Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: RankedVote – SurveyMonkey but focused on ranked-choice voting

rankedvote.co

Post reply on HN
Print view View original

261–270 of 326 posts

Re: Show HN: RankedVote – SurveyMonkey but focused on ranked-choice voting

#261

Post by bena »

autoexec wrote: post #238
ClayShentrup wrote: post #223

Earlier quoted context omitted.

This is simply wrong. Palin voters didn't get their second choice preferences counted. If they had instead tactically/ insincerely ranked Begich first, he would have won by a 5% majority against Peltola. They were punished by voting sincerely for Palin. Palin was a spoiler. People who voted for her threw away their vote, period.

Voting for a losing candidate does not mean throwing your vote away or not having your vote counted. People who picked Palin as their #1 choice won against Begich. Their vote was counted and it went to Palin exactly as they wanted. If you want to argue that there should have only been one republican on the ballot to prevent a 'spoiler' you'd have to take that up with the republican party

There it is, there's the thing they're really railing against. They wanted every Republican voter to vote for all of the Republicans, then all (or none) of the Democrats. Basically turning the ballot into RvD.

But they forgot one crucial thing: people. Even within the party, there are members and candidates others can't stand. And if you're going to give people choices, you have to be prepared for them to make choices you don't agree with.

Re: Show HN: RankedVote – SurveyMonkey but focused on ranked-choice voting

#262

Post by naniwaduni »

crazygringo wrote: post #53
naniwaduni wrote: post #45

Earlier quoted context omitted.

> And then in my example, the voter is tempted to approve just 1 in order to give them a better chance of winning. (Known as bullet voting.) This is fine. The probability with which a voter bullet votes vs votes three conveys the degree of preference for that candidates. This is not as informative as score voting on a wider range (because okay, degree of preference doesn't map perfectly linearly with probability of v…

Ranked choice absolutely conveys more information than approval. It's not as much as score voting, but it's mathematically false to say it's the same amount as approval. Among five candidates, approval would be 2^5-2=30 possibilities (almost 5 bits of information, assuming you can't approve/disapprove all because that's the same as not voting), while ranked choice would be 5!=120 possibilities (almost 7 bits of infor…

> Ranked choice absolutely conveys more information than approval. It's not as much as score voting, but it's mathematically false to say it's the same amount as approval. Among five candidates, approval would be 2^5-2=30 possibilities (almost 5 bits of information, assuming you can't approve/disapprove all because that's the same as not voting), while ranked choice would be 5!=120 possibilities (almost 7 bits of information, assuming ties aren't permitted). It's only with 3 candidates that they have equal information in terms of possibilities (2^3-2=6=3!).

This is a common intuition trap, but it's wrong. Consider the difference between score voting on a 16-point scale vs a million-point scale vs arbitrary-precision score voting. A ballot for the first conveys 4 bits per candidate; the second, about 20 bits per candidate; the last conveys infinite bits of information. You ... probably don't expect a substantial difference between election results with each option.

The trick is that the aggregation rule destroys information, which is kind of the point of aggregation. Different aggregation rules destroy different information. The difference between aggregated approval voting and aggregated arbitrary-precision score voting eats LSBs, sure, but the obvious probabilistic constructions for extracting approval votes from arbitrary precision score votes actually approximate the arbitrary-precision votes quite well, which is not what your intuition predicts, because it turns out LSBs are, well, less significant.

Also, the flat fact is that most single-victor elections in practice don't have more than three "top" candidates that you're practically interested in extracting information about.

> And individuals don't have intransitive preferences, only populations can. (Assuming mental competence.)

That's one hell of a bad assumption.

Re: Show HN: RankedVote – SurveyMonkey but focused on ranked-choice voting

#263

Post by buzzy_hacker »

supreme_loquat wrote: post #235
buzzy_hacker wrote: post #213

Earlier quoted context omitted.

> we are dumbfounded by those who prefer Ranked Choice Voting. Hello, I prefer RCV. Reasons: - it is a stepping stone to proportional representation (achieved by adopting both RCV and multi-member districts), which is much more important than which single-winner voting system is chosen - you need to be a fellow voting systems nerd to understand when you may want to vote tactically in RCV, so approximately zero people…

> it is a stepping stone to proportional representation (achieved by adopting both RCV and multi-member districts), which is much more important than which single-winner voting system is chosen There are proportional methods for STAR and approval as well, there's no reason they couldn't be the stepping stones from this perspective > you need to be a fellow voting systems nerd to understand when you may want to vote t…

> There are proportional methods for STAR and approval as well, there's no reason they couldn't be the stepping stones from this perspective

Proportional approval voting is extremely confusing (and expensive) to compute the winner, people won’t accept a system that they don’t understand how it works.

> Yet that doesn't try and stop people from duping voters into eg only ranking Palin, even though RCV is basically the only system where there is practically 0 reason to not rank additional candidates.

It’s the first election where people used it so they will learn about that. Also some people might just want to vote for Palin and are indifferent after that. While I would fill out the entire ballot, most people won’t bother, and that’s fine.

> if someone were to eg bullet vote in STAR to help their first choice win, is that really tactical voting, or just an honest vote/admission that they care so much about their first choice that all other opponents are equally bad?

No it does not necessarily mean that, it could plausibly mean that they do have a second preference, but are afraid of giving them points so that they don’t overtake their first preference. Think of a Bernie, Hillary, Trump election where some Bernie voters don’t want to give high points to Hilary, even though they much prefer her to Trump, for fear of her beating Bernie. RCV allows them to confidently vote Bernie > Hillary > Trump and not worry about this problem.

> I think the best use of approval is in primaries where the top two candidates proceed to the general.

I think it would be best to ditch primaries completely and have everyone run in a multi-winner general election with RCV. So few people vote in primaries, so the highly engaged extremists influence the outcome for everyone in the general. Also why do I need to vote twice for the same position, it’s annoying. It would be nice to only need to vote in one election, the general.

> With RCV the two dominant parties no longer have to worry about the spoiler effect, so there is even less incentive to reach across the aisle and receive influence from a third party than before. Whereas before you might want to take some token talking points from a third party and siphon some votes from them, with RCV you might be able to assume you'll get those voter's 2nd choices for free.

It seems like you are talking about two things at once:

1. a major party moderating their position to get votes from the other major party

2. a major party taking a (more extreme?) position to get votes from a third party

When you say “reach across the aisle”, to me that means #1 above, i.e. Democrats and Republicans moderating their positions to appeal to the median voter. I agree that would be good and RCV does encourage that!

RCV doesn’t encourage #2, and I think that’s also a good thing.

I would love a robust multi-party system, but for that we really need proportional representation. Trying to get it with single-winner elections, no matter the system, isn’t a good approach.

Re: Show HN: RankedVote – SurveyMonkey but focused on ranked-choice voting

#264

Post by naniwaduni »

dane-pgp wrote: post #122
naniwaduni wrote: post #45

Earlier quoted context omitted.

> And then in my example, the voter is tempted to approve just 1 in order to give them a better chance of winning. (Known as bullet voting.) This is fine. The probability with which a voter bullet votes vs votes three conveys the degree of preference for that candidates. This is not as informative as score voting on a wider range (because okay, degree of preference doesn't map perfectly linearly with probability of v…

> The probability with which a voter bullet votes vs votes three conveys the degree of preference for that candidates. I can see why that would be a useful assumption, but is there really empirical evidence to confirm that? A much bigger factor in determining what the voter does, I would guess, is whether they hear activists for their party telling them (perhaps disingenuously) that their candidate won't win if you "…

It's not a fantastic proxy, I agree, since the aggregation rule implicitly assumes linearity and it's almost certainly not. I think it's somewhat mitigated in practice by the kind of small intersection between situations where it's strategically safe for a party to push bullet voting and where it's useful to do so being pretty constrained to elections where the top candidates are mostly acceptable to the same voters, though? I find it pretty unintuitive to construct an approval vote election scenario with a rational spoiler effect, and that's probably less big a deal than the vote weighting issues that crop up with multi-point score votes, but that's definitely something you can quibble over.

The wider point imo is that compared to all forms of ranked choice voting, arbitrary-precision score voting is pretty obviously night-and-day superior (information-content-wise, at least), to which approval voting is a "close enough", at worst moderately inferior but at least theoretically equivalent, trade-off in favor of implementability.

Re: Show HN: RankedVote – SurveyMonkey but focused on ranked-choice voting

#265

Post by naniwaduni »

musingsole wrote: post #124
naniwaduni wrote: post #25

Earlier quoted context omitted.

No. Approval voting is actually a minimized score voting system, and although intuitively a two-point feels like this should convey less information, the way the aggregation rule works (straight addition—or specifically in the approval voting case, tallying), you're actually extracting significantly more information out of a set of approval votes than you would from a ranked choice vote. To put it another way: given…

Wow, thank you!!

It's a pretty interesting topic on which a common intuition (n bits is less information content than log(n!), right?) fails it turns out most of the work is being done in the aggregation algorithm, not in the data entry.

Which is emblematic of the biggest problem with RCV, which is that the aggregation algorithms are too opaque to realize at a glance that the results they give are ... not as good as doing a straightforward thing.

Re: Show HN: RankedVote – SurveyMonkey but focused on ranked-choice voting

#266

Post by crazygringo »

NickM wrote: post #158
crazygringo wrote: post #112

Earlier quoted context omitted.

> RCVs many glaring problems But all the voting methods have glaring problems, and every method is going to result in people complaining after their preferred candidate loses but would have won using another method. RCV isn't going to poison the well any more than other systems. STAR voting and approval voting, for example, open themselves up to the criticism that voting itself becomes a subjective process, which cou…

it's easy to understand and doesn't result in absurd outcomes I mean, I realize this is somewhat subjective, but "more people ranking a candidate first can cause them to lose" seems like a pretty absurd outcome to me. I'm not saying approval voting is perfect either, sure it has flaws, but at least it doesn't violate the monotonicity criterion (or a variety of other important criteria that RCV fails to satisfy).

> but "more people ranking a candidate first can cause them to lose" seems like a pretty absurd outcome to me.

It sure does seem absurd, and so thankfully there's no case where RCV does that! :)

There seems to be a lot of misinformation thrown around about this, where people seem to be conflating different issues.

I'm not sure what your source is, but you might be misunderstanding the Favorite Betrayal Criterion... which seems counterintuitive at first but isn't really. There's a lot of noise made about the fact that moving a non-preferred candidate higher than your preferred candidate can help your preferred candidate to win... but it's actually because you're ranking your preferred candidate's main opponent even lower in the process. And also it's basically impossible to do strategically because you'd need to know how everybody else voted first.

End of story, there's nothing absurd about cadidate A winning, but if some people downranked candidate A from 2nd to 3rd rank, then that candidate would lose.

Re: Show HN: RankedVote – SurveyMonkey but focused on ranked-choice voting

#267

Post by crazygringo »

fdsjgfklsfd wrote: post #253
crazygringo wrote: post #88

Earlier quoted context omitted.

I just want to make clear for others: this "ranking" is just bhauer's opinion. Since some people here might interpret it as being presented as objective fact or some consensus opinion. There are a lot of pros/cons to all of the richer voting systems, usually categorized as approval voting, ranked voting, and score voting, in terms of increasing information. But there is absolutely no consensus that approval or score…

There are many forms of "ranked" and there absolutely is consensus that the form that constantly gets pushed in the US is no good. It is marketed with lies and doesn't live up to any of its promises, as we just saw in Alaska. It didn't prevent the spoiler effect, it didn't make it safe for voters to rank candidates honestly, and it didn't choose the candidate preferred by the majority. Pretty much any other proposal…

If you're claiming there's consensus RCV with IRV is no good, you're going to have to show sources. Because everything I know tells me that's patently false. There's no such consensus in academia, there's no such consensus in the political world, there's no such consensus anywhere.

You say pretty much any other proposal is better, plenty of other people say the other proposals mostly look even worse.

I don't understand why you ascribe this to "hardheaded loyalty" instead of people just genuinely disagreeing with you, and that they have good reasons for it. Honestly, from the tone of your comment it seems like you're the one who's unwilling to seriously consider the alternatives.

Re: Show HN: RankedVote – SurveyMonkey but focused on ranked-choice voting

#268

Post by crazygringo »

ClayShentrup wrote: post #219
crazygringo wrote: post #88

Earlier quoted context omitted.

I just want to make clear for others: this "ranking" is just bhauer's opinion. Since some people here might interpret it as being presented as objective fact or some consensus opinion. There are a lot of pros/cons to all of the richer voting systems, usually categorized as approval voting, ranked voting, and score voting, in terms of increasing information. But there is absolutely no consensus that approval or score…

No it's not just his opinion, there are objective metrics like voter satisfaction efficiency that back it up. https://www.electionscience.org/library/approval-voting-vers...

No, because "voter satisfaction efficiency" is just as subjective of a criterion as the voting methods themselves are.

There's nothing "objective" about it. Just because it's a number you can calculate doesn't make it the right number. Some people think it's a good metric for evaluating voting systems, but many others don't, and there isn't even remotely any kind of consensus that it's a good or even useful measure.

Re: Show HN: RankedVote – SurveyMonkey but focused on ranked-choice voting

#269

Post by naniwaduni »

garaetjjte wrote: post #156
unholiness wrote: post #127

Earlier quoted context omitted.

Every voting system has some flaws (see: Arrow's impossibility theorem), and Instant Runoff Voting definitely saw its worst-case scenario rear its ugly head in Alaska: A Condorcet winner existed (Nick Begich), he lost, and voters whose first choice was Plain and second choice was Begich could theoretically, retrospectivally, all changed their first votes from Palin to Begich to have him win rather than Mary Peltola.…

>see: Arrow's impossibility theorem Arrow's theorem only applies to ranked systems. That said, cardinal systems are still susceptible to strategic voting (Gibbard's theorem)

Gibbard's theorem implies "strategy" in a sense (your optimal strategy depends on distribution of other people's votes) which is much less nefarious than many people's intuition on what "strategic voting" implies.

When people think about strategic voting coming from a FPTP context, the strategy they're thinking of is putting in a dishonest vote for a non-preferred candidate because that provides a better chance at matching their preferences than an honest vote. This is not what Gibbard's theorem is about. Strategic voting in cardinal systems, for example, tends to look basically like dynamic range compression and at an individual level generally only "tiebreaks" between candidates that were otherwise fairly evenly matched, while the FPTP strategy of bullet voting a non-preferred candidate is just straight up bad for your preferences.

Re: Show HN: RankedVote – SurveyMonkey but focused on ranked-choice voting

#270

Post by bena »

achr2 wrote: post #149
lucasmullens wrote: post #93

Earlier quoted context omitted.

> If it's a close three-way election, you might cause your favorite candidate to lose by ranking them first. Could you explain that? I can't think of a situation where putting the candidate I want to win as #1 would cause them to lose. Ranking other close candidates as #2/#3 could cause my #1 to lose, but I don't think voting for someone as #1 alone would possibly harm them.

Let's say you have three candidates A, B, C. A is left, B is center, C is right leaning. Polling shows that voters are leaning: - 15% BAC - 15% BCA - 25% ABC - 45% CBA If everyone voted as they desired, then the instant runoff would be between B and C, with B winning with 55% of votes. If a subset of C voters instead strategically vote ACB (6%), this would change the runoff to be between A & C, with C winning 60%.

First, a run off between A and C wouldn't have C with 60%. C would have 54%, the 6% of C voters would have to come from CBA. And if A is in the runoff, that 6% stays with A.

These are post facto strategies though. You can only implement that strategy if you can guarantee all of that information. Which you can only do after the election.

The first problem is that polling is an estimate. If you are wrong by a couple of percentage points, the strategy fails. If the results come back 16 (BAC), 16 (BCA), 23 (ABC), 39 (CBA), 6 (ACB), B now has 32 first votes and A only has 29.

Next is getting the right number of people to comply. You could easily just mess it up and have either not enough or too many people cast the strategic ballot. If only 4% or less vote ACB, the plan doesn't work. If 11% or more vote ACB, A wins.

In the end, strategic voting is a bit like voter fraud, talked about way more than is actually done. And implied to be a much bigger problem than it ever could be.

Post reply on HN