The 2024 election politics

The RCV proponents here keep repeating their same tired arguments. Here is some data. RCV did poorly in the 2024 election cycle despite proponents outspending opponents by 20:1

https://news.ballotpedia.org/2024/12/10/five-states-reject-ranked-choice-voting-measures-alaska-retains-system-after-recount/

Statewide measures

A recount in Alaska concluded on Dec. 9, confirming that the votes to retain RCV in the state, which was adopted in 2020 and implemented in 2022, had the majority by 743 votes. The defeat of Ballot Question 2, which would have repealed RCV, is the only victory from the Nov. 5 election for RCV proponents at the state level. Voters in Arizona, Colorado, Idaho, Nevada, and Oregondefeated measures that would or could (in the case of Arizona) have adopted the system in their respective states. The average “no” vote in states that voted against adopting RCV was 58.5%, with Idaho having the highest at 70.0%.

One state, Missouri, approved an amendment to preempt the adoption of ranked-choice voting at the state level. The vote margin was 68.4% to 31.6%.

Including Washington, D.C., proponents of RCV spent over $66.1 million on statewide measures to adopt or maintain RCV as of Dec. 10, which was 20 times the amount opponents of RCV spent on statewide measures ($3.3 million). Article Four was the top donor supporting RCV, giving $17.4 million total to two campaigns supporting RCV, $13 million in Nevada and $4.4 million in Alaska. Unite America was the second top donor supporting RCV, giving $15.4 million total to three different supporting campaigns: Colorado, Idaho, and Nevada. Nevada Alliance was the top donor opposing RCV, giving $2.1 million toward the campaign to oppose RCV in Nevada.

You are the only one repeating the same tired arguments, trying to make it a popularity contest rather than even attempt to discuss the actual merits and fundamentals of what RCV is.

Any voting system is a “popularity contest. I showed data that RCV is decisively losing in elections

Here are problems with RCV from Grok

  1. Violates “One Person, One Vote” and Monotonicity

. Non-monotonicity: Paradoxically, ranking a candidate higher can cause thatcandidate to lose, or ranking them lower can cause them to win.
Example: In the 2009 Burlington, Vermont mayoral race (widely cited by voting theorists), raising the Progressive candidate from 2nd to lst on some ballots would have made him lose instead of win. Estimates suggest 5—15 % of real-world RCV elections exhibit non-monotonicity strong enough to potentially flip the result.
This breaks voters’ intuitive expectation that supporting a candidate more should never hurt them.

  1. Ballot Exhaustion and Effective Disenfranchisement

. Voters who only rank one or a few candidates can have their ballots “exhausted” (discarded) in later rounds if all their choices are eliminated.

  1. Center-Squeeze and Elimination of Consensus Candidates

    RCV tends to eliminate moderate or broad-appeal candidates early if they lack enough passionate first-choice support, even if they would be the Condorcet winner (beats every other candidate head-to-head).

.Famous example: 2018 Maine gubernatorial race — moderate independent Eliot Cutler was eliminated early despite polling evidence he would beat both the eventual winner (Democrat Janet Mills) and the Republican in pairwise matchups.

  1. Complexity and Voter Confusion
    Voters must understand ranking strategy, and errors (overvotes, undervotes, or unintentional bullet voting) are higher than in plurality.

.Tabulation is far slower and more expensive: San Francisco and NYC have taken weeks to declare winners, compared to same-night results in most plurality elections.

Low-information voters are more likely to bullet—vote (rank only one candidate), which effectively turns their ballot into a plurality vote and increases exhaustion risk.

  1. Strategic Voting and “Buried” Candidates
    Sophisticated voters can still manipulate outcomes by bullet-voting or ranking insincerely (e.g., ranking a weak opponent higher to knock out a stronger rival).

In Australia (which has used RCV/IRV for a century), “how—to-vote” cards distributed by parties encourage disciplined ranking that often overrides voters’true preferences.

  1. Spoiler Effect ls Reduced but Not Eliminated
  • While RCV largely prevents classic third-party spoilers, a weak candidate who stays in longer than expected can still distort later rounds (the “turkey—raising” or “zombie candidate” problem).
  1. Partisan and Ideological Bias Concerns

.In practice, RCV has sometimes helped Democrats in the US (e.g., Maine 2018 House race, Alaska 2022 Senate and House races where Republican votes split and were redistributed toward Democrats), leading Republicans in several red states (Florida, Tennessee, Idaho, Montana, South Dakota, etc.) to ban or restrict it since 2020.

.Critics on the right call it a “left-wing reform” pushed by well-funded NGOs; critics on the left complain it still favors major parties in multi—winner versions (STV) and doesn’t go far enough toward proportional representation.

  1. High Implementation Costs and Central Count Dependency

.Requires centralized tabulation (no precinct-level results on election night), new voting equipment or software, and extensive voter education.
Audits and recounts are much harder than in plurality systems.

  1. Does Not Guarantee Majority Winners in All Cases

. Because of exhaustion, winners can receive less than 50 % of original ballots cast (e.g., 2021 NYC mayoral primary: Eric Adams won with 50.4 % of the remaining ballots, but only ~40 % of total ballots cast).

No, it wouldnt. As you example shows, changing who the voters vote for will affect who wins. Which is how elections work.

And one example from over 15 years ago? Really?

No, voters will have voted for a candidate who lost. Like in literally every election ever held.

Can you rank your three most favorite colors? If you can do so without getting confused, you can vote in a RCV election without getting confused.

So they’ve kept it for over a century? Why didnt you include this in your list of how everyone is abandoning the concept? And yes, party-line voting is common here too.

So RCV is better?

Fair enough. But it does guarantee the winner has more support than the second place option, just like any other election.

“Could”, not tends to. And even then, how terrible that the top votegetters are based on votes, and not party gamesmanship.

And as a third party candidate in a traditional vote, he would’ve suffered the same fate. He only beats one if the other is eliminated as an option.

Actually… I think this is a valid point and I think it happened in 2022 in Alaska. I had to dig deep (the “widely cited” Burlington race paper explains this, but it wasn’t clear to me) and the best example I could actually understand is in this reddit comment (and followup): https://www.reddit.com/r/explainlikeimfive/comments/1o1byqi/comment/nilcv4f/. Long story short, in a race between 3 candidates, the winner depends on the 2nd choice of the voters whose 1st choice was eliminated. Effectively the winner is determined not by who is the most popular, but by who is least popular. So “ranking the candidate A higher can cause the candidate A to lose” (to B) can be true if that candidate A was neither eliminated (C was eliminated) nor a 2nd choice of the voters whose votes get redistributed in the 2nd round. If more of A’s voters selected B as their first choice and A as second in such a way that B was eliminated instead of C, then A could win if more of B’s voters selected A in 2nd round, than C’s voters. It’s a paradox. And it’s rare, and it can’t be done on purpose.

1 Like

I’m pretty sure what you wrote here doesnt say what you were trying to say.

The point being made was that the winning candidate can be listed second on a number of ballots, but if he was listed first on those same ballots he would’ve lost. Because it could potentially change who gets eliminated first, which then affects what second choice ballots come into play, which then affects subsequent rounds. But that would be a fluke, because in general a second choice candidate B being counted means the first choice candidate A was eliminated - and with fewer first choice votes (instead slotting in candidate B) candidate A will still be eliminated.

I don’t think so. The point being made is that in some circumstances, ranking your most preferred candidate A in the second position behind the least popular candidate C could lead A to win the election (C is eliminated and votes redistributed to A). While ranking your preferred candidate A in the first position could get A eliminated if A is the least preferred by all voters in the first round.

You would of course be taking the risk of putting the least preferred candidate in the first position, and if too many people did that, you’d end up with the worst candidate winning. Which is why it can’t be done on purpose, but it can happen on accident as shown in at least two elections (2009 Burlington and 2022 Alaska).

Or maybe I’m wrong. I gotta go re-read the explanation. The wikipedia article on Alaska explains it pretty well I think. In this case not enough voters who voted for Begich (eliminated) selected Palin as their second choice, and enough of them selected Peltola, who won a majority in the second round even though her name was not on the majority of all the ballots. I actually don’t see any problem with this result. It seems like a problem, but it’s really not. The apparent issue was that a few thousand Begich voters didn’t write in anyone in the 2nd position. But it’s not a problem, because by not writing anyone in, they allowed other voters to decide who wins.

Yeah so I re-read this for the number of votes and I don’t think Grok is telling the truth. The Progressive candidate (Kiss) was in 2nd place in the first round. He picked up enough votes in the second and third rounds to win. Raising him from 2nd to 1st on any ballot would have simply given him more votes in the 1st round and he would still have won. This does not break voters’ intuitive expectation.

Some of the Grok slop probably sources from the following, emphasis added:

I think the problem is that you can’t cast votes against a candidate by ranking them. Any vote cast could only count for a candidate. Voting against requires not ranking a candidate at all, only ranking all the others instead. But also I’m not sure that sentence even makes sense, because if he was ranked last on those ballots, then those votes would have gone to Wright, who got 2nd place. They must have put Kiss last but omitted Wright entirely, implying that they actually voted against Wright, not against Kiss. The referenced math papers require a subscription.

But A will only receive LESS votes (and still be the least preferred) if you rank them 2nd and someone else 1st. And if C is the least popular when you list them 1st, they will remain the least popular when you dont list them 1st.

Switching your order can only result in A becoming the least popular and being eliminated, or elevating C out of last place and resulting in somoene else being eliminated.

Well, yeah, if everyone votes for the worst candidate, the worst candidate is going to win. That’s true in literally every election ever.

Let me rethink this a few times. I think I am explaining it wrong, but the reddit example makes sense so let me try with numbers.

Imagine an election in which candidate A gets 39 votes, B=31 C=30 in round 1. C is eliminated, votes are distributed evenly to each, A wins.

If before the election candidate A convinced 2 of B’s voters to vote for A instead, the result would be: A=41, B=29, C=30. B is eliminated. If 21 B voters put C in 2nd, C would win instead of A.

In other words, in order for A to win, some of his voters must put B in first position in order to eliminate C.

1 Like

Yes, there is a butterfly effect. But, it’s just as likely 21 B voters put C second, as it is 21 C voters put B second. So A has a solid path to a loss either way. As I said,

It was the second part of what you said that I have a problem with

I think the letters got switched up in my last example, but your conclusion was that the same candidate was eliminated either way, but in my example different candidates are eliminated, which changes the winner.

Long story short I think we both understand it. I’m just trying to decide if this flaw is sufficient to change my position on RCV :slight_smile: .

I guess my point was this is not direct cause/effect, it’s a butterfly effect. A different person winning is 2-3 steps past this one person changing the order of their candidates, you cannot strategize your vote to product a different winner unless you are doing so after the fact and can see all the other ballots cast. And I suspect you could find such potential in every RCV election if you dig through all the vote tallies. It is not a direct effect, like the effect of a third party spoiler syphoning off votes in a traditional election.

Again, it’s not a “flaw”. It’s a simple fact that when you change the votes cast, it’s going to change the results.

1 Like

Yes, scripta has the basic issue here correct.

You can understand the RCV outcome by thinking about a three candidate race where you first hold a primary between two particular candidates. A B C are in the race, but suppose we first have A vs B in an open party primary, the winner of which will face C in the general election.

It is well known that the outcome of the overall election can vary depending on which two candidates are in the primary, which then determines who faces the other side’s challenger in the final election.

For example, in a close primary between A and B, the other party’s C voters should vote not for the candidate A or B they would like to be elected as 2nd choice to C, but for the candidate most likely to lose to C. Likewise, primary party voters who want their party to win over C should not for their first choice of A or B, but for who they think can beat C. This leads to “strategic voting” issues where it makes sense to not vote your preference due to the election structure to achieve your overall goal.

This can happen in RCV too, but if everyone does it too much, you can elect some idiot minor candidate like president “the rent is too damn high”.

[Sideny] Morgenbesser, ordering dessert, is told by a waitress that he can choose between blueberry or apple pie. He orders apple. Soon the waitress comes back and explains cherry pie is also an option. Morgenbesser replies “In that case, I’ll have blueberry.”

2 Likes

But as you illustrated, it happens (or can happen) in all elections. Arguably, it’s more direct in the traditional system, as you showed. In RCV, it’s all blind, so your effort to manipulate is as likely to be counterproductive as it to get the result you want.

Right, OK, I’m back on the RCV boat :slight_smile:. My whole big idea behind RCV is that it corrects the potential problems with primaries: you want the best candidate to advance to the next round, but you have to pick one who has a chance of winning the next round, and your opponents want your worst candidate to advance. I.e. Bernie wasn’t likely to beat Trump, but Biden was, so D’s favoring Bernie are stuck voting for Biden, while any Rs pretending to be Ds (rare, I’m sure, but probably non-zero, as are Ds registered as Rs) could be voting against Biden. With RCV you can rank them in order you actually want without having to strategize too deeply.

1 Like

Another new, sure fire voting system runs into implementation problems. LOL

One of the world’s premier security organizations has canceled the results of its annual leadership election after an official lost an encryption key needed to unlock results stored in a verifiable and privacy-preserving voting system.

The International Association of Cryptologic Research (IACR) said Friday that the votes were submitted and tallied using Helios, an open source voting system that uses peer-reviewed cryptography to cast and count votes in a verifiable, confidential, and privacy-preserving way. Helios encrypts each vote in a way that assures each ballot is secret. Other cryptography used by Helios allows each voter to confirm their ballot was counted fairly.

The implementation did exactly what it promised - verifiable, confidential, and privacy-preserving. If you lose the key, you can’t open the lock.

3 Likes

Verifiable? There were no results to verify.

Confidential. Suppose so. Even the voters do not know how they voted.

privacy preserving. How so? The public knows they participated in this fiasco.