrists. But that was the exception in US history, not the norm.

The general US approach to a party going off the rails is some combination of:
* let them lose elections till they decide to be less extremist
* let then win and mess up so badly that they delegitimize themselves for some period
* put lots of veto points in the system so no one party can do too much damage
* try to have some norms that allow bipartisan action, at least when no one's looking

It's pretty messy, but it's hard to come up with something better.

With our system, policy can careen back and forth between ideologies. And it probably should! A party in sole control of the government becomes corrupt, out of touch, or uninspiring. And a party in perpetual opposition becomes incompetent and unrealistic. Sometimes the only way to get rid of a bad idea is to try it out so everyone can see it fail.

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Voting system and reform thread

Post by mèþru »

I am trying to at least address the criteria, not fully solve it. Look at my draft in an earlier post.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
Torco
Smeric
Smeric
Posts: 2372
Joined: Thu Aug 30, 2007 10:45 pm
Location: Santiago de Chile

Re: Voting system and reform thread

Post by Torco »

KathTheDragon wrote:
Torco wrote:whoever gets the least nopes wins.
It's extremely unlikely, but what happens in the event of a tie?
same thing as happens in any other voting system where enough people vote that ties don't happen... i guess parliament decides or something?
alynnidalar wrote:Having a president and vice president from different parties does not sound like a great idea. There's a reason the US stopped voting separately for those two positions.
ya, i would imagine it makes magnicidio a much more attractive notion.

User avatar
mèþru
Smeric
Smeric
Posts: 1984
Joined: Thu Oct 29, 2015 6:44 am
Location: suburbs of Mrin
Contact:

Re: Voting system and reform thread

Post by mèþru »

I'm pretty sure murdering is grounds for impeachment, and any death of a president provokes conspiracy theories about vice presidents anyway, even if they are of the same faction of the same party.
ìtsanso, God In The Mountain, may our names inspire the deepest feelings of fear in urkos and all his ilk, for we have saved another man from his lies! I welcome back to the feast hall kal, who will never gamble again! May the eleven gods bless him!
kårroť

User avatar
Torco
Smeric
Smeric
Posts: 2372
Joined: Thu Aug 30, 2007 10:45 pm
Location: Santiago de Chile

Re: Voting system and reform thread

Post by Torco »

ya but if the vice president was a democrat a lot more people would be awfully happy if the donaldmeister ate some plutonium or whatever

Post Reply