Show HN: RankedVote – SurveyMonkey but focused on ranked-choice voting • Hacker News Board

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

#111

Post by lambertsimnel » Thu, Oct 06, 2022, 6:46 PM UTC

gamegoblin wrote: post #60

I used to be a hardcore pro-RCV supporter (I liked Yang, and Yang made RCV one of his main issues, outside UBI), but I've become a little more jaded in the last few years. Empirically, it seems like RCV-IRV (RCV from here on) not much better than Plurality. For example, look at all the countries and cities with RCV and without. It's not at all clear to me that those with RCV are much better than those without. Both N…

NZ and a great many European countries use neither plurality voting nor RCV-IRV for their nationwide legislative elections. (France and some US states use a two-round system, which is essentially a bastardised form of RCV-IRV.)

Some European countries (at least France and Ireland) use RCV-IRV or a two-round system for their presidential elections, but Australia, NZ and several European countries are constitutional monarchies, and therefore don't have presidential elections.

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

#112

Post by crazygringo » Thu, Oct 06, 2022, 6:48 PM UTC

NickM wrote: post #94

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…

it would be a shame to spend our effort squabbling over which alternative, than giving support to the one that has the most momentum and chance of succeeding now I think the bigger shame would be to spend so much effort promoting RCV, only to have people become jaded and disillusioned with voting method reform in general when they realize RCVs many glaring problems. The biggest problem with RCV IMO is that it looks l…

> 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 could "poison the well" even more. When people have to worry about what it means to give a 3 vs a 4 to a candidate, or whether to approve 1 or 2 or 3 people?

Ranking candidates is an objectively accurate statement of preference (in contrast to STAR/approval which are subjective), and while instant runoff isn't perfect, it's easy to understand and doesn't result in absurd outcomes or people not understanding how to vote.

You can criticize RCV for sure, but if you want to defend another method, you have to show how its drawbacks aren't even worse.

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

#113

Post by dane-pgp » Thu, Oct 06, 2022, 6:49 PM UTC

scythe wrote: post #35

bhauer wrote: post #32

Ranking the voting systems: STAR Voting > Approval Voting > Ranked Choice Voting > Plurality ("pick only one") voting. Ranked Choice Voting is marginally better than plurality voting, but it has problems. The chief defect with Ranked Choice Voting is its non-monotonicity, whereby increasing your support for your genuine favorite can actually hurt their odds of winning. This may be what happened in Alaska [1]. STAR Vo…

Regarding the Score/Approval/STAR trio, I always felt that there was a clear fourth possibility missing from the list, so I invented it: https://pastebin.com/PCgSaBS8 Still, I'm not much of a political campaigner, so it hasn't gone very far.

That's a clever system, thanks for sharing it. What I appreciate most about it, though, is the explicit goal of "being easier to count by hand", which I think many advocates of voting-reform overlook.

You even say "by sorting ballots into each of thirteen piles corresponding to the thirteen possible rankings of three candidates", which shows you've actually thought through what the process would look like for the people doing the counting.

When people are proposing voting systems, they really need to consider that their process is being carried out in an adversarial environment, where different groups have incentives to cheat, or accuse the voting system of producing the "wrong" result (e.g. due to the unintuitive outcome, or the complex tactical thinking needed to optimally fill in the ballot), or claim that voting machines were running backdoored firmware (which is a destabilizing claim whether it is in fact true or false).

That last point is perhaps the most important. I think that as time goes on, the pressure to automate the counting process and give instant election results (or worse, do the elections online) is only going to grow, and any voting reform which (if done manually) slows down the counting process, or for any other reason in practice requires people to trust machines that are selected and maintained by officials from one political party, is going to do even more harm to democracy than FPTP does.

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

#114

Post by kirse » Thu, Oct 06, 2022, 6:50 PM UTC

elihu wrote: post #87

kirse wrote: post #54

Earlier quoted context omitted.

Yea, it's been interesting to watch how RCV has filtered into the public consciousness over the past 5-10 years and slowly increased in terms of how many people parrot it as the answer to all our voting ills. There was a site posted to HN years back that mathematically analyzed a ton of these voting schemes, I cannot seem to find it on Google but it was some sort of hybrid scheme that apparently has the best overall…

> There was a site posted to HN years back that mathematically analyzed a ton of these voting schemes, I cannot seem to find it on Google but it was some sort of hybrid scheme that apparently has the best overall outcomes. It might have been "To Build a Better Ballot": https://ncase.me/ballot/ https://news.ycombinator.com/item?id=13161396

Good looking out, RangeVoting.org was the site I was thinking of but that HN post drove the discussion.

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

#115

Post by unholiness » Thu, Oct 06, 2022, 6:53 PM UTC

TimPC wrote: post #105

bhauer wrote: post #32

Ranking the voting systems: STAR Voting > Approval Voting > Ranked Choice Voting > Plurality ("pick only one") voting. Ranked Choice Voting is marginally better than plurality voting, but it has problems. The chief defect with Ranked Choice Voting is its non-monotonicity, whereby increasing your support for your genuine favorite can actually hurt their odds of winning. This may be what happened in Alaska [1]. STAR Vo…

Isn’t giving each candidate a score even more gameable than ranking candidates. If I like candidate A and think candidate B is their best opposition I can give candidate A 5.0 and candidate B 0.0 even if my real opinion of candidate B is 3.0. Lowering candidate B’s score artificially definitely helps candidate A.

Absolutely. Naive exaggeration is a fantastic way to make your vote count more than others in score voting. It also encourages more negative campaign tactics, since convincing a voter to rank your opponent 0 rather than 4 is literally 4 times as powerful as convincing them to rank you a 5 rather than a 4.

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

#116

Post by fasthands9 » Thu, Oct 06, 2022, 6:53 PM UTC

godelski wrote: post #69

fasthands9 wrote: post #59

Earlier quoted context omitted.

One thing that I think gets lost in political debate around voting is that RCV also changes the behavior/selection of the candidates. Its not just that the candidates are identical and you vote with a different mechanism - but candidates have an incentive to moderate. This may mean the outcome is the same (in terms of which candidate wins) but their positions represent the median voter more. I would imagine approval…

That's kinda the entire point. One of the (many) metrics for voting systems is called Voter Satisfaction Efficiency (VSE). This is basically a global L2 distance from each voter to each candidate. To maximize VSE we want to minimize that distance. In other words, we want winning candidates to be most representative of a population's opinions and values. You can see a comparison of different systems here[0]. Note that…

If I'm not mistaken this link is exactly what I am describing. It simulates how close the winning candidate is to voters based on how people would vote. But it explicitly doesnt simulate how candidates would shift due to changing incentives.

In other words - it treats elections as a one side game where voters pick strategy (voting methods/preferences) and the candidates are stable. But it is a two-sided game where politicians change opinions/policies based on system as well.

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

#117

Post by lostdog » Thu, Oct 06, 2022, 6:57 PM UTC

godelski wrote: post #71

gamegoblin wrote: post #68

Earlier quoted context omitted.

This is actually not the case - RCV suffers from the "center squeeze" effect [1], in which moderates are squeezed out in the early rounds. That said, things like Approval have the opposite problem, in which moderates are given an advantage (some approval people will say this is a feature, not a bug). For an edge case, consider an election with 3 factions: socialists, libertarians, and fascists. There is also a 4th ca…

> That said, things like Approval have the opposite problem, in which moderates are given an advantage (some approval people will say this is a feature, not a bug). This is a weird way to say that representing the population is a bug. Your example election would be a compelling argument if population preference was more of a multimodal distribution rather than closer to a unimodal distribution. Despite the US's curre…

Voters will realize that approval gives benefits to centrists, and they will change their voting strategies and stop approving centrists. None of the math analyses shared here take that into account.

For example, say three fictitious candidates are running for dogcatcher, Bernie, Hillary, and Trump. Does your hardcore Bernie supporter friend believe that approving Hillary doesn't hurt Bernie? Will they choose not to approve centrist candidates, leading to extreme results in elections?

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

#118

Post by dane-pgp » Thu, Oct 06, 2022, 6:59 PM UTC

c7b wrote: post #62

Feature suggestion: the system automatically tells you when you have a https://en.wikipedia.org/wiki/Condorcet_paradox situation and offers to decide the outcome based on a coin flip instead. Only half joking there, the amount of negative results about ranked-choice voting in social choice theory is quite amazing. I'm not sure whether I'd place it on the efficiency frontier of the 'complex but has desirable propertie…

While we're offering troll feature suggestions, I have another:

Add a calculation for how many person-hours such a ballot would have taken if carried out by hand (based on scientifically validated assumptions about how long one person takes to sort through a stack of N votes).

This would also let you give a figure for how much money had been saved by people entrusting their democratic process to an inscrutable and unaccountable online system, and therefore the size of the financial incentive to do so.

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

#119

Post by theptip » Thu, Oct 06, 2022, 6:59 PM UTC

godelski wrote: post #80

theptip wrote: post #77

Earlier quoted context omitted.

> How long would it take to teach a member of the general public to evaluate IRV results? I'm puzzled by this. You can do it with pencil and paper in a few minutes? It's simple addition. You can easily build a website that visualizes each step if you think addition is difficult for people. It's trivial to compute in a spreadsheet. > might leave people disillusioned This is speculative; if it is true, you can find evi…

> It's simple addition. Not quite. The tallying algorithm is `while (no candidate > 50% of votes): pop min(candidates), reallocate_votes(min(candidates))`. On the other hand cardinal methods are `argmax(sum(candidates))`. For comparison, FPTP is `argmax(sum(candidates))` too. The only difference is that FPTP is a one-hot vector. I'd argue that the average person intuitively understands cardinal methods without any tr…

You're right, my bad. Computing `reallocate_votes()` is not as trivial as I suggested.

I do wonder if many people actually care to work through that step; in all of the public discourse around the Alaska vote, it's simply presented as a table with the reallocations already summed. For example, https://fivethirtyeight.com/features/what-democrats-win-in-a....

The general algorithm of "reallocate the votes from the candidate that is being eliminated" is conceptually simple, even if it's annoying to tabulate. I do stand by my claim that it would be very easy to build a website that can visualize this process without getting bogged down in the tabulation. See every site that's giving analysis on the Alaska results for examples.

I want to see actual UX studies on this; for example has anyone polled in Alaska to see whether people understand the result?

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

#120

Post by manholio » Thu, Oct 06, 2022, 7:09 PM UTC

crazygringo wrote: post #88

bhauer wrote: post #32

Ranking the voting systems: STAR Voting > Approval Voting > Ranked Choice Voting > Plurality ("pick only one") voting. Ranked Choice Voting is marginally better than plurality voting, but it has problems. The chief defect with Ranked Choice Voting is its non-monotonicity, whereby increasing your support for your genuine favorite can actually hurt their odds of winning. This may be what happened in Alaska [1]. STAR Vo…

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…

> I just want to make clear for others: this "ranking" is just bhauer's opinion.

It's also a non-sensical claim, any voting system "ranking" depends on specifics of the electoral system. For example, larger multi-winner districts using some sort of proportional or Condorcet system are always better that any ranked or approval result in a single winner districts.