My vote on voting systems

https://blog.danielh.cc/blog/voting

by max__dev • 2 years ago

70 51 2 years ago

blog.danielh.cc

jszymborski2• 2 years ago

No voting method meets every criterion but two of my favourites are

- Schulze Method [0]

- Ranked Pairs [1]

The Schulze method allows for a simple ranked choice ballot, and satisfies more criteria than other RCV methods. Downside is that it is hecking complicated so it can feel like an opaque process. With distrustful electors it's a no go imo.

The Ranked Pairs method satisfies a similar number of criteria as Schulze, and meets a weaker version of later-no-harm. It's also a very intuitive method. The main downside is that the ballots become impossibly long, scaling quadratically with the number of candidates.

[0] https://en.m.wikipedia.org/wiki/Schulze_method

[1] https://en.wikipedia.org/wiki/Ranked_pairs

nearting• 2 years ago

Ranked pairs does not have a quadratic ballot length, since all of the preference information can be computed from a single ranking of the candidates.

jszymborski2• 2 years ago

How did I never realize this. Thank you.

EDIT: I was also swapped "later-no-harm" with "independence of irrelevant alternatives". I should stop writing comments before my morning coffee.

intalentive• 2 years ago

Unanimous.ai has an interesting approach: instead of static snapshots, make voting a dynamic, real-time process where preferences are elicited through action.

yellowapple2• 2 years ago

It's weird to me how little score/approval voting has caught on (at least here in the US) amid the push to replace FPTP. It's a lot simpler to explain than RCV, and has a lot fewer downsides overall than both RCV and FPTP.

ztetranz• 2 years ago

Approval voting might be good but just at a human level, I think it can leave the voter feeling unsatisfied. If there is one candidate that I really like and another that I could only accept reluctantly as an alternative to someone worse then it feels bad that I need to "approve" of those two equally.

ClayShentrup5• 2 years ago

what matters is satisfaction with the election outcome, not the five minutes you spent filling out a ballot.

https://rpubs.com/Jameson-Quinn/vse6

ekimekim2• 2 years ago

Being vulnerable to strategic voting is a huge downside that outweighs other considerations.

As the article mentions, in the real world score voting would just be approval voting where you put a max score on some choices and 0 on others.

And in approval voting you need to think about how others will vote and pick your cutoff point based on who you think has a chance - do you vote "yes" for the center-right party to avoid the hard right party getting in? Or do you vote "no" to help the center-left party beat the center-right party? (swap those directions to personal preference)

RCV isn't perfect, but in all but the smallest elections there's really no practical strategic voting considerations. You just state your true preference order.

Of course, I'll take any of them over FPTP.

yellowapple2• 2 years ago

Being vulnerable to strategic voting is a huge downside that outweighs other considerations.

I disagree that strategic voting as a downside outweighs the downsides of RCV or FPTP - especially when FPTP itself is susceptible to strategic voting, too. None of the three satisfy the condorcet winner criterion (that is: none of them guarantee the winner would beat every other candidate head-to-head), but it seems less likely / more contrived for score/approval voting to fail it.

And in approval voting you need to think about how others will vote and pick your cutoff point based on who you think has a chance - do you vote "yes" for the center-right party to avoid the hard right party getting in? Or do you vote "no" to help the center-left party beat the center-right party? (swap those directions to personal preference)

That's why I'd personally go with a simple three-level score vote: "yeah", "meh", or "nah". If people really want to shoot themselves in the foot by ignoring the "meh" option, then so be it, but at least the option is there for people to vote "meh" for candidates that are merely acceptable/tolerable (and reserve "yeah" for ideal candidates and "nah" for unacceptable/intolerable candidates).

ClayShentrup5• 2 years ago

As the article mentions, in the real world score voting would just be approval voting where you put a max score on some choices and 0 on others.

utterly false.

https://www.rangevoting.org/HonStrat https://www.rangevoting.org/Honesty https://www.rangevoting.org/RVstrat6

And in approval voting you need to think about how others will vote and pick your cutoff point based on who you think has a chance

1. this is true of _every_ deterministic voting method. it's mathematically proven.

2. even approval voting is extremely accurate. see voter satisfaction efficiency calculations from harvard stats phd jameson quinn, who served with me on the board of the center for election science. https://electionscience.github.io/vse-sim/

RCV isn't perfect, but in all but the smallest elections there's really no practical strategic voting considerations. You just state your true preference order.

ludicrous. IRV is extremely vulnerable to strategic voting. my aunt who voted biden even tho she preferred warren (to stop trump) would do the same thing with ranking: rank biden 1st to prevent warren from losing to trump.

something similar happened in the first alaska house special election. palin was a spoiler. she caused the democrat mary peltola to win, even tho fellow republican nick begich was preferred to peltola (and to palin) by a large majority. in other words, palins supporters hurt themselves by honestly ranking palin 1st. they would have been strategically better off burying palin and ranking begich 1st.

https://rcvchangedalaska.com/

and the VSE metrics show that even 100% honest IRV performs worse than approval voting with all voters being strategic.

IRV is simply one of the worst voting methods in existence. not to mention being radically overcomplicated and opaque.

i'm sorry that you posted this before having even an inkling about how any of this works.

bongodongobob• 2 years ago

I think fundamentally it's because you have to convince the elected to pass laws that potentially undermine the way they got there, essentially asking them to kick out the ladder from underneath them. Couple that with the average person thinking that FPTP is the most fair because explaining why it isn't is counterintuitive, nothing is ever going to change.

int_19h2• 2 years ago

Explaining why FPTP is unfair in the abstract is not difficult at all.

The problem is that US specifically has a political climate which can be summed up as, "the ends justify the means". While both parties still ritually extoll the importance of democracy and fairness, as soon as it's down to one particular wedge issue, winning is more important than preserving the integrity of the system.

To a large extent, this is because there's basically zero trust between differing factions - it is assumed that your political opponents will squeeze out every advantage they can out of the system when they are in power, no matter how unfair, and so when you are in power, if you don't do the same, you basically hand them victory on policy long-term. Arguably was originally Republicans who triggered this cycle during the Obama presidency, but it's hardly important now - the point is that once trust is broken, it quickly becomes a positive feedback loop.

So, getting back to electoral system, your average voter looks at those proposals to replace FPTP with RCV or whatever, and the first thing they ask is not, "is this more fair?", but rather, "how will this affect the balance of power?". And in any given constituency, the answer is that it will take power away from the current majority (or plurality) and hand it to their opposition. This is both obvious and easy to explain, so that's the usual agitprop angle, and it sells very well.

slyall• 2 years ago

The problem is the US is still running off best practice from 200 years ago and is famously hostile to change or copying another from other countries.

Any change is also going to disadvantage the two current parties so they will be hostile to it. Which with the hard-to-amend US constitution makes things very hard to change.

Some form of proportional representation or maybe STV in multi-member districts is actually what you'd do. Single member districts with fancy voting systems just gets you a token 3rd party representation.

theluketaylor3• 2 years ago

The mathematical challenges of voting systems is one of the reasons I've come to support multi-member districts. When only a single person can represent a jurisdiction you're always going to have people disappointed in the choice and we have have plenty of proofs that perfect voting isn't possible, so we need to aim for best reasonable compromise.

Voting is even more about identifying who didn't win than who did win. In a multi-member district a much smaller slice of the electorate 'loses', making it much easier to build broad consensus around results.

Some of the sharper edges of many ranked choice or scoring voting systems are blunted by multi-member districts. They are not winner-take-all, so small changes in expressed preference that would have changed the winner in a single member district just re-orders the winners. Specific candidates could still find themselves losing when another system would have elected them, but the electorate as a whole still ends up with representation they can tolerate.

jfengel• 2 years ago

Multi member districts does circumvent one of the assumptions of Duverger's Law, which predicts a two-party system.

But you still need to work around collusions between candidates. Candidates will be pushed to run on a slate of N candidates for N seats, and promise to support each other. Voters who want any of their positions will avoid somebody who would interfere with their own top priority. The result acts the same as a two party system, where all N candidates support the same party.

You can get some benefit by combining it with List Voting, where you go ahead and reify the party system. The party gets to send whoever it wants, proportional to the votes.That makes it harder to run multiple candidates that take all of the spots. (Not impossible: they can break into multiple virtual parties. But that's risky and coordinating votes is difficult.)

Even at that I'm not sure it really helps. Lawmaking is still fundamentally a "single member" operation: a bill either passes or it doesn't. You still get pushed to be in either the majority or minority coalition.

johngladtj• 2 years ago

Even at that I'm not sure it really helps. Lawmaking is still fundamentally a "single member" operation: a bill either passes or it doesn't. You still get pushed to be in either the majority or minority coalition.

The solution for that is to not make it a simple binary pass/fail, but to have how it passes and fails affect the outcome.

A simple and consistent way of doing this is by making it so that all laws have a built in expiration date, and how big the majority for passing the bill affects how long the law is in effect from.

A 50% + 1 margin of victory should result in the law becoming void the moment the parliament composition changes.

Perhaps a 60% margin might allow it to go for 10 years, a 70% margin for 15 years, and so on.

Constitutional changes should essentially require unanimity ( and a referendum on top of it) since they are the only permanent laws.

Some laws would be more or less "permanent" since every 30 years or so you'd like get a 99%+ majority for a "theft is illegal" type low, but other more controversial would likely be on the ballots every few years.

theluketaylor3• 2 years ago

I'm really against mixed-member proportional (and pretty much all form of proportional representation) since you end up with party hacks unaccountable to voters. When parties control the order of the lists there is also really strong incentive by representatives to tow the party line to avoid relegation.

tanewishly• 2 years ago

I must be not quite following your explanation... In most of the EU, there are multi-party systems. From what I can tell, that is because of multi-member districts. Eg. The Netherlands has only one district for parliamentary elections (150 seats). There are more than 2 parties in Parliament there.

dragonwriter3• 2 years ago

But you still need to work around collusions between candidates

No, you don’t, if it’s a multimember proportional system, like STV. There are some multimember systems your concern applies to, like the “vote for N, top N win” at-large system used for some US city councils, school district boards, and similar bodies: those are simply means to have a multimember body while avoiding either diffuse or geographically-concentrated minorities from being represented. (And the prospect of that particular kind of multimember arrangement being used to neutralize the effect of black voting rights, specifically, is why multimember districts are prohibited by federal law in Congressional apportionment, currently.)

dang• 2 years ago

We replaced the linkbait "you" in the title with a nicely first-person phrase from the end of the article. If there's a better* title, we can change it again.

* better = more accurate and neutral, preferably using representative language from the article

https://news.ycombinator.com/newsguidelines.html

mdp2021• 2 years ago

# "Voting system[s hardly] work"

hooverd• 2 years ago

I think STAR voting wins out in quality as a voting system and understandability to people who don't know what a voting system is.

boxed• 2 years ago

I think there's also a different and deeper question: what does the voting system do to the society it is running on. This is a very different kind of discussion than "does it map preferences well", but arguably more important.

tromp• 2 years ago

However, they might decide to vote strategically to amplify their vote — giving Desert a score of 10 while giving everything else a score of 1 would effectively make their vote stronger.

It makes more sense for the voter to give their favorite candidate a 10, their least favorite a 0, and grade every other candidate on the spectrum between their two extreme picks. So that still seems to be as informative as you could hope for.

throwaway_12242• 2 years ago

I think the state of democracy in the "factory bamboozling" era is such that we need a system that reflects direct voter preferences less, not more.

The trick is coming up with a system that diverges towards the interest of the masses, rather than the elites or the corrupt or the foreign. Basically, the system must get the voters' preferences wrong, but get them wrong right.

This is really akin to the nature of representative democracy. Elected leaders manage the voters' interest rather than always doing what is polling better on that day. It's obviously a tricky task, but it's the hallmark of the best-functioning democracies.

Right now we're about as far from the goal as we've ever been. Our systems have been hacked and we don't have an effective counter.

int_19h2• 2 years ago

This is a very long-winded way to say that democracy is a bad idea if it's not just for show.

throwaway_12242• 2 years ago

"Just for show" would be foregone conclusion elections like Russia's or ignored elections like Venezuela's. Voters must still be central to the system and get the final say for it to be democratic.

The system just needs to be wiser and more resilient. For instance, the Electoral College was originally intended to be more than a bunch of sworn minions. In Europe, many parliamentary democracies have strong party infrastructures that prevent their takeover by "memetic" forces. There are endless possible improvements.

cogman10• 2 years ago

What we need, IMO, is direct legislation. Many states have this and I think federally having the ability to put up and vote on a bill would do a lot to correct many of the issues we currently face.

Barring that, RCV would do wonders for the current political ecosystem.

valiant55• 2 years ago

We should do away with voting and select our representatives randomly, similar to jury duty. We'd need a society that places value on civil service and is better educated and willing to listen to experts to hash out legislation.

Elections are only good for finding the most popular candidate instead of the most qualified for office. Keep terms medium length (e.g. four years) and you'll have people focus on the greater good instead of personal benefit because at the end of their term they have to live in the society they helped build.

exmadscientist2• 2 years ago

The direct legislation initiatives we just had here in WA were a total crapshow and I hope we never see their like again. They got tarred with the same brush because they came from the same sponsor (which is somewhat reasonable), even though they were not all equally popular or even sane. Then some were written in the negative and some were written in the positive, so it was very difficult to figure out what you were actually voting for. And one of the special interest groups figured out their favorite position was "all no", which is really easy to advertise, even if many voters in that block would have actually preferred yes on one of the initiatives.

And of course the Legislature had gone in and muddled with them even before they made the ballot (which, to be fair, is explicitly the Legislature's job in the WA initiative system, but it didn't help the confusion).

It was not a pretty election. I am convinced that direct democracy is not the answer, and that was with just four relatively inconsequential initiatives.

ajuc• 2 years ago

Proportional + threshold is simpler and better in practice than most of these. Yes it's not a direct equivalent (you have to change the problem so it has a good solution).

Also the problem in some countries isn't the voting system alone, it's the unlimited donations by billionaires, unchecked gerrymandering and general institutional decay.

System doesn't matter when people get away with ignoring it.

smitty1e3• 2 years ago

12-year election officer here. This idea is great for small groups of informed voters.

However, I submit that there are (at least) three other crucial discussions to have:

- Does my status as an informed voter bias me toward a system for which I have done all the homework?

- Does my script that works fine in my dev system with relatively clean test data scale to an enterprise deployment and real-world loading?

- How significant in improvement is needful to pay for the switching costs? Twice as good? An order of magnitude?

Because I submit that, for all the virtues we informed types can see in RCV, it is prone to malicious attacks, e.g. flooding the ballot with a bunch of relatively unknown candidates.

Furthermore, if there is any sort of contention involved, the complexities are going to lead to (likely unfounded) claims of tinkering.

I'm a huge fan of the ditt-kuh-pow! (DTTCPOW, the Dumbest Thing That Could Possibly Work):

- Make Election Day a federal holiday

- Paper ballots

- In-person voting except in strictly controlled, marginal cases (e.g. deployed military).

Everyone trading security for convenience is misguided, in my opinion. You want to mark a ballot and cast it in secret so that, whatever you have to say to appease the shrieking neighbor, you and your ballot are as physically secure as possible.

Staff the process so that pollbooks are <5k people. We get it done with ~20 people in my county for 13hrs on election day. ~240 precincts in the county. Not cheap, and yes, exhausting to staff. So, what? Staff, equip and support them and elections, can be a positive experience for all involved.

exmadscientist2• 2 years ago

FWIW, after living in an all-mail voting state (WA), I never want to go back to voting in person. It is amazing having a couple weeks to mark a ballot at my leisure, do as much research as I want, and then just drop it in a box (or mail it, I guess, but the drop boxes have definitely been more reliable for me).

But it's true I don't have a shrieking neighbor or brutal spouse to worry about.

seniorivn• 2 years ago

serious question, how do you know your vote is counted and is not intercepted/counterfeited before that?

Spivak• 2 years ago

Even when you use the digital voting machines it spits out a paper ballot that you deposit into the ballot box so the country already agrees with you.

It's not security vs convenience it's security vs turnout. Our turnout is embarrassingly low, especially on non-presidential election years, we should be expending voting access to methods that have even less friction. If we can conduct banking via an app or website we can vote via an app or website too. We can start talking about reigning it in once we don't have a third of the country not voting.

You deal with the abusive spouse problem by letting an in-person vote override the mail-in ballot while everyone else gets an easy way to vote.

smitty1e3• 2 years ago

Turnout is also a problem, but one needs to ensure that knowledge of non-participation doesn't become proxy votes cast by people In The Know.

Rest assured: any workable angle you can think of an many you have not have already been employed to cheat elections.

This is not a partisan dig; rather, simple reality.

EdwardDiego• 2 years ago

NZ switched from FPP to to MMP [0] after a few elections where a) the government was formed by a party that achieved <40% of the popular vote and b) parties that received 15 - 20% of the popular vote achieved only a tiny modicum of representation in Parliament.

I'm not sure MMP is the ultimate ideal, but I much prefer it to FPP. It's a very rare government that isn't a coalition government, and while we get the occasional tail wagging the dog moment, we also get a more diverse set of parties in power, and also more diverse representation.

Often what voters want to know before voting is who a party is willing to go into coalition with, there's usually no big surprises (e.g., the righter-wing ACT party and the centre-right (by NZ standards, they're still left of the Democrats at least when measured by public healthcare etc.) National always dance together, Labour and the Greens typically get along via a confidence and supply agreement even if the Greens are not formally in coalition) but then you have New Zealand First, who are openly mercenary about their choices, and the original Maori Party that went into coalition with National which doomed them in the next election.

[0] https://en.wikipedia.org/wiki/Mixed-member_proportional_repr...

throw0101d• 2 years ago

I'm a huge fan of the ditt-kuh-pow! (DTTCPOW, the Dumbest Thing That Could Possibly Work):

This acronym needs to be more widely deployed.

KenSF• 2 years ago

My vote is 'any 1 election runoff system'[score, star, approval, ranked choice, ...]. Every 1ER is better than first past the post. It would be nice if vote nerds would shut the frak up about this one is better than that on. Open primaries; top 4 or 5 go through to the general; general uses any 1ER. Were I writing a law to put this in place, I would say we should revisit which 1ER after X number of years or Y voting cycles.

grumpymuppet• 2 years ago

I love the promise of ranked choice voting (and other schemes). Properly implemented it seems like it could be a magic bullet to balancing some of the political divide and drawing out candidates and ideas that would satisfy more than just the extremes...

But, oh boy: we have trouble with simple plurality voting. How would you explain the outcome to people if it was implemented broadly?

My vote on voting systems

Related Stories