Preview - Is Democracy Mathematically Impossible?
Added 2024-08-26 14:22:40 +0000 UTCI'm so excited to share with you a draft edit of the new Veritasium video, which is all about the math behind different voting systems (note that a few sections aren't quite final). I'd love to hear your feedback!
Thanks.
Petr and Team Veritasium
Comments
Great to see this topic being covered, thanks. We've had preferential voting in Australia since Federation, but it's sad how many people still vote for a likely winner thinking that voting otherwise is a wasted vote. Thus we have a two-party system that is not representative. So I wonder if you shouldn't include proportional representation as part of this video even though it's about voting systems, rather than representation systems. Most European countries use it and more US states are moving to it. Tasmania and the ACT use it. Btw, our politicians are as adversarial as they come, so singing 'kumbayah' is definitely a temporary phenomenon!
Simon Cole
2024-08-28 09:13:29 +0000 UTCI just noticed the typo in the word unanimity around 13:00. Too late unfortunately, it's already published 😕
Stephen Woo
2024-08-28 03:29:03 +0000 UTCWhile mathematically interesting, I feel that you spend too much time casting doubt on ranked choice voting and rush way to quickly through interesting alternatives. The net result of this structure may make this video a piece of propaganda for those who would prefer that the US continue to use first past the post. Those people exist and are loud already --- I don't understand their motivation; maybe they are just ignorant or listening to people who believe they can more easily buy elections in a plurality voting system. Whatever their motivation, I hope you will think about re-organizing this video to more strongly stress the importance of voting reform in the USA. Clearly, ranked choice voting would be better than what the US has now.
TTST
2024-08-28 03:14:16 +0000 UTCI agree it is tiring, and confusing, and frankly, I did not think it made sense. But I want to know why the theorem is true, not just a "statement from authority".
Patrick W. Gilmore
2024-08-27 18:36:57 +0000 UTC14:10: This confused me. For instance, I do not get why a pivotal voter has so much power. At 15:35, you say you will change "other voters" preference _without_ changing A>C preference. If the A/C pref is not changed, how does P have any effect at all? How does the societal preference flip if no one else in society changes their A>C preference? Plus just listening does not help. Graphics might make it clearer, but the text should change too.
Patrick W. Gilmore
2024-08-27 18:36:18 +0000 UTC20:24: You say "just tally up what percentage of voters approve of each candidate". But just before that, you mentioned giving a -10 to +10 approval rating. I could not figure out how you "tally up" a percentage. I think the illustration is for the case of binary approve/disapprove voting?
Patrick W. Gilmore
2024-08-27 16:54:37 +0000 UTCmisspelling at 12:26: UNANIMATY should be UNANIMITY
Malcolm Dickinson
2024-08-27 12:37:04 +0000 UTCExcellent video, great subject matter, and explained with beautiful clarity! My only suggestions pertain to the sound. (1) background music is too loud throughout, and in some cases too dramatic. (2) the "balloon stretching about to burst" sound effect (at 4:27, 4:31, 4:39, 5:00, etc.) seems like a poor choice. (Perhaps it was intended as a "card shuffling" sound effect but that's not how my ears perceived it, and it was distracting.)
Malcolm Dickinson
2024-08-27 12:30:31 +0000 UTCGreat video although i think you should ditch the long abc and cab part which is very tiring mathematically and skip straight to the conclusion
Eran W.
2024-08-27 05:17:51 +0000 UTC@0:20 "the methods we're using to elect our leaders are fundamentally irrational" Would you be able to begin the video with expectations and premises? The term "irrational" leaves a lot to the imagination. @3:41 Kudos for the bestagons @14:09 The Arrow proof is difficult to follow as written. The mathematical notation was created to condense and simplify generalized concepts so that they could be written more concisely by hand, not to be read out loud. That said, we could still pause the video repeatedly as we try to follow along. Was that the intention? @15:42 Is the "Independence of Irrelevant Alternatives" rule being defined in such a way that it cannot exist? Was it presumed that the introduction of "P" could _ever_ be completely independent of the popular evaluation of "C" > "A"? It's as if a value system was presumed in which people could evaluate their choices independent of any social dynamics. Yes, trying to account for the value systems of social dynamics would probably make the math impossibly messy, but dismissing it entirely would result in voting model that has been distanced from how most people operate. @18:38 "So, is Democracy doomed? Well, Arrow's Impossibility Theorem seems to say so." Would it be better to say that democracy is an ideal that can never be reached (by this definition, at least)? For this theorem to doom democracy, then democracy would have to exist, but since Arrow's five rules for a democratic voting system can never all exist at the same time, then his version of democracy could never exist in the first place. @18:45 "...there is no ranked choice method to rationally aggregate voter preferences." What is the rational way, then? The appeal to rationality was also at the beginning of the video, but what would be the rational way? Yes, I'm asking you to say and "such and such would be rational, otherwise not". @21:10 "The use of 'First past the post' voting feels, quite frankly, ridiculous to me given all of its flaws." It needs to be logistically feasible too. Given America's technological state when that system was first put to use (no electricity, no plumbing, no trains, no steam engine, no telegraph, no gas lights, very little education, but you do have parchment, quill, abacus, and people on horseback), it was unheard of. Greek and Roman voting was a local affair and therefore feasible during their time. America's national voting system (even if it only extended to the Electoral College) was the largest successful attempt at voting in human history up to that point. And yes, we can also do even better now, and we should try :) Great video! That approval based voting system at the end is elegant in its simplicity.
chromicacid
2024-08-27 02:32:59 +0000 UTCLove the new video and topic! I am a volunteer with an organization (https://rankmivote.org/) that is working to get Ranked Choice Voting instituted at the state level and added as an option at lower levels via a state constitutional amendment in Michigan. The language we iron out with the gracious help of pro bono lawyer hours will be needed so we have the verbiage exactly as it would be ratified before collect signatures in 2025 for to get it on the ballot in 2026. We'll be holding discussions within the group, and a more tabular comparison of the methods could help us outline the trade-offs involved between the proposed solutions. For example, does the Borda method at least guarantee unanimity? Another factor is what if some voters just mark one candidate and the candidate is eliminated in a ranked removal round or needs to be added to other values in the Condorcet approach (a concern is what if voters don't understand or don't want to rank each choice)? In your opinion, would you scuttle a ranked choice approach for a rated choice approach? A big selling point of the current campaign is to make the change less scary reporting each person still only has a single (transferable) vote. Does the Condorcet and rated choice give the impression (or even mathematically result in) a person having as many votes as they opt to fill in a value for including write ins? The new constitutional language should also work for electing more than one winner such as selecting 5 members for city council. Are there any details you could add to the video to touch on any differences you are aware of? One option would be proportional choice voting.
Mark Jochum
2024-08-27 02:21:11 +0000 UTCSomething mundane: I couldn't find any dictionary that spells it "unanimaty", only "unanimity". The fifth of Arrow's rules might need some more explanation. Generally it is not necessarily true that introducing a new option needs to leave my evaluation of previous options unmodified. Say I favor candidate A over candidate B because they want fewer infuence of unions in the work market. But now enter a very charismatic new candidate who I might not favor over A and B but whose arguments change my opinion about unions and hence my preference between A and B. In my personal opinion, only if C doesn't modify my preference between A and B, the voting system should respect that. If Arrow's fifth rule is supposed to be stronger than that, I would consider it questionable. I agree that Arrow's theorem gets a lot of time compared to the options mentioned after that.
Lionel Pöffel
2024-08-27 00:03:21 +0000 UTCI somewhat agree with @mpartel about the Arrow proof at ~14:08. But a really well-done visual might help. At 21:16: "But just because things aren't perfect, doesn't mean we shouldn't try. Being interested in the world around us, caring about issues, and being politically-engaged is important." Germane to that quote ... I use a site called ISideWith that's a political stance & voting guide. I like it because it uses empirical data (in the form of your answers to 200+ questions) to identify your alignment with political ideologies, themes, topics, parties and candidates. Because of that site I have a *much* better sense of where I fall on the spectrum like what was shown at ~ 19:03 when talking about Duncan Black. (I always knew the general region, but the site honed in significantly.) The site also is global, not just US-centric. Though a resident of the US, I have actually switched my location on occassion to France so I could take their specific questions and see how I align with that country. (I'm a Francophile.)
Kimberly Green
2024-08-26 21:33:46 +0000 UTCThe US system with the electoral college adds another layer of problems to the mix.
Gregor Shapiro
2024-08-26 16:05:17 +0000 UTCObservations: - The proof at 14:08 was difficult to follow, and I don't think visuals will help decisively. - It's not clear whether Arrow's theorem deserves so much attention. Arrow's theorem reminds me of how the impact of the Halting Problem is easy to mispercieve: yes, you can't have a fully general prover for halting, but you *can* prove halting for a vast number of practical programs. The halting problem, for all its elegance and generality, doesn't have much to say about how difficult it is to prove halting in your everyday programming. Similarly, it's not clear how "impactful" Arrow's theorem is in different voting systems (as you allude at the end). Suggestion: I'd like to see the final part expanded and the middle part shrunk. That is, I'd love to trade details about Arrow's theorem for more empirical (and theoretical) data about the consequences of different voting systems.
mpartel
2024-08-26 15:14:02 +0000 UTC
