My city voted against STAR voting (a Rated Voting system) because the people who were lobbying against it were the ones in positions of political power--and they benefit from first-past-the-post voting. They knew that if we used STAR voting, their extreme views would be likely supplanted by more moderate candidates. 😢
The pivotal voter thing doesn’t make sense, because it’s not actually an individual person. You shuffle up the votes, and one random ballot happens to be the pivotal voter. No one knows who it is ahead of time, and even more important, that vote being pivotal relies on all other voters having made a specific decision. Had they voted different, that would no longer be the pivotal vote.
Yeah, the "dictator" ends up being a random person, who doesn't even know they're the dictator for that election. It's one of those math things that doesn't translate to real life as well. Having one determining vote is how voting works sometimes, that's why you always use an odd number of voters, so in the event of a tie, one vote will break that tie.
In this field it's hard to make the explanations make sense intuitive, because the conclusions are sometimes not but also the terminology is confusing. Unfortunately Arrow's theorem does mean that it's not just a pivotal voter who is unknown before the election. It means that the rule of the election is "look at how this person voted, and say that is how society votes". Unless it is such a rule (the dictator can be selected randomly, but it's not just a coincidence, it is a pre determined person outside of the vote counting process) you cannot have both unanimity and independence of irrelevant alternatives. It's a looser rule in special cases, but no generalized ranked system passes.
It does make sense, though it is difficult to highlight constructively. First you show that any 'dictatorship' (i.e., a voting system with a pivotal voter) satisfies the other 4 conditions. Then, you assume that a voting system satisfies all rules except potentially the non-dictatorship rule. You can then show rigorously (with the type of argument that was presented in the video) that there MUST be a pivotal voter. You can't know which person it is because you are making a general argument for all voting systems. But you did show that every such voting system must have a pivotal voter.
love the CGP Grey references, he introduced voting systems to my mind sphere. Thanks Derek for further expanding on the topic, would love to see different scenarios for approval voting and +/-10 points voting
18:40 that pivotal voter isn't really a dictator though, because he requires all the other voters to set up the scenario in which he would become, unknowingly, the pivotal voter. That pivotal voter would never know or be able to abuse their dictatorial status in practice? The grain of sand that tips the scale isn't a dictator. Edit: I understand that this question is largely philosophical and has little bearing on a scientific debate/discussion but I felt it's worth examining for the sake of the larger conversation about where autonomy of the voter ends.
They become, mathematically, the dictator that changed the outcome for everyone. Doesn't matter if there's only 2 choices, you'd never know who the pivotal vote would be, but with 3 or more choices you can construct scenarios where you can find the single ranking that influences the outcome (which is what is shown in the video, the position doesn't matter). Doesn't mean it has to happen every time, heck the more voters there are, the less likely it will be. But if you can mathematically show the case can exists, the voting system isn't "perfect" enough to adhere to that condition.
100% agree. One can only be a dictator with the foreknowledge that their vote is pivotal. If you can't leverage your "power", you have none. You're best off just voting your heart.
The pivotal voter only becomes a thing if there's discussion and negotiation where everyone knows how many votes they have (like what happens in Congress) versus just having an election day where everyone goes in and votes and nobody has any idea what the vote count is. Even in the first case, that only happens if only 1 voter is willing to negotiate and the others are immovable. But that's more a testament to how negotiating in politics can get you favors.
I absolutely loathe the first-past-the-post voting. My entire life we've ended up with a system where everyone votes AGAINST the candidate they dislike the most rather than voting FOR the candidate they actually support the most. The amount of disinformation that has been spread when we've tried referendums to move to proportional representation or alternatives is infuriating.
why is that a problem? Sometimes that's the choice you face as a country - it says more about the political climate in your country than your voting system. And arguably getting rid of the guy you don't want to be in charge is the single most important requirement of a voting system.
you can only have one unified entity that is the administration. it can never be representative. so called representative systems are not representative at all, since there is coalition building that is entirely out of control of the voters.
Support the Equal Vote Coalition and help us get past this challenge! Ballot initiatives, city council ordinances and pressure from citizens can make it happen. The good thing is not all politicians staunchly align with their party. City council members understand the issues with FPTP and some are open to change.
@@rasta77-x7o You mean 30% of the FIRST PREFERENCE votes. Preferential voting will always be objectively better because it considers multiple candidates for individual votes. If you believe in a democratic society then you should not be rooting for fptp...
That approval voting actually excites me. When you started explaining it I started seeing a light at the end of the tunnel. That's pretty cool I would totally participate in that type of election
In Europe, with proportional voting, parties form coalitions after elections, because it's very rare for one party to win a majority of districts. In the US, coalitions are formed IN parties. That's an inevitable results of 'winner takes all' elections, and the fact that it's impossible to get a majority to agree on every significant issue.
@@caldodgethis is a total myth. Even duverger himself noted that countries which use a top two runoff tend to escape duopoly. There's absolutely zero reason to expect single winner districts to produce a duopoly. What produces a duopoly is the plurality voting method.
Plus the person put in charge is not voted for directly, but the coalition selects them. They can't go against the will of the government or people, and the government reflects the mix of the people in the country much better, even if the voting system would be slightly off.
I was looking for the words but you summed up my thoughts perfectly. I think the American perspective on democracy is culturally different from some of the Europe and prone to misunderstandings. It's complicated to explain every system and would probably take a lot of videos to cover I believe.
Who needs districts if your country is x times smaller than the average US state?😅🥳 Still, election results can go against the will of the majority, albeit partially. Forming a coalition can mean that the least liked party will be in government for example and the second most liked is not. Speaking from experience, having no say in who becomes prime minister or who takes a seat in the senate can be a bit of a bummer as well. Looking at our own version of MAGA-ism, even though our guy has been bothering our parliament for far longer than Trump, our version of democracy is definitely far from perfect at this moment. But that has more to do with the way people "do" politics these days and the ways people (don't (care to)) inform themselves i guess. Now, can you guess where i'm from?😅
In a time when voter confidence in our election system is at an all-time low, the title of this video is very irresponsible. "It's crooked, but it's the only game in town?" Why not just regurgitate everything Fox News says? 🙄 You're the Ralph Nader of this election.
Exactly this. All problems with better way to vote are like "But imagine there is only 3 people voting, and they exactly line up with this or that!" It never happens IRL. And in the so very unlikely even that it does, just redo the vote, people will change their mind. Heck, maybe even remove the least favorite one. But, yeah, that's the problem with maths, they forget the real world application. Country population are counted in millions, one person won't change a thing because such a huge number of people can't think the same thing. And voting for one candidate only is really bad. Even more if you have, like here in France, a two turn based election! It's awful really.
What's Mathematically Impossible is to have a winner-takes-all voting system that's representative of the electorate. If you want to represent the electorate accurately you have to let go of the winner-takes-all system and adopt a representative democracy where parties are awarded seats in parliament proportional to their vote share in a general election.
@@Nonixificationthey could just push the laws that a majority does agree with. You know, like getting pizza, indian or Chinese food... You can still get a majority. Lost of countries have systems that don't use the "winner takes all" system. And it also leads to less extreme choices. But it means that multiple parties will need to work together, and negotiate, to get something they all can somewhat agree on. The "winner takes all" will always lead to a two-party system, and always to very bad politics IMHO.
well his work did lead to us now understanding why rated voting systems are probably the only systems that work- it may not have been his intention but it was informative enough for us to avoid the slander that rated voting systems could have accumulated, people will criticise anything without any understanding of it just because it could inconvenience them. It is like how when you do a multiple choice question it may be easier to prove why everything else is wrong than to prove why the answer is right.
I think you have achieved wisdom. I don't think anyone here will put one and one together, for the videos you've made during this year. You have my compliments.
I liked the CGP Grey reference. His videos about voting systems are phenomenal. They’re quite simple, but I love how he talked about them. Good job Derek! This is a great video!
and they actually left out a key potential issue with ranked choice too, in Australia a few years back a candidate with only 17 number 1 votes was elected to the senate due to preference deals. He served with honour, but that's the clearest actual real world problem with the system, preference deals between candidates
My thoughts precisely. I mean YEAH there are SOME inaccuracies, but then again it's not like first pass the post system ACTUALLY reflects the will of the people. Heck, make it simple: As many Candidates as people present themselves, every voter gets 3 choices. 1st Choice, 2nd Choice, 3rd Choice. The one with the most points wins. Simple as that. Afterall while that person might not be the MOST favourite candidate, they ARE the candidate that most people are reasonably happy with. Heck, give the people the choice to forgo their 2nd and 3rd choice, a la 'Bernie or Bust!' style if they REALLY want to. But in that case they shouldnt be surprised if they end up with a candidate they dont like at all.
I mean if you watch his videos, you know it's usually involves math or science. Thus his title makes sense within the context of his video. Mathematically a democracy would be something where it's free and fair every time. 99 percent free and fair would not make the mark of a democracy. He didn't even cover the peciluar way that the US elects its president. The electoral college works 93 percent of the time. The truth is, in reality, all the methods he talked about are practiced by many different democracies we all recognize as democracies, whether they use first past the post, ranked choices, or an electoral college system. If you actually clicked on a Veritasium video and expected the host to make a political argument that democracy is impossible, you must have never seen any of his videos before. I knew when I clicked on the video he would make a mathematical argument why, not some actual political argument that democracy is impossible.
As a European, I was patiently waiting the entire video for him to mention that of course, democracy can take completely different forms wherein you don't have to elimimate a list of individuals down to one final victor, so that democracy does not necessarily have to stand or fall with this issue, but I was sorely disappointed.
@@keltzar1 If party A gets 48% and party B and C gets 26% each, then party B and C can make a cooperative government and win, or if they cant agree, then Party A wins.. There will be alot more than 3 parties and there is alot of negotiation too see who will cooperate. If noone wants to cooperate, you might end up with a "minority government" wich would be is Party A ended up winning in the first example. Party B and C are not the rulers of the country, but they have a majority for votes and such, while the ruling party has a minority. It can also happen that Party B and C cooperates, but decides to stop cooperating after a few years, then suddenly the winner is Party A. All that being said.. I know nothing.. They teach us nothing in school, and what i believe i know i have deduced and infered from the elections i have voted in.
@@keltzar1 Imagine there's 100 seats up for election. If I get 2% of the vote I get 2 seats. etc. So you end up with a much more diverse Parliament. This type of system also has its problems and critics and on occasion lead to pretty chaotic and unstable Governing.
7:55 idk if I agree with this. Cuz if bohr gives a bad speech and gets less votes, presumably a portion of curies voters would also be dissuaded from voting bohr as second choice due to the bad speech. That is to say, I don't think it'd be a 50/50 split of curies votes in round 2
Well i think it would make sense to assume that before bohrs speech the second choice of curie voters was distributed 25:45 (Same as the initial ratio of first votes between Einstein and Bohr) and after the terrible speech it was closer to 50:50 (Like the Einstein Bohr First Vote Ratio after the speech). So while you're right that some Curie voters switched their second Vote from bohr to einstein it still makes perfect sense as long as the second votes of curie voters are similarly distributed as the First votes between Einstein and bohr
Maybe I'm missing something, but I definitely disagree with what is presented in the video. "Clearly this isn't something we want in a voting system"...? I don't see this as clear at all. Thats kind of the whole point in RCV, even if you're not the people's first choice, you are more closely aligned to how people want you to govern. It also seems like the example has two first rounds of voting. No one would know what the numbers are before Bohr gave his bad speech or whatever.
Also, why are Bohrs voters going to Einstein and not the more moderate Curie? If a candidate performs poorly, voters tend to go to their 2nd choice, which tends to be close to their political position. This example assumes voters swing all the way to the other side which makes no sense.
I disagree with that example also. Why did those voters who were turned off by Bohr's speech all go over to Einstein (who was vastly more unpopular overall) and none go to Curie? Even if just a third of the switching voters went to the centre, Curie would have still won.
Cool, but doesn't this (especially Arrow's Impossibility Theorem) assume that the votes of the citizens are _known_ ahead of time, so that the 'pivotal voter' can change the vote one way or the other? In reality, elections wouldn't actually hinge on a single voter, as they wouldn't be able to know how others ranked their choices _exactly._ In the end, the problem with the current system is not really certain people affecting the vote more than others, but the outcome does not really represent the majority's wishes, especially when you consider the Electoral College and gerrymanderin in the US (Europe and other places luckily doesn't have Electoral College, but I believe gerrymandering is still a problem). Particularly with approval voting, both problems disappear mostly, but it is crystal clear that *ANY* ranked choice method is almost _guaranteed_ to be better than (meaning, a more representative democracy than) what we have now. So it is INSANE to insist on the current system. The only logical argument would be that it is 'easier to understand', and that makes this that much more insane! Hundreds of millions of people, afraid of primary school level maths and logic...
It's not about changing votes or outcomes (or actually exerting the power of a "dictator"). It's just about those 5 demands being mutually incompatible. There cannot exist a voting system that satisfied all of them. That's all.
No it becomes apparent after the fact (or rather during the counting). A basic outline of the proof: 1. You assume you have a voting system where IIA and transitivity hold for any possible combination of votes. 2. You show that in this system, by transitivity and IIA, if every vote ranks B either at the top or the bottom, so must the social welfare function. 3. You show that given step 3, there is some “pivotal” voter that determines B’s ranking relative to AC. Note at this step, the “pivotal” voter is just a tipping point (point of no return), not a true dictator. 4. Now, you show that regardless of how every other voter ranks A relative to C, by assuming IIA the social welfare functions relative rank of A and C must mirror this pivotal voters relative rank of A and C. This shows that there is some voter that fully determines the relative rank between A and C, regardless of how anyone else votes (i.e. is a dictator over AC). Call this voter voter x. 5. By similar logic, there is some voter y that fully determines the relative rank between A and B, and some voter z that fully determines the relative rank between B and C. 6. Without loss of generality suppose voter y has A>B, voter z has B>C. Then the social welfare function has A>C by part 5. But this means that voter x is not a dictator over A and C, contradicting part 4. The only way to resolve this contradiction is that voter x is voter y and voter z. But this means that voter x is a relative dictator over AC, AB, and BC, making them a full dictator. The dictator in this sense isn't consciously choosing to manipulate the voting, but their individual vote has more meaning than every single other vote which is against the common rules set out (that all votes are equal). The video isn't a value proposition of FPTP vs RCV (and any other single winner election). Just that all of them are flawed whenever there are more than 2 choices. Of course PR doesn't have any of these specific issues (but does have multiple other issues).
Why does it matter, anyhow? Conflating "democracy" with "deciding who wins an election" is missing a *major* point when the politicians will just do whatever they want once elected regardless of "promises"...
The main issue with Ranked Choice Voting is that voters will not be able to make educated guesses of each candidate and the possible combinations of runners. Also, a lot of people will not rank more than two candidates, so there will be a faux majority, because others did rank more than 2.
ACKTHUALLYYY they are not the bestagons. Saw a really good video by Con Hathy debunking CGP's and saying the bestagons were instead, generally speaking, triangles.
My man said "It's the best thing we got and only game in town" after a whole video of showing how there are many different ways of doing democracy, the worst possible one is used in most places in the world and the best we could figure out is not used anywhere
Exactly.. he was completely right with everything and then spat into the face of it at the end. CLEARLY a better way is available but everyone just says “uhhh bad German man!1!!” without even looking into it.
@@loslingos1232 Not sure what you mean about the bad German. The video also hasn't mentioned direct democracy, referendums, council rule etc. Or anything about lobbying, media capture, manufacturing consent etc.
The practical problem of a hierarchical democratic organization of the whole state (federal government at the top, then state governments, then local governments, then companies and large families, i.e., clans) is more interesting than these theoretical musings. let's hope we get another video on this topic. After all, in theory, theory and practice are the same. In practice, they are not.
Exactly, proportional representation is a proven and effective form of democracy. Unlike the winner-takes-all system, which often distorts the true will of the voters, proportional representation ensures that the distribution of seats in a legislature more accurately reflects the percentage of votes each party receives. This results in a more representative and inclusive government, where even smaller parties and minority voices have a chance to be heard. The winner-takes-all system, on the other hand, often leads to disproportional outcomes and can entrench a two-party system, stifling diversity in political representation. In contrast, proportional representation encourages a more balanced and fair political landscape, making it a far more democratic approach.
The thing about proportional representation is that proportional representation can also result in extremely disproportional effective power. In the extreme case you can end up with a small party and two large ones just shy of 50% of the seats, and despite this, the small party has the same amount if effective power as the large ones, as any combination of 2 parties has a majority, and no single party has. Now imagine the two big parties being opposides that barely agree on anything, and the small party representing some narrow interest group that only carss about that and nothing else, and you end up with a small "kingmaker" group with way more influence than it's proportional vote.
@@kobyma2 In Denmark, the political landscape is both diverse and dynamic. There are approximately 14 political parties to choose from, at the moment, each offering a range of candidates that you can vote for personally. This personal voting system allows you to give your preferred candidate a significant boost, even if they are not at the top of the party list. Coalitions in Denmark are highly flexible. If a party cannot secure the necessary mandates within its current coalition, it is free to seek support from other coalitions or parties. This flexibility encourages continuous negotiation and compromise, making the Danish political system both inclusive and representative.
Isn't this all assuming that by voting for a party, you support all their stances on all political issues/topics at this exact percentage? A voter's opinion isn't represented correctly in the first place since you don't vote for each topic separately
Honestly Love how this Educational channel makes there Ads in last unlike most who put in right in the quarter part of there videos which is possibly I do not like the most when it’s not presented creatively.
CGPGrey has done so much damage to voting reform selling Ranked Choice Voting. This video does his best to undo that damage by asserting that Ordinal methods are trash. CGPGrey also has an Approval video but insinuates it's only good for picking lunch. imho STAR Voting is best. If I can't have STAR then Approval is best.
I am an engineer, not a mathematician. Let's not make practical the enemy of perfection. We should use the method with the least chance of undesired results. After all, little in life is perfect.
The least one with undesirable results? Tell me, how is democracy helping White people to secure their borders, any of them for that matter, on all of our countries. How is it the fertility rate? Happiness? Health? The safety of our families? It would be one thing if we just stopped having almost any children, but we are bombarded with legal and illegal immigrants across the whole of the West. We are 7% of the human population world wide, we are the minority and we are closing in to 0% in a 100 years, or becoming a tiny r4ce of sl4ves for 100's of years more. I don't understand this cheap, weak and cowardly 3rd rate philosophy which older men cling on to. You are leaving us with nothing but D3ATH.
The least one with undesirable results? Tell me, how is democracy helping White people to secure their borders, any of them for that matter, on all of our countries. How is it the fertility rate? Happiness? Health? The safety of our families? It would be one thing if we just stopped having almost any children, but we are b0mbarded with legal and illegal immigrants across the whole of the West. We are 7% of the human population world wide, we are the minority and we are closing in to 0% in a 100 years, or becoming a tiny race of sl4v3s for 100's of years more. I don't understand this cheap, weak and cowardly 3rd rate philosophy which older men cling on to. You are leaving us with nothing but D34TH.
The least one with undesirable results? Tell me, how is democracy helping White people to secure their borders, any of them for that matter, on all of our countries. How is it the fertility rate? Happiness? Health? The safety of our families? It would be one thing if we just stopped having almost any children, but we are b0mbarded with legal and illegal immigr4nts across the whole of the West. We are 7% of the human population world wide, we are the minority and we are closing in to 0% in a 100 years, or becoming a tiny race of sl4v3s for 100's of years more. I don't understand this cheap, weak and cowardly 3rd rate philosophy which older men cling on to. You are leaving us with nothing but D34TH.
18:30 a thought: if the election is confidential, it means you cannot really know if you are the pivotal voter or not. Everyone is potentially the pivotal voter. So can't you then say that everyone is 1/15 th part dictator, so everyone is basically the same again? Meaning: the collective is the Dictator, which is sort of the translation for Democracy?
or put differently: if an election is close, there is always one voter or small group of voters that would count as dictator. But since there is no order of votes, all those votes are equally 'dictatorial'. Which is usually the Majority
No, this is incorrect, and Veritasium did a terrible job finishing up the proof (which some other commenters and Wikipedia clarified). The dictator is part of the election system. The proof shows that any election system satisfying the 5 postulates has the same one guy as a dictator, regardless of what everyone else thinks. Who the dictator is can only depend on the issue we're voting for (so you could e.g. have a dictator on taxes, a different dictator on food, a dictator on the military) and not any of the voters' preferences. Which means you could theoretically analyze the voting system before the election and figure out who the dictator is. Now it's true you could randomly select the dictator for every election so nobody would know who it is and voting would still be worthwhile, but random systems were excluded from consideration earlier in the video.
I live in a "first past the post" country that forces a two-party system and penalizes voting your conscience unless it aligns with one of those parties. While there may be flaws in Ranked Choice Voting that could emerge in fringe cases, it is so obviously superior to our current system that it is hard for me to worry too much about the nuance of how it might not be 100% perfect 100% of the time. Any (democratic) system is better than what we have now.
The US has had first past the post for over 250 years and only had a "forced" two party system for fewer than 100. Coincidentally the prevalence of the "two party system" began around the time universal suffrage was introduced.
RCV solves the spoiler problem, but it does not solve the representativeness problem. It just makes smaller parties inoffensive to the "big two". It actually performs terribly and erratically when more than two parties are competitive, which is exactly the scenario where you would want a better voting system to matter.
@@scopeguywhat do you mean by this? With very few exceptions, we’ve only ever had two parties really in contention for races. Third parties have EXISTED, sure, but their main effects have been bringing new ideas to the two major parties’ discussions, and splitting the vote between more moderate and more extreme members of the same “side” of political thought
I was going to mention that. It doesn’t has to be: winner takes all. What if candidates get power in proportion to their vote? They can rule by consensus. The real problem I see with Democracy is weakness against other powers ruled by Autocracy.
@@atscub you are on the money! I vote for A because I like rules from A and no one else, likewise for people who voted B etc etc OH WAIT that's called a free market, lets skip voting and just pay for goods/services from people who are servicing us the way we like
@@shawnradkeExcept then you immediately deny the principle of democracy by ceding power to people with the most money. Not to mention the free market fails at providing ideal choices all the time.
Except that doesn't work when you're deciding a LEADER, you f+cking m+ron. Also, proportional representation sucks, and coalitions are a terrible idea.
@@bminecreeperyeah and also monopolies exist, they will just bankrupt anyone trying to offer a service so they are the only one that people can purchase from
This is an older video yet perennially relevant. This discussion has come up in 2024 within my company's chat with some folks in the US, and others in Europe, and some on the US west coast getting to use Ranked Choice. I shared this video in the discussion when I argued that first-past-the-post is rational if not optimal, and forces coalition choices to internal within the 2-party choices, and that the 2 US parties change coalitions internally over time.
The Bohr-Curie-Einstein example is odd. The example assumes Bohrs poor performance leads to votes for Einstein and not Curie, despite Curie being much more closely aligned politically to Bohrs voters. Obviously this cooouuuuld happen, but its very unlikely; far more likely Bohr's loss is Curie's gain. And ultimately, the voters still expressed their preferences, and that person was elected. Preferential voting has flaws (too many candidates for voters to reasonably assess for one, like that Minnesota example; also potentially causing a candidate that almost no one liked to become elected due to weird preference flows), but tbh I'm not sure this example demonstrates them super well.
Yes, I felt that this example was contrived and unconvincing. If that's the best argument against Instant Runoff, it looks like a resounding endorsement instead of a criticism. Can't be worse than FPTP.
Arrow's impossibility theorem shows weird things CAN happen, but not how likely it is to happen. The condorcet loop for example, is pretty much impossible in practice and that's presented as the methods big downside
Odd as it sounds, such a scenario is easy to write (based on US politics). Imagine a third party, the Liberals, who united the disaffected into a 20% voting block. We'll also say that everyone kinda likes the Liberal, while everyone else dislikes the Democrat and Republican candidate. Thus, the candidate with 20% who will drop out handily, is the one everyone was okay with. And the winner will be despised by 60%. Step 2: now the Democrat does a solid (if dirty) attack on the liberal and takes 5% of the vote from him (D45, R40, L15). Looks like he's golden, right? Well, no. The Liberals now all switch their second choice to Republican. Which means that this 5% gain ensures a 45-55 loss against the Republican. Step 3: The LIberal now points this out, and tells blue voters that their party has *zero* percent chance of winning. However, if they vote Liberal, they'll win. Not everyone is convinced, but he gets his 5% back, and an additional 10%. So now the polls say R40, D30, L30. That is to say, the very fact that he lost votes, is the reason he now draws even. So now consider: it's a dead heat between L and D. If the Liberal scrapes ahead, he'll gain all the D-votes and the Republican will lose the final round in a 40-60 landslide. But if the Democrat picks up a few votes, then the Liberal gets eliminated, and even if the Democrat gets his 10% back, the Republican will still beat him in a 60-40 landslide. A few votes, therefore, will decide whether the Republican wins in a landslide (against D) or loses in a landslide (against L). Finally then consider their optimal strategies. The Democrat may be wisest to just drop out. He can't win, his best chance is to back the liberal and ask for some concessions. The LIberal's best option may be to demoralize Democrat voters who refuse to switch. If a few percent don't show up... sure, he'll lose those votes in the final round, but he'll make the final round in the first place. And the Republican's situation is the strangest: ideally, he would give some of his own votes to the Democrat, so that the Democrat passes the first round, only to be smashed in the final round. The GOP may use a well-selected faux-pas that only the small portion of his base that picks D second will switch for. If it's close, GOP operatives may even *vote democrat* to make sure the GOP wins. I think all of that is correct. And I think you'll agree, this is a realistic scenario. Yet it has several strange results in it.
I came to comment on this too. I felt this wasn’t a very good example of a problem with ranked choice voting. In the instant runoff example at least everyone who voted was able to have their vote counted toward the eventual winner. Can weird things happen? Sure, but at least your vote wasn’t completely invalidated & affected the final outcome.
@@valentetorrez3398 I found some stuff confusing, but the overall point is that there are various things you'd want in an election (particularly where one guy gets the job)... and you can't get all of it. So you're gonna have to decide which needs are most important.
15:45 I don’t understand? By changing position of A and C such that all A are below C, aren’t you either not changing A at all or making all A worse than B by a larger margin or narrowing the margin such that A is now only slightly preferable to B? To me, the clear winner is C; they have first choice on half and never last choice. Meanwhile B and A both have last choice for half with A never having first choice. C > B > A. Because half say B>>A. And half say A>B. Half say C>>B and half say B>C. Please somebody explain differently?
It’d work as you said if you were using a Broda-like ranked system (point based - where the “distance” matters). They are using the Condorcet method where you only look at the head-to-head matchups. I agree here it’d feel more fair and natural to use the former here.
Thank you, I'm not insane. The example doesn't work. He arbitrarily puts the final ranking as A-B-C, but then the voting on the left shows a pattern that would only allow for C-B-A. And if you change the final ranking to that, then there is no contradiction.
7:32 If Bohr becomes more unpopular why would his voters switch to Einstein and not Curie who is closer to Bohr ideologically but with none of the baggage of his controversial comments? In that case it would be Einstein who gets eliminated in the 1st round instead of Curie.
One example of this could be in the recent UK election. You have Labour, Conservatives and Reform. Some reform voters would be more inclined to vote Labour than Conservative because they were unhappy with the Conservatives who were in government at the time.
"The political spectrum is a U, the ends almost meet" - quote by some guy, I forgot who. I'm from Germany and especially in the easters states, voters frequently switch between the far left and far right. Those two parties share the "anti-establishment, blame [insert part of society], anti-institutionalist, propagandistic views". One just blames the rich, the other blames the immigrants. So, it's definitely not impossible.
The movement of votes is unimportant. The point is that if the distribution is 25-30-45 curie would win, but if the distribution was 31-30-39 Bohr would win, even though Curie got the same amount of votes.
I find it interesting that every problem presented relies either on the premise that the voters need to elect only one winner or on the idea that a candidate who wasn't anyone's first choice but was everyone's second choice isn't a good compromise. These things are just presented as facts without establishing why this should be necessary.
I hoped that in the end, Michael reflect how this theory could really represent real elections struggles. And how power dynamics fumbles with the math principles in the "fairness"
It also makes a huge assumption that “democracy = elections with voting” and excludes any other form of collective decision making, of which there have been many throughout history (like lots of indigenous societies in the Americas, or the Catalonian Anarcho-Syndicalists who used systems of mandated recallable delegates, or a variety of activist groups who use some form of modified consensus, etc)
@@shredded_lettuce The drawback of these other forms of democracy is that they even further empower the masses and disempower the individual. Democracy was a bad word until the rise of mass society in the 19th century. What we need to work toward is a world without "the masses." In mass democratic society there will always be a tyranny that manipulates social reality to guarantee its hold on power. People who live and think locally are free to organize themselves however they please. Most of the time, it turns out this is some combination of monarchy and aristocracy. Everyone in such a local world exists in an organic matrix of direct obligations and privileges, duties and benefits. It's how families have always operated. I would dare call it the natural mode of human organization.
@@shredded_lettucehe's either clueless, or more likely, pretending to be so he can get views by making people outraged that democracy is actually "mathematically impossible" when it isn’t. It's perhaps his weakest video yet, certainly the most repugnant.
@@shredded_lettuce if it was used by indigenous americans, its probably worthless then. Maybe a defeated and crushed people arent the one to take advice from.
"the pivotal voter is a complete dictator" NO! This is such faulty logic. They're only the pivotal voter because of the votes of the rest of society. It's literally an exercise in teamwork, they could not have done it alone.
I think he means that when a lot of people vote for group X and a lot of people vote on group Y, group Z could still win( by less votes). That means that votes cast on less popular candidate would be more powerful.
The problem runs even deeper than that. The pivotal voter is not any kind of dictator because they do not "dictate" anything. Regardless of what personality is appointed to the post, it is the ruling class as a whole that dictates how the system itself is run. Individuals are not that important. In fact, it is that system that produces individuals who take up particular posts.
What I understood was that the pivotal voter becomes a dictator because they rearrange their vote (in the example at 17:19, moving A above B). If any one person with B above C had done the same, then assuming all over votes don't change, they would be the "dictator" instead. ("Dictator" is a bit of a confusing word choice, since it is a different definition of dictator than usual. They aren't being called a fascist, rather that their vote becomes more relevant.) AND if this was the end of it, it wouldn't be an issue. The majority of people put A above B, and either prefer A to C or have no preference. The issue is that now the transitive property and the independence of irrelevant alternatives cancel out, as when we start having voters preference C to A, suddenly it creates a loop. It's all a bit obtuse, and I think the three-voter example (pizza, burgers, or sushi) was easier to follow. But honestly, all this just comes down to: "What do you do in a tie?"
My favourite voting system is STAR voting which stands for Score Then Automatic Run-off. Basically, voters score each candidate from 0-5. Then the scores across all ballots are tallied. The two candidates with the highest scores then move forward into a second round of counting where each ballot is treated as a single vote for one of the two candidates depending on which one the voter gave a higher score. The system is practically guaranteed to produce a winner with majority preference and at the same time allow all voters to express their honest preferences without a spoiler effect. You might think that voters will just bullet vote (scoring 5 for their favourite candidate and 0 for everyone else) but this is actually a really risky strategy as it means if your candidate doesn't get to the run-off, you're basically saying you don't have a preference over anyone else. So, it incentivises honest voting.
Everyone will just rank their favourite candidate full marks and everyone else zero. It's the best strategy. It will just default back to first past the post.
I'm not a big fan of STAR because it implies that a simple head to head preference actually reflects society's best candidate. If one candidate gets 4/5 from everyone because they're reasonably moderate and agreeable, but then another gets 60 votes 5/5 and 40 votes 1/5, because they're on an extreme that is slightly favored by the group, the extremist wins. And the all around agreeable candidate loses. Which I don't think is great.
Say for example both Einstein and Bohr developed a stance where they inadvertently proposes the same idea of universal basic income. But they name it differently and the source of the idea is politically different. Bohr then proposes that the income should come from eliminating veterans' pension. Some Bohr voters would look at every candidates and found similarities in Einstein's economic policies. These are the voters that will switch. Also, some people really dislike being called a fence-sitter, or a centrist. Those also would be the ones switching.
All this only applies to democracies that vote for 1 candidate or 1 party to rule, but the whole problem ceaces to exist, with proportional representation. Would've been nice if you mentioned that.
Yeah there are so many different forms of Democracy. I think the title should say be "Why First past post and preferential democracies are mathematically impossible". Would've loved for him to go over sortition.
Its like you didn't watch the video. Anyways, the problem is the assumption of unqualified franchise, compounded by the fact that politicians can now import new constituents.
@@churblefurbles Why? He is right? You could argue that these issues happen in negotiations inside coalition forming to a lesser degree but it is a whole new system to investigate with potentially (dis)advantages.
This is the second time this week I've seen a video on this exact topic with a Grey reference. This is because he did a video about this exactly years ago
The most important thing is to minimise gerrymandering and for all voices of society to be represented, without favour for one over another, rather than aiming to be mathematically equal
The funny thing about voting is that most of the time we don't get to pick the options we're voting on, rather we're delivered a curated selection of options and are left to choose the one we find most palatable.
Then join the body that curates the options? For example, here in the UK, join your local party. Then you’re one of the ones picking the candidates. In fact, since we have a lot of safe seats, joining the local branch of the party that controls your seat effectively makes you one of the few dozen or so people in that constituency whose vote counts for anything. I’m sure there are similar possibilities in the US and other places.
"The funny thing about voting is that most of the time we don't get to pick the options we're voting on, rather we're delivered a curated selection of options and are left to choose the one we find most palatable." and the ones who arent curated are smeared endlessly by the establishment and the media, so to th extent that they often become the most hated people in the country
I fail to see the paradox of perfectly balanced preferences not leading to a clear winner in an election. I didn't even use the same method i just noticed that one party liked each food most, second most and least. How is ANY fair voting system to have a clear winner there? Such occasions should lead to coalition government
Maybe this comment has been made and I can be directed to it, but I'm a new teacher (grade 7s) and we're learning about government in BC, Canada- DEFINITELY going to show my class this but just seeing if any teachers have already developed a lesson plan around it? Thanks Veritasium!
British Columbia tried to have a referendum for preferential voting, but they didn't explain it well AT ALL so most people just voted for "no change" rather than be confused by what might happen if we changed. I really wish they'd redo that vote but explain things better. Every single person I've talked to hates the current first past the post system in Canada.
Just because first-past-the-post's spoiler effect is awful doesn't mean there aren't worse alternatives. As Kenneth Arrow proved, any kind of preferential voting system will have some degree of arbitrariness. Ranked choice voting looks great at first, but it has a ton of major, non-obvious problems. (Look up "Instant Runoff Voting: Looks Good--But Look Again" by Stephen H. Unger if you're interested.) And people being confused is a valid reason to not go with a preferential voting system; it makes voting tallies far more opaque to the public, which is a problem when people already have a low level of trust in their governments. Score voting systems (including approval voting) and STAR voting are much better alternatives (and they _actually_ remove the spoiler effect, unlike ranked choice voting, which only seems to at first).
I wouldn't trust any British Columbia voters to make a right choice on anything. Once there was a vote for keeping HST so everyone pay less tax or switch back to PST+GST, they voted for paying more tax.
@@RobinClaassen To be kind to Unger, he's analyzing Ranked Choice in the context of abstract statistical anomaly. It's what Veritasium was saying at 7:30 - that a wing party dropping votes to the other wing can get the moderate eliminated first. Fun for math. Ridiculous for humans. Voters would be difficult to shift like that and a candidate would be crazy to try. To be less kind to Unger, I also read some of his other articles. He sounds a lot like "The best system is the one that gets me what I want!" But I hope most Veritasium viewers would understand that this short-sightedness is not a good way to design your voting rules.
@@notme222 I do find it somewhat difficult to get a sense of how common some of the situations IUnger is talking about are likely to be (e.g. those situations in which all voters reversing their preference orders would result in the same candidate winning). I was more or less talking his word for it that they were plausible. In that post, he gives the impression of being unbiased. If you're saying that he gives the impression of having motivated reasoning in his other writings, that's a reason to question his assertions. That said, some of his examples are more clearly likely to be common occurrences. For example, in any situation in which there are at least 3 candidates with high levels of support (each having some chance of winning), there will always be some voters who will be incentivized to downgrade their ranking of their most preferred candidate because they'll want to make sure that the first candidate with a high level of support to be eliminated isn't ideologically opposed to them (such ballots would be likely to have an also-ideologically-opposed candidate as their second pick They'll instead want someone who they ideologically agree with to be eliminated first, and may therefore feel afraid to give their preferred candidate their first choice if they think that that might result in stopping that candidate from being eliminated first. It's essentially the same spoiler effect that we get with first-past-the-post systems. So it shares that problem with first-past-the-post, while also having many problems that first-past-the-post doesn't have. Including the results being sensitive to the level at which they;re tallied (e.g. county, state/province, or national), and those tally results being more opaque to the public. And given how much preference information voters are expressing on their ballots, it's frustrating how deaf to most of that information its vote-tallying algorithm is. It only reads one line at a time, ignoring the rest of the information, and never even reads most preference information expressed on most voters' ballots. When it does go past the first line of preferences, it arbitrarily only does so with some voters but not others. While you can express more preferences on a ranked choice voting ballot than a approval voting ballot, the approval voting algorithm is actually the one that will take more of your preferences into account.
@Franimus if you had 3 candidates A B and C, and then let the voters rank how good a candidate is overall on a 1-5 rating. The 2 candidates with the highest average go to a runoff, where everybody revotes for whatever candidate they want, whoever gets a higher approval wins
To be clear, the runoff's "revote" happens automatically because the ballots already show which of the two finalists each voter prefers, so voters only need to go to the polls once. For example if you had ranked A (5 stars) B (3 stars) and C (0 stars) and then it turned out that A and C were finalists, your vote would be given to A in the runoff. If B and C had been finalists your vote would go to B. It has the major advantage over classic RCV that it only ever requires 2 tabulation steps (the score and the run-off) and that both can be easily counted by hand. In theory it also eliminates the need for primaries, though it still works fine if you have them.
Here is a better title: "Why Winner-takes-it-all Democracy is Mathematically Impossible". The first theorem only applies if there is a single winner in every district. Having actual proportional representation does not have these flaws.
Not quite true. The underlying problems still exists even with proportional representation, if based on rankings. It's just less noticeable because the PR "simulates" cardinal information in aggregate.
Yup. Just have a runoff between the two most popular options. It works in most places. It leads to good outcomes most of the times. Usually the more centrist of the two wins. Problem solved. Or, like Sweden, you have party systems, so you vote for parties, that also works OK, but that would likely require constitutional changes jn the US. A two stage system just needs open primaries, on the same day, later in the cycle.
Approval voting solves literally everything discussed in this video, and requires virtually no change whatsoever in the way elections are done. Just "click or check" every person you like. Person with most clicks/checks wins. It's so simple and so obvious it is maddening we don't do it this way. In fact, a huge number of "spoiled ballots" are because someone checked more than one person, thus invalidating their ballot. A little disappointed Veratisium didn't spend more time explaining Approval Voting. How it requires virtually no adaptation to implement, and how it completely demolishes rank choice in every conceivable way. It has the finality of first past the post with the variance of options of ranked choice.
Lost me around 16:00. I don’t accept that one of the rules should be that an extremely divisive candidate that splits the vote between love and hate should either be completely on top or completely on the bottom. Intuitively, the middle IS where they belong. If our criteria insist they have to win or lose, our criteria are wrong. Electing them or dropping them would make a LOT of people unhappy.
You have a point. When at paradox every given option is equally good, in order to decide one have to refer to other set criteria outside of the parodoxical ones. If {paradox} then: other formula.
I've seen lots of comments about the Bohr voters switching to Einstein. The reason he gave of *why* people switched distracts from the point. The real problem being highlighted is that the system incentivizes "strategic voting". Suppose that in polls leading up to the election, the results are 25% - 30% - 45%. If we assuming that people are voting according to their views, Curie would win (as explained in the video). So, the Bohr voters could (in theory) organize to have 6% of them rank Einstein 1st, which would result in Bohr winning. Obviously it's not good when voters can deliberately down-rank their preferred candidate to make them win.
This all comes down to if his critique of the RCV is a purely mathematical one or does it have any real life value. For one, the example he gave is completely illogical, even in purely mathematical terms, it will never happen. Bohr voters will never switch to Einstein. I am disappointed that a scientist would ignore logic just trying to make a point to fit his thesis. The explanation you gave, while perfectly logical, are so difficult to achieve it in real life, it might be almost impossible. Even without considering polling errors, the amount of effort needed to ask a CORRECT portion of your voters to hold their nose and strategically vote for another candidate is so huge, it's not realistic, purely theoretical. This, and the pivotal voter paradox, plus the terribly misleading title, makes this video one of the worst video this channel has put out. I mean, I still love this channel, but he can and should do better.
@@zhenfengliu8032Yes, there's a logic error thwre. Assuming that most of the voters jumped from Bohr to Einstein but still more of Curie's 2nd choice would move to him rather than to Einstein suggests Bohr is the moderate candidate, not Curie. And Bohr wins as a moderate, as more people oppose Einstein.
This feels like a mathematical proof that suggests problems in edge, corner or point cases make an entire system unusable. I’ve never interpreted “impossible” to mean “it only works 99% of the time”.
The recent 2022 Alaska special election had Palin spoil the election for Begich. There's not even any need to speculate about what voters would have done, since it's all in the voting data. Begich was preferred to both Palin and Peltola but Palin stole too many hardliner votes away from him.
While it's perturbing that the current voting systems might be fundamentally flawed, I'm intrigued by the prospect of approval voting systems. I hope we soon see more real-world trials for the same.
This video only speaks about Winner-takes-them-all voting systems. Most democratic countries don't use them. If the voting distributes about 4 parties like this: A: 35% B: 30% C: 20% D: 15% that's a totally fine result. You can fill the seats of the parliament proportionally. Then A and B can work together or A and C or even B, C and D can work together if they can agree on the same laws. Many democracies work that way and it's not mathematically impossible. Besides that you could have even other ways of democracy. You could for example get rid of votes and instead pick random people for a parliament. Put people in groups of 6 and throw a w6 dice to choose one. When you have reduced the candidates to 1/6, do the same again with that group until you have enough people for the seats of your parliament.
Approval has a serious flaw. In the election between Ms Loved, Mrs Hated and Mr Indifferent, you obviously approve Loved and don't approve Hated; but if polls are close you don't know if you should approve Indifferent (which might beat Hated) or not (because he might beat Loved). This is called "later no harm" and can cause approval votes to behave as if they were FPTP. The real best solution is something the video doesn't even mention, proportional representation.
Primer does an excellent video showing the impact and downside different voting systems have. He uses blobs for simulations and while yes they are programmed simulations they attempt to show how things can play out and the different strategies that voters could "use"... use being in quotes because honestly strategy only works if you have a large enough group to agree to it. Also @3:48 nice CGP Grey ref!
Before I watch, is it Arrow's Impossibility Theorem and various apportionment paradoxes? EDIT: I was not disappointed - it was. Although I learned quite a few things about the history of the field. I had no idea that attempts to replace our silly plurality-rules format had gone so far back (and things similar to Arrow's had been noticed all the way back then). Moreover, I was greatly pleased to learn that Arrow's Impossibility is much less of a problem for IRV in actual practice. Of course, I wouldn't stop supporting ranked-choice voting/IRV just because of that (it's better than plurality and we can't let the perfect be the enemy of the good-enough). I too, support rated-choice voting over ranked-choice, but with two caveats: First, I think we should establish ranked-choice first simply because it's simpler and people have heard of it. It is good enough, after all. Second, there is one form of rated-choice which I can *not* stand. That's approval voting. The problem with it is, that you preference for all tolerable options is considered identical. You wouldn't be able to spoil your least objectionable option, but you could spoil your favorite option. In ranked-choice, you make very clear which one is your favorite, and that one will get a leg up. If ranked-choice was available in 2016, I could've cast a vote I was proud of without spoiling things for the lesser-of-two-evils we had then (I still did, because I don't live in a swing state and the popular vote doesn't matter, but I digress). I couldn't have done that if all options voted for are rated equal; I would have had no way of communicating just how awful I considered my least objectional option to be. For rated voting to work *well* the voter *must* be allowed to give a higher rating to some candidates than to others.
There is no single dimension that ranks all candidates, because by definition people will prefer different areas. Approval voting definitely makes more sense, since that means the most people agree on the candidate.
The problem with Approval voting is that it's actually just first past the post wearing a mustache. If we assume that A and B are the two "best" candidates, and everyone assumes one of the two will win, with B having a very small advantage. How should you vote if you like A more than B, but you approve of both? Or if you approve of neither? Either way you should say you approve of A and disapprove of B, because otherwise your vote is entirely worthless. Now lets say that while A and B are both similar, C is very different and is also a decent candidate. Unlikely to win, but likely to draw about 35% approval. If both A and B are perfectly matched and both vote selfishly against each other, C will win. That's the spoiler effect. On the other hand, if supporters of A or B choose to approve of the other to make sure C doesn't win, then whichever candidate's base is more "selfish" will win. In that case, the most fair thing to do would be to form a coalition between A and B, have an internal vote of their supporters to see whether they prefer A or B, and then have the loser drop out of the race so the winner can beat C for sure. And that's what we call party primaries in the US.
@@driftwisp2797 Comparing approval with FPTP is simply wrong, because the strategy in approval voting is never the same strategy in FPTP. There is never a reason not to approve your favorite candidate under approval, whereas this is common under FPTP. So any analysis or claims that approval falls back to FPTP is false, regardless of any convoluted scenario or logistics on top of it.
@@1ucasvb There is never a scenario in Approval where voting for your first choice is wrong, that's 100% true. Approval is "FPTP but it's always okay to vote for your first choice". There are, however, many scenarios in which "approving" of your second choice is wrong. In fact, in almost any scenario where your first choice and your second choice both have any realistic chance of winning, voting for your second choice too is hurting the chances of your first choice winning. If each person only casts a vote for their first choice, it has become first past the post. If people are casting multiple votes, it means they think only one person they're voting for has a reasonable chance of winning, so it's unlikely to ever have an impact on the election. When it comes down to it, the main difference between Approval Voting and First Past the Post, outside of "any convoluted scenario" is that after the election is over you have a better idea of how well the candidates who had no chance of winning actually did, and you can see whether the candidates were generally unpopular (people throwing away their vote by not approving of either, like they do in FPTP) or generally popular (people throwing away their vote by voting for both). Great for statisticians, not changing the election.
Approval is ok. But tbh I'm not sure it's better than ranked choice. The tick box version places a binary decision on the voter for each candidate, rather than a ranking. So individual voters have a hard time expressing who they REALLY want. The rated system with numbers (e.g. -10 to +10) solves that problem, but creates a complexity issue that is likely to be politically gamed. How can a voter know how much approval to give someone? What if their +5 is enough to put that candidate over their +10 choice? It's not super complex, but it's enough that politicians will go with (+10 for me, -10 for everyone else). Which becomes problematic not mathematically, but electorally.
Checking off/ giving % to any and all candidates you approve of is cool theoretically. But many people will realize they are seriously kneecapping their major picks for each minor pick included. Eventually we're back to 2 party system. The ranked voting in your example was polarized to the extreme, but candidates would realize the best way to succeed is to be accepted by most rather than polarizing
Eh, CGP Grey has a mix of great videos (hexagons are bestagons is pretty convincing if unimportant, spaceship you was incredible in early Covid lockdown, and his various explanations of voting systems are great primers) with some pretty glaringly bad takes (the British royal family is actually cool and good, self-driving cars are the solution to traffic).
Thank you for sharing this video. The majority (59% in a recent poll) have never even heard of proportional voting systems, and the benefits (although not perfect) that they can provide. I highly encourage everyone that believes we can improve our democracy, to share the ideas of proportional voting to their family and peers.
What benefits? They are just empty claims when in reality it works the same no matter what you do. You end up with 2 opposing groups, the only difference is you are huffing your own farts thinking that you are brighter than others because you have your head so far up your own butt you aren't able to observe reality.
Funny, but the point of the video was kind of the opposite. It's math - not morals or any real division - that prevents us from having good representation. But it doesn't have to be.
My question for Rated Voting systems would be the impact of the “integrity” of the negative ratings. If a bunch of people voted 10 for their person and -10 for every opponent, but the opposition didn’t go to FULL -10 for their less-preferred ratings, would it encourage people to just approach the system and vote 10’s and negative 10’s? If the people who disagree with you are giving your person -10, wouldn’t a lot of people just try to game the math by only rating with the extremes? Is that accounted for in some way?
There are plenty of other attempts to remedy (though not solve) this problem. For example the winning party could be required to get a true majority (51%). If they receive less than 51% of the vote, they have to form a coalition. They must pick at least one other party to cross the 51% threshold. As a side effect this also incentivizes them to maintain a semi-decent relationship with competing parties instead of going completely off the rails during campaigning. This system has held up very convincingly in the real world. While it isn't a perfect solution, it's a very good improvement.
You describe a parliamentary system which typically has many parties. The United States almost always has only two parties of any consequence, voting for a 3rd party is basically a wasted vote BUT it will bleed off votes for one of the two remaining parties. In practice this turns into a SINGLE PARTY having actual control of a particular branch of government.
Yes, that's actually how real democracies work. Systems like the one of the USA are fundamentally broken. What you described is completely normal for most countries. If you have e. g. A: 35% B: 30% C: 20% D: 15% as an voting result, A and C can just work together. There is no need at all for a single winner. Therefore the mathematically problem from the video completely vanish. It's not democracy that is impossible. It's the Winner takes it all system that the USA uses.
@@Duconi "What you described is completely normal for most countries." As you have noticed, the United States is exceptional, that is to say, not like "most countries". "A and C can just work together." In which case they might as well be the same party. This is what happens in the United States. It is a natural outgrowth of "democracy" and eventually it will be just "A" a single party rule.
@@thomasmaughan4798 A and C can have different positions also maybe over time a new party E comes up that fits better to A and gets popular. With voting you can therefore also define priorities. One party maybe has environmental protection as focus another on wants to improve the health system. They can work together but voters can vote for what's most important to them. If you just have 2 parties you don't have a choice to prioritise. You can just pick one of the 2 parties and take whatever they offer. It's not possible to pick a party with different solutions or that wants more than one of the 2 parties offer.
@@Duconi " If you just have 2 parties you don't have a choice to prioritise." Many parties exist in the United States. There's typically 8 or so choices for President of the United States. But since there can be only one President; any vote for anyone other than the winner is basically wasted. This is true for Congress as well. Having someone NOT a Democrat or Republican is rare. It is very expensive for someone of other parties, Libertarian for instance, to run for office. What tends to actually happen is that rather than create a new party, an existing party is evolved or adapted to new priorities. Democrats have been captured by the "far left" and in 2016 the Republicans were captured by.. well I don't remember exactly, almost libertarian. Each party is thus *already* a coalition of sometimes very different priorities. Typically there's only one or two top priorities; for Democrats it is "woke" and abortion; for Republicans it is "liberty" and first and second amendments. This can also be viewed as "herd" versus "not-herd".
For those of you who are curious, the "post" is a metaphor from horse racing. The finish line in a horse race is an actual post in the ground. It goes back centuries. Originally there were two posts and the race was "start at post 1" then "round post 2" finally "end at post 1". Once we invented ovoid tracks with gates to hold the horses, there was only the one "post" remaining.
CGP Gray covered this quite a while ago from the perspective of voting and how to resolve it. It's long been one of my favorite mini series. This is a fascinating continuation to me!
As a math PhD student I will clear up everyone’s confusion. So a lot of people are confused about the pivotal voter being a dictator, which even I was prepared to go on a rant about how the non-dictator rule was a terribly chosen axiom. I feel like this could have been explained a little better, but here I am to explain it for you guys. Upon further inspection, the reason the pivotal voter creates a contradiction is not because of it being the median voter, but actually because if you notice that it creates a loop again with B > C > A > B in that case, which is actually a transitivity issue. The non-dictatorship rule is actually just a way to accompany multiple rules into one. The actual theorem states that for any transitive and complete preorder that the following three things cannot happen at once: Pareto Efficiency (if a is preferred to b for everyone, then a is preferred to b overall), Non-Dictatorship (for any overall ordering there does not exist a single voter which if that voter prefers a to b then a is preferred in the overall ordering), and Independence of Irrelevant Alternatives (if all voters prefer a to b, then the preference of c should not change the overall preference of a over b). The issue is not that the voter is the median voter or the tiebreaker, it is that his vote has a theoretical potential to cause a Condorcet loop. Therefore, in practice it can be very useful, but can become detrimental if the specific conditions are met, which would throw the democracy into collapse (unless there is a secondary voting method in place to counteract the Condorcet loop, which in that case, we should have used that method instead in the first place). If you guys want me to make a UA-cam video explaining the issue in finer detail please let me know.
You just confused me even more. Pretend you're talking to an AVERAGE voter who doesn't understand math. Use names, not letters. Too many (non-colored) letters. Basically, what math is telling us is the idiot undecided voters who don't even know what their own preferences are decide our elections? I knew that already, lol.
07:52 Idk why this is a problem. After all, you select your preferential candidate. While Bohr might have given an unpopular speech/policy proposal, he is still the most popular of the three. And if people are voting Einstein from Bohr, they think Curie is bad/incompetent. Voting comes down to population's favourite candidate at the end. The second choice exists to say who people prefer anyway. So it's good all around. I dont see the problem. But I agree with the commentary about how politicians try not to stand out too much so as to win 2nd or 3rd choice preference.
I was Latif’s Uber driver a couple of months ago and we had a nice chat and he told me all about Radiolab. So cool to see him in this video! “Is that my guy? That’s my guy!” Was my exact response 😂
When it was inherited seats, most were "cross-bench"; they stood with more than one party depending on what stance had. Now that's it's nearly all "life-peers" it's a joke. They get their life peerage from being part of the "inner circle" of the governing political party, or by being a large donor; it's full on corruption. We've gone from a review house that did the best for the long term (usually anyway) because their title and family line of that title depend on it, to one where they only care about their short term profiting (either monetarily or socially).
Transitivity seems to rely on an assumption that comparison is made based on any one dimension. For example like number, if number a > b and b > c, then surely a > c But political preference doesn’t not work like that. It is more like an aggregation of comparisons along multiple dimensions. Therefore preferring a over b and b over c does not necessarily mean that a is preferred over c (given that there are many dimensions to consider )
Most would agree if your preference is not transitive, you are not rational. While people do behave like that, it is easy to prove that with cycles (i.e. without transitivity), impossibility will arise in some sort.
I have reservations about the Swiss results. I think it makes difficult decisions nearly impossible. Just this year they voted for adding 13th pension to the year, despite the government's recommendation to vote against it.
@@BakhtiyarNeymanyou either want democracy or competent governance. Those two aspects rarely overlap. A good democracy fundamentally necessitates a weak executive. A weak executive will never get any thing done in terms of policy implementation.
@@BakhtiyarNeyman It's what the people want and therefore democracy. I would agree that democracy isn't perfect as people are incompetent, but then you would first have to come up with a new system that is fairer than democracy, which doesn't exist (yet).
But it becomes disingenuous. For example, I would vote -10 for the candidate i don't like and 10 for the one I do like, even if I'm really choosing my candidate begrudgingly. It's overly complicated and doesn't really add information more than a ranked choice would. Just my opinion.
The main problem in representative democracy is that the representatives start to have their own distint interests that are often at odds with the people they are supposed to represent. The main barrier to improving any voting system is that the people who could make the change are necessarily the people who have just won under the existing system
A proper voting system pretty much devastates the incumbent advantage. The method I've recommended uses a 7- or 9-winner nonpartisan primary for a reason: less than 15% of the voters can throw you out of office. With 7 candidates equally spaced across the voters on the single-dimensional ideal axis, if 15% of the voters on one "side" move the hypothetical winner down on their ballot, the next most similar candidate wins. They can't tip the election from Trump to Bernie, but they can tip the election from e.g. Biden to Obama.
@@johnmoser3594 The solution is much simpler. You have a parliament with e. g. 100 seats. Just distribute the seats matching to the votes. Meaning, if the people voted for A: 35% B: 30% C: 20% D: 15% the get seats as following: A: 35 seats, B: 30 seats, C: 20 seats, D: 15 seats. Problem solved. So now this group of 100 people is discussing new laws. If for example Party A and C want the new law, they together will have 55% of seats, so the law passes (except it's a law that requires a 2/3 or 3/4 majority). There is no need for a winner-takes-it-all system.
@@Duconi Your approach leaves several questions. What do you mean "seats matching to the votes"? Do you mean by party, where we vote for parties in which the elite aristocracy decides who will represent us and what the policies will be? That's strongly disconnected from the voters. Have a look at the US Democratic party, with people like Bernie Sanders, Ocasio-Cortez, and Joe Manchin, on varying parts of the political spectrum; the Democratic Party would exclusively choose people like Joe Manchin, internally. This kind of system gave rise to the Nazi party because it moves almost all democratic power out of the hands of the voters. Do you mean by single transferrable vote with 100 winners? That's viable, but it's a complex process. Do you have geographical concerns? If we're going with a blanket STV election across the entire nation, it's technically possible for all of the elected politicians in the US to come from Baltimore City. How do you deal with single-winner elections such as for governor? You've only got one winner. Runoffs are completely broken. The process I described addresses each part of this by making the elected candidates reflect voters instead of party-elite-chosen aristocracy, both with multi-winner elections and single-winner elections; and by giving some degree of geographical control, with the electoral process completely eliminating gerrymandering (tweaking the district lines is no longer a free action: to obtain, something of equal value must be given up in exchange, so you can manufacture a moderate liberal district to get rid of a far left-wing district but only if you replace a strongly conservative district with a moderately conservative district in exchange; and that only works for single-winner elections, otherwise you're basically just moving the political ideals between districts rather than blending them to a smooth and creamy consistency of uniform ideology).
True! People react to what is in their faces right now - bills to pay etc. They don't realize by actually holding politicians accountable that they may get a fairer shake. In their desperation they turned to Trump who wasn't a politician, but they apparently had never heard of a POS 'business man' before.
Great video which shows, that a mathematical approach to look at a problem, which concerns society, manages to almost miss every point, that matters. Like the undemocratic influence, which big coporations have over politics. And the fact that those companies are arguably more important then our governments and are organized in an completly undemocratic way. Or that many people think democracy just means the majority or their representatives decide, even though respecting human rights are just as important.
Ramon Llull being pronounced Rayman Lull, too... the closest approximation to English phonetics and phonotactics is something like: ruh-Moan Yoo-ee (GA IPA: /rʌ'mown 'juwi/)
not always, enough real life examples. Also not about the specifics, but mostly about having 25, 30, 45 changing to 31, 30, 39 and what could happen. A lot of people are angry, and will vote for ultra right or ultra left and even switch between them on a whim.
There's an argument that the political spectrum is not a line, but a circle - that far left and far right are closer to each other than they are to the moderates in the middle.
Not necessarily. They say "the political spectrum is U, not a straight line. The ends almost meet". Many people in eastern Germany vote either far left or far right, nothing in between. The common denominator that's relevant for them? Anti-establishment.
Yeah that was a poor example and poor explanation of a real effect. If we imagine three candidates A, B, and C in a spectrum where A and C all preference B and B is 50/50 split, if you're a C voter you _can_ increase the chance A wins if you don't vote strategically for B. If the votes fall 40/25/35 for A/B/C respectively, B gets eliminated and A wins. If C voters are strategic and change the result to 40/35/25 now C gets eliminated and B wins. A better result for C voters by ranking their favourite candidate lower, failing the no favourite betrayal criterion. In the real world this seems to be surprisingly rare because true centrist parties are rare. You would more often find B voters split their votes 90/10 one way or the other. And even the C voters might be split 90/10.
@@TheWhitePianoKeyProductions Why would it be an issue for 31, 30, 39 to result in a Bohr win? If the preference of the public is to go extreme over center, then extreme should logically win. It'd mean the public generally believes Bohr > Einstein > Curie.
15:00 I have watched this example a couple of times and I still don't fully understand it. It states that if everyone ranks a particular candidate either first or last, society as a whole must also rank that candidate first or last. But as far as I understand the condition of transitivity was never broken, since the group as a whole didn't agree initially nor in the altered vote that either A>B or B>A (and C>B or B>C by extension). What the altered vote changes is that there is an unanimous agreement that C is preferred over A (C>A). Changing A>C to C>A doesn't break the condition of transitivity and therefore the change in result from A>B>C to C>B>A meets all five conditions. If there's something incorrect about my thought process here I'd really like to know.
I agree, the goal of the system is to elect one winner not to sort out the exact order of preferences among the remaining choices. from the example he gives it is clear choice c should be the winner, the exact order of b versus a is interesting academically but not really relevant operationally.
It's condition 5 that is broken. B is the irrelevant alternative. Take out B, and C wins by unanimity. Introduce B and suddenly A wins by the rules of transitivity.
@@OneLine122 I don't see your point there. Why should A win? It seemed to me that the A>B>C was just a thesis, which in my opinion is wrong. If you make the thesis C>B>A there is no problem. If you take B out, its C>A ( b>"C>A" on one side and "C>A">b on the other). That means it doesn't matter whether B is in the vote or not.
@@OneLine122 My point it that the condition of transitivity is never broken, since, as I mentioned before, the group as a whole didn't agree initially nor in the altered vote that either A>B or B>A (and C>B or B>C by extension).
Yeah except CGP Grey convinced everyone (including myself) to be an advocate for Ranked Choice Voting...which is a trash system. This video is at least an attempt at undoing that damage. Ordinal systems are trash. Cardinal systems are good.
My city voted against STAR voting (a Rated Voting system) because the people who were lobbying against it were the ones in positions of political power--and they benefit from first-past-the-post voting. They knew that if we used STAR voting, their extreme views would be likely supplanted by more moderate candidates. 😢
STAAAAAAARS
@@kentslocum Class struggle continues
RAAAAATS
More moderate candidates is good isn't it? Why go extreme?
@@Irondragon1945 resident evil nemesis??
Good video, now let’s watch it
😂
Another banger by riot kassadin
🤣
Not if I watch it first
@@gustavosedano294 "The balance of power must be reserved" - Kassadin F. Kennedy
The pivotal voter thing doesn’t make sense, because it’s not actually an individual person. You shuffle up the votes, and one random ballot happens to be the pivotal voter. No one knows who it is ahead of time, and even more important, that vote being pivotal relies on all other voters having made a specific decision. Had they voted different, that would no longer be the pivotal vote.
You don’t need one when it’s rigged…. Biden should be president for another 4 years. Long live King Joe.
Yeah, the "dictator" ends up being a random person, who doesn't even know they're the dictator for that election.
It's one of those math things that doesn't translate to real life as well. Having one determining vote is how voting works sometimes, that's why you always use an odd number of voters, so in the event of a tie, one vote will break that tie.
Why do you want to prove that dictator Trump is a way to go? Russians money again?
In this field it's hard to make the explanations make sense intuitive, because the conclusions are sometimes not but also the terminology is confusing. Unfortunately Arrow's theorem does mean that it's not just a pivotal voter who is unknown before the election. It means that the rule of the election is "look at how this person voted, and say that is how society votes". Unless it is such a rule (the dictator can be selected randomly, but it's not just a coincidence, it is a pre determined person outside of the vote counting process) you cannot have both unanimity and independence of irrelevant alternatives. It's a looser rule in special cases, but no generalized ranked system passes.
It does make sense, though it is difficult to highlight constructively. First you show that any 'dictatorship' (i.e., a voting system with a pivotal voter) satisfies the other 4 conditions. Then, you assume that a voting system satisfies all rules except potentially the non-dictatorship rule. You can then show rigorously (with the type of argument that was presented in the video) that there MUST be a pivotal voter. You can't know which person it is because you are making a general argument for all voting systems. But you did show that every such voting system must have a pivotal voter.
love the CGP Grey references, he introduced voting systems to my mind sphere. Thanks Derek for further expanding on the topic, would love to see different scenarios for approval voting and +/-10 points voting
STAR voting is a nice system of rated voting
But one man alone is going to destroy it
or, just vote for any candidate you want. +10 and -10 adds complexity which isn't needed for approval voting
18:40 that pivotal voter isn't really a dictator though, because he requires all the other voters to set up the scenario in which he would become, unknowingly, the pivotal voter. That pivotal voter would never know or be able to abuse their dictatorial status in practice?
The grain of sand that tips the scale isn't a dictator. Edit: I understand that this question is largely philosophical and has little bearing on a scientific debate/discussion but I felt it's worth examining for the sake of the larger conversation about where autonomy of the voter ends.
They become, mathematically, the dictator that changed the outcome for everyone.
Doesn't matter if there's only 2 choices, you'd never know who the pivotal vote would be, but with 3 or more choices you can construct scenarios where you can find the single ranking that influences the outcome (which is what is shown in the video, the position doesn't matter).
Doesn't mean it has to happen every time, heck the more voters there are, the less likely it will be. But if you can mathematically show the case can exists, the voting system isn't "perfect" enough to adhere to that condition.
@yak-i4c this is exactly what I was thinking while watching. Ignoring the definition of "dictator" to prove that one exists.
I hope this comment gets more traction. Can't ignore the definition of dictator in order to prove that a dictator exists.
100% agree. One can only be a dictator with the foreknowledge that their vote is pivotal. If you can't leverage your "power", you have none. You're best off just voting your heart.
The pivotal voter only becomes a thing if there's discussion and negotiation where everyone knows how many votes they have (like what happens in Congress) versus just having an election day where everyone goes in and votes and nobody has any idea what the vote count is. Even in the first case, that only happens if only 1 voter is willing to negotiate and the others are immovable. But that's more a testament to how negotiating in politics can get you favors.
I absolutely loathe the first-past-the-post voting. My entire life we've ended up with a system where everyone votes AGAINST the candidate they dislike the most rather than voting FOR the candidate they actually support the most. The amount of disinformation that has been spread when we've tried referendums to move to proportional representation or alternatives is infuriating.
why is that a problem? Sometimes that's the choice you face as a country - it says more about the political climate in your country than your voting system. And arguably getting rid of the guy you don't want to be in charge is the single most important requirement of a voting system.
Exactly, Biden should be president for another 4 years.
Like this comment if you want a man in office
US or UK?
you can only have one unified entity that is the administration. it can never be representative. so called representative systems are not representative at all, since there is coalition building that is entirely out of control of the voters.
The worst problem with FPTP is that the current two parties will forever block changing it.
yep it's quite contradictory. how do you vote for a better voting system under a bad voting system?
Well in Australia the 2 party preferred system is worser that FPTP, the current dude only got 30% of the votes and got in.
Exactly this.
Support the Equal Vote Coalition and help us get past this challenge! Ballot initiatives, city council ordinances and pressure from citizens can make it happen.
The good thing is not all politicians staunchly align with their party. City council members understand the issues with FPTP and some are open to change.
@@rasta77-x7o You mean 30% of the FIRST PREFERENCE votes. Preferential voting will always be objectively better because it considers multiple candidates for individual votes. If you believe in a democratic society then you should not be rooting for fptp...
That approval voting actually excites me. When you started explaining it I started seeing a light at the end of the tunnel. That's pretty cool I would totally participate in that type of election
exactly!
In Europe, with proportional voting, parties form coalitions after elections, because it's very rare for one party to win a majority of districts. In the US, coalitions are formed IN parties. That's an inevitable results of 'winner takes all' elections, and the fact that it's impossible to get a majority to agree on every significant issue.
@@caldodgethis is a total myth. Even duverger himself noted that countries which use a top two runoff tend to escape duopoly. There's absolutely zero reason to expect single winner districts to produce a duopoly. What produces a duopoly is the plurality voting method.
Plus the person put in charge is not voted for directly, but the coalition selects them. They can't go against the will of the government or people, and the government reflects the mix of the people in the country much better, even if the voting system would be slightly off.
I was looking for the words but you summed up my thoughts perfectly.
I think the American perspective on democracy is culturally different from some of the Europe and prone to misunderstandings. It's complicated to explain every system and would probably take a lot of videos to cover I believe.
Who needs districts if your country is x times smaller than the average US state?😅🥳 Still, election results can go against the will of the majority, albeit partially. Forming a coalition can mean that the least liked party will be in government for example and the second most liked is not.
Speaking from experience, having no say in who becomes prime minister or who takes a seat in the senate can be a bit of a bummer as well.
Looking at our own version of MAGA-ism, even though our guy has been bothering our parliament for far longer than Trump, our version of democracy is definitely far from perfect at this moment. But that has more to do with the way people "do" politics these days and the ways people (don't (care to)) inform themselves i guess.
Now, can you guess where i'm from?😅
This. Everyone's vote is heard proportional to the number of people who voted for the party. It's only winner take all that is stupid.
We don't need perfection, we need improvement. An outcome more reflective of the will people will be better than the one that is less reflective.
In a time when voter confidence in our election system is at an all-time low, the title of this video is very irresponsible. "It's crooked, but it's the only game in town?" Why not just regurgitate everything Fox News says? 🙄 You're the Ralph Nader of this election.
at the very least I don't want my ballot to essentially be a true or false between two parties.
Exactly this.
All problems with better way to vote are like "But imagine there is only 3 people voting, and they exactly line up with this or that!"
It never happens IRL. And in the so very unlikely even that it does, just redo the vote, people will change their mind. Heck, maybe even remove the least favorite one.
But, yeah, that's the problem with maths, they forget the real world application.
Country population are counted in millions, one person won't change a thing because such a huge number of people can't think the same thing.
And voting for one candidate only is really bad.
Even more if you have, like here in France, a two turn based election!
It's awful really.
@@williamyoung9401 definetly
@@williamyoung9401 You leave MSNBC out of this!
Love the CGP Grey Cameo
"Hexagons are the bestagons"
More like CCP Gay
@@mphRagnarokskibidi toilet
@@mphRagnarokyou’re not sigma lil bro
Time stamp?
6:20 please don't play the audio oh no you did
I loved it 🥰 - now I want to visit Minnesota ❣️
I know right?
3:39 I love that CGP Grey reference, perfectly fits with his voting videos. Great video!
Hexagons are perfectagons. :)
What's Mathematically Impossible is to have a winner-takes-all voting system that's representative of the electorate. If you want to represent the electorate accurately you have to let go of the winner-takes-all system and adopt a representative democracy where parties are awarded seats in parliament proportional to their vote share in a general election.
Yes. Imagine 100 candidates... A winner would then be determined by 1% + 1 vote. Clearly not the choice of the majority of the population.
United States is not a Democracy.
Good idea BUT if there is no majority in parliament then none of the law can be pushed through
@@Nonixificationthey could just push the laws that a majority does agree with. You know, like getting pizza, indian or Chinese food... You can still get a majority.
Lost of countries have systems that don't use the "winner takes all" system. And it also leads to less extreme choices. But it means that multiple parties will need to work together, and negotiate, to get something they all can somewhat agree on.
The "winner takes all" will always lead to a two-party system, and always to very bad politics IMHO.
@@Nonixification I really hope I am misunderstanding you here, because what you said is very silly
14:13 Imagine being awarded a Nobel Prize for essentially saying, "Look everyone, I've mathematically proven why we can't have nice things."
Introducing 5 axioms, misinterpreting one of those axioms to show a contradiction and get a Nobel prize for it.
What a joke
@@sweatyeti good summary of 20th century voting malaise? In any case, better tests are here…
@@lukamiler5824 Which one? The one about the pivotal voter or...?
@@lukamiler5824 Could you explain the flaws in the theorem please?
well his work did lead to us now understanding why rated voting systems are probably the only systems that work- it may not have been his intention but it was informative enough for us to avoid the slander that rated voting systems could have accumulated, people will criticise anything without any understanding of it just because it could inconvenience them. It is like how when you do a multiple choice question it may be easier to prove why everything else is wrong than to prove why the answer is right.
I think you have achieved wisdom. I don't think anyone here will put one and one together, for the videos you've made during this year. You have my compliments.
I liked the CGP Grey reference. His videos about voting systems are phenomenal. They’re quite simple, but I love how he talked about them. Good job Derek! This is a great video!
Glad that someone else noticed the reference!
Veritasium even made him a hexagon!
CGP Grey's video was the first thing that came to mind when I saw this on my home page and the reference made me so happy while watching!
Not only the 'Hexagons are the bestagons' comment, but the hexagon having the similar glasses as Grey was nice touch.
Title: “Why Democracy is Impossible”
Video: “Why ranked choice voting only works 99% of the time”
and they actually left out a key potential issue with ranked choice too, in Australia a few years back a candidate with only 17 number 1 votes was elected to the senate due to preference deals. He served with honour, but that's the clearest actual real world problem with the system, preference deals between candidates
My thoughts precisely. I mean YEAH there are SOME inaccuracies, but then again it's not like first pass the post system ACTUALLY reflects the will of the people.
Heck, make it simple: As many Candidates as people present themselves, every voter gets 3 choices. 1st Choice, 2nd Choice, 3rd Choice.
The one with the most points wins. Simple as that. Afterall while that person might not be the MOST favourite candidate, they ARE the candidate that most people are reasonably happy with.
Heck, give the people the choice to forgo their 2nd and 3rd choice, a la 'Bernie or Bust!' style if they REALLY want to.
But in that case they shouldnt be surprised if they end up with a candidate they dont like at all.
It's getting sad to see how greedness is affecting good science channels...
Conclusion: Derek's cIickbait strategy works 99% of the time.
I mean if you watch his videos, you know it's usually involves math or science. Thus his title makes sense within the context of his video. Mathematically a democracy would be something where it's free and fair every time. 99 percent free and fair would not make the mark of a democracy. He didn't even cover the peciluar way that the US elects its president. The electoral college works 93 percent of the time.
The truth is, in reality, all the methods he talked about are practiced by many different democracies we all recognize as democracies, whether they use first past the post, ranked choices, or an electoral college system. If you actually clicked on a Veritasium video and expected the host to make a political argument that democracy is impossible, you must have never seen any of his videos before. I knew when I clicked on the video he would make a mathematical argument why, not some actual political argument that democracy is impossible.
As a European, I was patiently waiting the entire video for him to mention that of course, democracy can take completely different forms wherein you don't have to elimimate a list of individuals down to one final victor, so that democracy does not necessarily have to stand or fall with this issue, but I was sorely disappointed.
As an American how do things tend to work over in Europe?
@@keltzar1 If party A gets 48% and party B and C gets 26% each, then party B and C can make a cooperative government and win, or if they cant agree, then Party A wins..
There will be alot more than 3 parties and there is alot of negotiation too see who will cooperate. If noone wants to cooperate, you might end up with a "minority government" wich would be is Party A ended up winning in the first example. Party B and C are not the rulers of the country, but they have a majority for votes and such, while the ruling party has a minority.
It can also happen that Party B and C cooperates, but decides to stop cooperating after a few years, then suddenly the winner is Party A.
All that being said.. I know nothing.. They teach us nothing in school, and what i believe i know i have deduced and infered from the elections i have voted in.
Thanks, that's 20min of my life I'm saving.
@@keltzar1 Imagine there's 100 seats up for election. If I get 2% of the vote I get 2 seats. etc. So you end up with a much more diverse Parliament. This type of system also has its problems and critics and on occasion lead to pretty chaotic and unstable Governing.
how anarcho syndicalist of you
Watching this the election after is wild!
7:55 idk if I agree with this. Cuz if bohr gives a bad speech and gets less votes, presumably a portion of curies voters would also be dissuaded from voting bohr as second choice due to the bad speech. That is to say, I don't think it'd be a 50/50 split of curies votes in round 2
Well i think it would make sense to assume that before bohrs speech the second choice of curie voters was distributed 25:45 (Same as the initial ratio of first votes between Einstein and Bohr) and after the terrible speech it was closer to 50:50 (Like the Einstein Bohr First Vote Ratio after the speech). So while you're right that some Curie voters switched their second Vote from bohr to einstein it still makes perfect sense as long as the second votes of curie voters are similarly distributed as the First votes between Einstein and bohr
Maybe I'm missing something, but I definitely disagree with what is presented in the video. "Clearly this isn't something we want in a voting system"...? I don't see this as clear at all. Thats kind of the whole point in RCV, even if you're not the people's first choice, you are more closely aligned to how people want you to govern.
It also seems like the example has two first rounds of voting. No one would know what the numbers are before Bohr gave his bad speech or whatever.
i'm also not buying the example. the voters had 2 rounds to move away from Bohr
Also, why are Bohrs voters going to Einstein and not the more moderate Curie? If a candidate performs poorly, voters tend to go to their 2nd choice, which tends to be close to their political position. This example assumes voters swing all the way to the other side which makes no sense.
I disagree with that example also. Why did those voters who were turned off by Bohr's speech all go over to Einstein (who was vastly more unpopular overall) and none go to Curie? Even if just a third of the switching voters went to the centre, Curie would have still won.
Whoever does your animations, those pencil-looking drawing parts, great job
Cool, but doesn't this (especially Arrow's Impossibility Theorem) assume that the votes of the citizens are _known_ ahead of time, so that the 'pivotal voter' can change the vote one way or the other? In reality, elections wouldn't actually hinge on a single voter, as they wouldn't be able to know how others ranked their choices _exactly._
In the end, the problem with the current system is not really certain people affecting the vote more than others, but the outcome does not really represent the majority's wishes, especially when you consider the Electoral College and gerrymanderin in the US (Europe and other places luckily doesn't have Electoral College, but I believe gerrymandering is still a problem). Particularly with approval voting, both problems disappear mostly, but it is crystal clear that *ANY* ranked choice method is almost _guaranteed_ to be better than (meaning, a more representative democracy than) what we have now. So it is INSANE to insist on the current system.
The only logical argument would be that it is 'easier to understand', and that makes this that much more insane! Hundreds of millions of people, afraid of primary school level maths and logic...
It's not about changing votes or outcomes (or actually exerting the power of a "dictator"). It's just about those 5 demands being mutually incompatible. There cannot exist a voting system that satisfied all of them. That's all.
No it becomes apparent after the fact (or rather during the counting).
A basic outline of the proof:
1. You assume you have a voting system where IIA and transitivity hold for any possible combination of votes.
2. You show that in this system, by transitivity and IIA, if every vote ranks B either at the top or the bottom, so must the social welfare function.
3. You show that given step 3, there is some “pivotal” voter that determines B’s ranking relative to AC. Note at this step, the “pivotal” voter is just a tipping point (point of no return), not a true dictator.
4. Now, you show that regardless of how every other voter ranks A relative to C, by assuming IIA the social welfare functions relative rank of A and C must mirror this pivotal voters relative rank of A and C. This shows that there is some voter that fully determines the relative rank between A and C, regardless of how anyone else votes (i.e. is a dictator over AC). Call this voter voter x.
5. By similar logic, there is some voter y that fully determines the relative rank between A and B, and some voter z that fully determines the relative rank between B and C.
6. Without loss of generality suppose voter y has A>B, voter z has B>C. Then the social welfare function has A>C by part 5. But this means that voter x is not a dictator over A and C, contradicting part 4. The only way to resolve this contradiction is that voter x is voter y and voter z. But this means that voter x is a relative dictator over AC, AB, and BC, making them a full dictator.
The dictator in this sense isn't consciously choosing to manipulate the voting, but their individual vote has more meaning than every single other vote which is against the common rules set out (that all votes are equal).
The video isn't a value proposition of FPTP vs RCV (and any other single winner election). Just that all of them are flawed whenever there are more than 2 choices.
Of course PR doesn't have any of these specific issues (but does have multiple other issues).
nice to see you here
Why does it matter, anyhow?
Conflating "democracy" with "deciding who wins an election" is missing a *major* point when the politicians will just do whatever they want once elected regardless of "promises"...
The main issue with Ranked Choice Voting is that voters will not be able to make educated guesses of each candidate and the possible combinations of runners. Also, a lot of people will not rank more than two candidates, so there will be a faux majority, because others did rank more than 2.
Who's watching after Trump became president
me
Yep
Me and it's scary, I'm not even American and this will have repercussions in the world
@@lillycue be fr
@@lillycuewhy scary? Where do you get your news from
3:53 of course CGP Gray is saying Hexagons are the best-agons 😂
ACKTHUALLYYY they are not the bestagons. Saw a really good video by Con Hathy debunking CGP's and saying the bestagons were instead, generally speaking, triangles.
My man said "It's the best thing we got and only game in town" after a whole video of showing how there are many different ways of doing democracy, the worst possible one is used in most places in the world and the best we could figure out is not used anywhere
He’s referring to democracy as a system.
Not the winner takes all nature of American politics
Democracy good because everything else not as good
Exactly.. he was completely right with everything and then spat into the face of it at the end.
CLEARLY a better way is available but everyone just says “uhhh bad German man!1!!” without even looking into it.
@@loslingos1232 Not sure what you mean about the bad German. The video also hasn't mentioned direct democracy, referendums, council rule etc. Or anything about lobbying, media capture, manufacturing consent etc.
The practical problem of a hierarchical democratic organization of the whole state (federal government at the top, then state governments, then local governments, then companies and large families, i.e., clans) is more interesting than these theoretical musings. let's hope we get another video on this topic. After all, in theory, theory and practice are the same. In practice, they are not.
All this assumes that we are voting for a singular position to be filled by just one person. There is still proportionally representative voting.
could be different parties too tho
Exactly, proportional representation is a proven and effective form of democracy. Unlike the winner-takes-all system, which often distorts the true will of the voters, proportional representation ensures that the distribution of seats in a legislature more accurately reflects the percentage of votes each party receives. This results in a more representative and inclusive government, where even smaller parties and minority voices have a chance to be heard. The winner-takes-all system, on the other hand, often leads to disproportional outcomes and can entrench a two-party system, stifling diversity in political representation. In contrast, proportional representation encourages a more balanced and fair political landscape, making it a far more democratic approach.
The thing about proportional representation is that proportional representation can also result in extremely disproportional effective power. In the extreme case you can end up with a small party and two large ones just shy of 50% of the seats, and despite this, the small party has the same amount if effective power as the large ones, as any combination of 2 parties has a majority, and no single party has. Now imagine the two big parties being opposides that barely agree on anything, and the small party representing some narrow interest group that only carss about that and nothing else, and you end up with a small "kingmaker" group with way more influence than it's proportional vote.
@@kobyma2 In Denmark, the political landscape is both diverse and dynamic. There are approximately 14 political parties to choose from, at the moment, each offering a range of candidates that you can vote for personally. This personal voting system allows you to give your preferred candidate a significant boost, even if they are not at the top of the party list.
Coalitions in Denmark are highly flexible. If a party cannot secure the necessary mandates within its current coalition, it is free to seek support from other coalitions or parties. This flexibility encourages continuous negotiation and compromise, making the Danish political system both inclusive and representative.
Isn't this all assuming that by voting for a party, you support all their stances on all political issues/topics at this exact percentage? A voter's opinion isn't represented correctly in the first place since you don't vote for each topic separately
Honestly Love how this Educational channel makes there Ads in last unlike most who put in right in the quarter part of there videos which is possibly I do not like the most when it’s not presented creatively.
3:54 I love the CGP Grey nod in a video he definitely has covered some of the topics of, and also love how you put your own spin on these videos
I love the "Hexagons are the bestagons" :D
Hexagon is the bestagon :p
CGPGrey has done so much damage to voting reform selling Ranked Choice Voting.
This video does his best to undo that damage by asserting that Ordinal methods are trash.
CGPGrey also has an Approval video but insinuates it's only good for picking lunch.
imho STAR Voting is best. If I can't have STAR then Approval is best.
I am an engineer, not a mathematician. Let's not make practical the enemy of perfection. We should use the method with the least chance of undesired results. After all, little in life is perfect.
Exaclty, especially when we don't even agree on which necessary/sufficient conditions.
The least one with undesirable results? Tell me, how is democracy helping White people to secure their borders, any of them for that matter, on all of our countries. How is it the fertility rate? Happiness? Health? The safety of our families? It would be one thing if we just stopped having almost any children, but we are bombarded with legal and illegal immigrants across the whole of the West. We are 7% of the human population world wide, we are the minority and we are closing in to 0% in a 100 years, or becoming a tiny r4ce of sl4ves for 100's of years more. I don't understand this cheap, weak and cowardly 3rd rate philosophy which older men cling on to. You are leaving us with nothing but D3ATH.
So we don't Even try to make better things? I'm also an engineer, and I think that things that can be improved should be.
The least one with undesirable results? Tell me, how is democracy helping White people to secure their borders, any of them for that matter, on all of our countries. How is it the fertility rate? Happiness? Health? The safety of our families? It would be one thing if we just stopped having almost any children, but we are b0mbarded with legal and illegal immigrants across the whole of the West. We are 7% of the human population world wide, we are the minority and we are closing in to 0% in a 100 years, or becoming a tiny race of sl4v3s for 100's of years more. I don't understand this cheap, weak and cowardly 3rd rate philosophy which older men cling on to. You are leaving us with nothing but D34TH.
The least one with undesirable results? Tell me, how is democracy helping White people to secure their borders, any of them for that matter, on all of our countries. How is it the fertility rate? Happiness? Health? The safety of our families? It would be one thing if we just stopped having almost any children, but we are b0mbarded with legal and illegal immigr4nts across the whole of the West. We are 7% of the human population world wide, we are the minority and we are closing in to 0% in a 100 years, or becoming a tiny race of sl4v3s for 100's of years more. I don't understand this cheap, weak and cowardly 3rd rate philosophy which older men cling on to. You are leaving us with nothing but D34TH.
18:30 a thought: if the election is confidential, it means you cannot really know if you are the pivotal voter or not. Everyone is potentially the pivotal voter. So can't you then say that everyone is 1/15 th part dictator, so everyone is basically the same again? Meaning: the collective is the Dictator, which is sort of the translation for Democracy?
or put differently: if an election is close, there is always one voter or small group of voters that would count as dictator. But since there is no order of votes, all those votes are equally 'dictatorial'.
Which is usually the Majority
Given an election system, the pair-dictator is unique. Anyone can just study the mechanism to find out.
@@nimrod06 unique, but interchangeable. If you have a certain amount of equal votes you can interchange them all.
@@mistasomen that's totally not true.
No, this is incorrect, and Veritasium did a terrible job finishing up the proof (which some other commenters and Wikipedia clarified).
The dictator is part of the election system. The proof shows that any election system satisfying the 5 postulates has the same one guy as a dictator, regardless of what everyone else thinks. Who the dictator is can only depend on the issue we're voting for (so you could e.g. have a dictator on taxes, a different dictator on food, a dictator on the military) and not any of the voters' preferences. Which means you could theoretically analyze the voting system before the election and figure out who the dictator is.
Now it's true you could randomly select the dictator for every election so nobody would know who it is and voting would still be worthwhile, but random systems were excluded from consideration earlier in the video.
Very nice topic and also explained very well. From India
I live in a "first past the post" country that forces a two-party system and penalizes voting your conscience unless it aligns with one of those parties. While there may be flaws in Ranked Choice Voting that could emerge in fringe cases, it is so obviously superior to our current system that it is hard for me to worry too much about the nuance of how it might not be 100% perfect 100% of the time. Any (democratic) system is better than what we have now.
The rated voting system sounds even better imo
The US has had first past the post for over 250 years and only had a "forced" two party system for fewer than 100. Coincidentally the prevalence of the "two party system" began around the time universal suffrage was introduced.
RCV solves the spoiler problem, but it does not solve the representativeness problem. It just makes smaller parties inoffensive to the "big two". It actually performs terribly and erratically when more than two parties are competitive, which is exactly the scenario where you would want a better voting system to matter.
@@scopeguywhat do you mean by this? With very few exceptions, we’ve only ever had two parties really in contention for races.
Third parties have EXISTED, sure, but their main effects have been bringing new ideas to the two major parties’ discussions, and splitting the vote between more moderate and more extreme members of the same “side” of political thought
What’s the country?
3:52
Hexagons are the bestagons 🔥🗣
PENTAGONS ARE THE WENTAGONS 🔥🗣️
Looks to me like he is trying to summon someone.
Yes, yes they are. That is most definitely the most iconic CGP Grey phrase.
I saw too 🤩 +1 for CGP Grey!
did someone just mention benzene?
Now lets introduce a term "coalition" to this video.
I was going to mention that. It doesn’t has to be: winner takes all.
What if candidates get power in proportion to their vote? They can rule by consensus.
The real problem I see with Democracy is weakness against other powers ruled by Autocracy.
@@atscub you are on the money! I vote for A because I like rules from A and no one else, likewise for people who voted B etc etc
OH WAIT that's called a free market, lets skip voting and just pay for goods/services from people who are servicing us the way we like
@@shawnradkeExcept then you immediately deny the principle of democracy by ceding power to people with the most money. Not to mention the free market fails at providing ideal choices all the time.
Except that doesn't work when you're deciding a LEADER, you f+cking m+ron.
Also, proportional representation sucks, and coalitions are a terrible idea.
@@bminecreeperyeah and also monopolies exist, they will just bankrupt anyone trying to offer a service so they are the only one that people can purchase from
This is an older video yet perennially relevant.
This discussion has come up in 2024 within my company's chat with some folks in the US, and others in Europe, and some on the US west coast getting to use Ranked Choice.
I shared this video in the discussion when I argued that first-past-the-post is rational if not optimal, and forces coalition choices to internal within the 2-party choices, and that the 2 US parties change coalitions internally over time.
The Bohr-Curie-Einstein example is odd. The example assumes Bohrs poor performance leads to votes for Einstein and not Curie, despite Curie being much more closely aligned politically to Bohrs voters. Obviously this cooouuuuld happen, but its very unlikely; far more likely Bohr's loss is Curie's gain.
And ultimately, the voters still expressed their preferences, and that person was elected.
Preferential voting has flaws (too many candidates for voters to reasonably assess for one, like that Minnesota example; also potentially causing a candidate that almost no one liked to become elected due to weird preference flows), but tbh I'm not sure this example demonstrates them super well.
Yes, I felt that this example was contrived and unconvincing. If that's the best argument against Instant Runoff, it looks like a resounding endorsement instead of a criticism. Can't be worse than FPTP.
Arrow's impossibility theorem shows weird things CAN happen, but not how likely it is to happen. The condorcet loop for example, is pretty much impossible in practice and that's presented as the methods big downside
Odd as it sounds, such a scenario is easy to write (based on US politics). Imagine a third party, the Liberals, who united the disaffected into a 20% voting block. We'll also say that everyone kinda likes the Liberal, while everyone else dislikes the Democrat and Republican candidate.
Thus, the candidate with 20% who will drop out handily, is the one everyone was okay with. And the winner will be despised by 60%.
Step 2: now the Democrat does a solid (if dirty) attack on the liberal and takes 5% of the vote from him (D45, R40, L15). Looks like he's golden, right?
Well, no. The Liberals now all switch their second choice to Republican. Which means that this 5% gain ensures a 45-55 loss against the Republican.
Step 3: The LIberal now points this out, and tells blue voters that their party has *zero* percent chance of winning. However, if they vote Liberal, they'll win. Not everyone is convinced, but he gets his 5% back, and an additional 10%. So now the polls say R40, D30, L30. That is to say, the very fact that he lost votes, is the reason he now draws even.
So now consider: it's a dead heat between L and D. If the Liberal scrapes ahead, he'll gain all the D-votes and the Republican will lose the final round in a 40-60 landslide. But if the Democrat picks up a few votes, then the Liberal gets eliminated, and even if the Democrat gets his 10% back, the Republican will still beat him in a 60-40 landslide.
A few votes, therefore, will decide whether the Republican wins in a landslide (against D) or loses in a landslide (against L).
Finally then consider their optimal strategies. The Democrat may be wisest to just drop out. He can't win, his best chance is to back the liberal and ask for some concessions.
The LIberal's best option may be to demoralize Democrat voters who refuse to switch. If a few percent don't show up... sure, he'll lose those votes in the final round, but he'll make the final round in the first place.
And the Republican's situation is the strangest: ideally, he would give some of his own votes to the Democrat, so that the Democrat passes the first round, only to be smashed in the final round. The GOP may use a well-selected faux-pas that only the small portion of his base that picks D second will switch for. If it's close, GOP operatives may even *vote democrat* to make sure the GOP wins.
I think all of that is correct. And I think you'll agree, this is a realistic scenario. Yet it has several strange results in it.
I came to comment on this too. I felt this wasn’t a very good example of a problem with ranked choice voting. In the instant runoff example at least everyone who voted was able to have their vote counted toward the eventual winner. Can weird things happen? Sure, but at least your vote wasn’t completely invalidated & affected the final outcome.
@@valentetorrez3398
I found some stuff confusing, but the overall point is that there are various things you'd want in an election (particularly where one guy gets the job)... and you can't get all of it.
So you're gonna have to decide which needs are most important.
15:45 I don’t understand? By changing position of A and C such that all A are below C, aren’t you either not changing A at all or making all A worse than B by a larger margin or narrowing the margin such that A is now only slightly preferable to B?
To me, the clear winner is C; they have first choice on half and never last choice. Meanwhile B and A both have last choice for half with A never having first choice. C > B > A. Because half say B>>A. And half say A>B. Half say C>>B and half say B>C.
Please somebody explain differently?
The video is just stupid
It’d work as you said if you were using a Broda-like ranked system (point based - where the “distance” matters). They are using the Condorcet method where you only look at the head-to-head matchups. I agree here it’d feel more fair and natural to use the former here.
Thank you, I'm not insane. The example doesn't work. He arbitrarily puts the final ranking as A-B-C, but then the voting on the left shows a pattern that would only allow for C-B-A. And if you change the final ranking to that, then there is no contradiction.
7:32 If Bohr becomes more unpopular why would his voters switch to Einstein and not Curie who is closer to Bohr ideologically but with none of the baggage of his controversial comments? In that case it would be Einstein who gets eliminated in the 1st round instead of Curie.
Because the video has to make contrived circumstances for ranked choice to not be obviously superior.
@@F0XD1E There's something to that -- I think it's worth asking, when identifying flaws in a system, how likely those flaws are to affect the outcome.
One example of this could be in the recent UK election. You have Labour, Conservatives and Reform. Some reform voters would be more inclined to vote Labour than Conservative because they were unhappy with the Conservatives who were in government at the time.
"The political spectrum is a U, the ends almost meet" - quote by some guy, I forgot who.
I'm from Germany and especially in the easters states, voters frequently switch between the far left and far right. Those two parties share the "anti-establishment, blame [insert part of society], anti-institutionalist, propagandistic views". One just blames the rich, the other blames the immigrants. So, it's definitely not impossible.
The movement of votes is unimportant. The point is that if the distribution is 25-30-45 curie would win, but if the distribution was 31-30-39 Bohr would win, even though Curie got the same amount of votes.
This is most fascinating, thank you for this knowledge .
I find it interesting that every problem presented relies either on the premise that the voters need to elect only one winner or on the idea that a candidate who wasn't anyone's first choice but was everyone's second choice isn't a good compromise. These things are just presented as facts without establishing why this should be necessary.
I hoped that in the end, Michael reflect how this theory could really represent real elections struggles. And how power dynamics fumbles with the math principles in the "fairness"
It also makes a huge assumption that “democracy = elections with voting” and excludes any other form of collective decision making, of which there have been many throughout history (like lots of indigenous societies in the Americas, or the Catalonian Anarcho-Syndicalists who used systems of mandated recallable delegates, or a variety of activist groups who use some form of modified consensus, etc)
@@shredded_lettuce The drawback of these other forms of democracy is that they even further empower the masses and disempower the individual. Democracy was a bad word until the rise of mass society in the 19th century. What we need to work toward is a world without "the masses." In mass democratic society there will always be a tyranny that manipulates social reality to guarantee its hold on power. People who live and think locally are free to organize themselves however they please. Most of the time, it turns out this is some combination of monarchy and aristocracy. Everyone in such a local world exists in an organic matrix of direct obligations and privileges, duties and benefits. It's how families have always operated. I would dare call it the natural mode of human organization.
@@shredded_lettucehe's either clueless, or more likely, pretending to be so he can get views by making people outraged that democracy is actually "mathematically impossible" when it isn’t. It's perhaps his weakest video yet, certainly the most repugnant.
@@shredded_lettuce if it was used by indigenous americans, its probably worthless then.
Maybe a defeated and crushed people arent the one to take advice from.
"the pivotal voter is a complete dictator" NO! This is such faulty logic. They're only the pivotal voter because of the votes of the rest of society. It's literally an exercise in teamwork, they could not have done it alone.
I think he means that when a lot of people vote for group X and a lot of people vote on group Y, group Z could still win( by less votes). That means that votes cast on less popular candidate would be more powerful.
The problem runs even deeper than that. The pivotal voter is not any kind of dictator because they do not "dictate" anything. Regardless of what personality is appointed to the post, it is the ruling class as a whole that dictates how the system itself is run. Individuals are not that important. In fact, it is that system that produces individuals who take up particular posts.
it depends who the pivotal voter is
he is the man that choses power over everyone else's choice. he is an "oopsie dictator"
What I understood was that the pivotal voter becomes a dictator because they rearrange their vote (in the example at 17:19, moving A above B). If any one person with B above C had done the same, then assuming all over votes don't change, they would be the "dictator" instead. ("Dictator" is a bit of a confusing word choice, since it is a different definition of dictator than usual. They aren't being called a fascist, rather that their vote becomes more relevant.)
AND if this was the end of it, it wouldn't be an issue. The majority of people put A above B, and either prefer A to C or have no preference. The issue is that now the transitive property and the independence of irrelevant alternatives cancel out, as when we start having voters preference C to A, suddenly it creates a loop.
It's all a bit obtuse, and I think the three-voter example (pizza, burgers, or sushi) was easier to follow. But honestly, all this just comes down to: "What do you do in a tie?"
My favourite voting system is STAR voting which stands for Score Then Automatic Run-off. Basically, voters score each candidate from 0-5. Then the scores across all ballots are tallied. The two candidates with the highest scores then move forward into a second round of counting where each ballot is treated as a single vote for one of the two candidates depending on which one the voter gave a higher score.
The system is practically guaranteed to produce a winner with majority preference and at the same time allow all voters to express their honest preferences without a spoiler effect. You might think that voters will just bullet vote (scoring 5 for their favourite candidate and 0 for everyone else) but this is actually a really risky strategy as it means if your candidate doesn't get to the run-off, you're basically saying you don't have a preference over anyone else. So, it incentivises honest voting.
Is it a ranking? Where you can only assign each value once? And must assign all values 1-5? Or could you say, vote 5, 5, 5, 0?
STAR voting deserves a chance to shine!
Yes, that's actually just ranking system with the possibility of a tie in your options. IMHO it's the best system.
Everyone will just rank their favourite candidate full marks and everyone else zero. It's the best strategy. It will just default back to first past the post.
I'm not a big fan of STAR because it implies that a simple head to head preference actually reflects society's best candidate.
If one candidate gets 4/5 from everyone because they're reasonably moderate and agreeable, but then another gets 60 votes 5/5 and 40 votes 1/5, because they're on an extreme that is slightly favored by the group, the extremist wins.
And the all around agreeable candidate loses. Which I don't think is great.
This is pretty fascinating
7:44 Why would voters switch to Einstein instead of Curie? Didn’t you just tell us that Einstein and Bohr have very different stances?
Because e = mc^3 is just so wrong people hate Bohr :P
Yeah, it really doesn't seem like this flaw of RCV is a practical concern.
That's mentioned at the end of the video that in real life, that particular error is less likely
I think it is because the voters want to vote against Bohr and Einstein is the best way to punish Bohr
Say for example both Einstein and Bohr developed a stance where they inadvertently proposes the same idea of universal basic income. But they name it differently and the source of the idea is politically different. Bohr then proposes that the income should come from eliminating veterans' pension. Some Bohr voters would look at every candidates and found similarities in Einstein's economic policies. These are the voters that will switch.
Also, some people really dislike being called a fence-sitter, or a centrist. Those also would be the ones switching.
All this only applies to democracies that vote for 1 candidate or 1 party to rule, but the whole problem ceaces to exist, with proportional representation. Would've been nice if you mentioned that.
Yeah there are so many different forms of Democracy. I think the title should say be "Why First past post and preferential democracies are mathematically impossible". Would've loved for him to go over sortition.
Its like you didn't watch the video. Anyways, the problem is the assumption of unqualified franchise, compounded by the fact that politicians can now import new constituents.
Thanks for saving me 15 minutes 😂
@@churblefurbles Why? He is right? You could argue that these issues happen in negotiations inside coalition forming to a lesser degree but it is a whole new system to investigate with potentially (dis)advantages.
That's what I was going to say too!!
@4:00 I see CPG Grey glasses
For a hexagon, which is the bestagon, no less
This is the second time this week I've seen a video on this exact topic with a Grey reference. This is because he did a video about this exactly years ago
Was just about to comment the same
I also saw Grey's voting vids
Me: "Wait is that...? Yeah it's him."
The most important thing is to minimise gerrymandering and for all voices of society to be represented, without favour for one over another, rather than aiming to be mathematically equal
The funny thing about voting is that most of the time we don't get to pick the options we're voting on, rather we're delivered a curated selection of options and are left to choose the one we find most palatable.
The establishment wins no matter what. All the "options" you get to "choose" from help them regardless
Then join the body that curates the options?
For example, here in the UK, join your local party. Then you’re one of the ones picking the candidates. In fact, since we have a lot of safe seats, joining the local branch of the party that controls your seat effectively makes you one of the few dozen or so people in that constituency whose vote counts for anything.
I’m sure there are similar possibilities in the US and other places.
You can just stand in the election and wala u picked a choice
we do though, its just every finds the 2 big ones and never even try to support the others, especially in the west and even more in the US
"The funny thing about voting is that most of the time we don't get to pick the options we're voting on, rather we're delivered a curated selection of options and are left to choose the one we find most palatable."
and the ones who arent curated are smeared endlessly by the establishment and the media, so to th extent that they often become the most hated people in the country
11:27 it’s also called rock paper scissors
Was thinking the same lol
I fail to see the paradox of perfectly balanced preferences not leading to a clear winner in an election. I didn't even use the same method i just noticed that one party liked each food most, second most and least. How is ANY fair voting system to have a clear winner there? Such occasions should lead to coalition government
Love the CGP Grey reference at 3:40. You even made him the bestagon hexagon!
Yeah love the reference he makes to cgp grey's election videos
Yeah it’s awesome
Maybe this comment has been made and I can be directed to it, but I'm a new teacher (grade 7s) and we're learning about government in BC, Canada- DEFINITELY going to show my class this but just seeing if any teachers have already developed a lesson plan around it?
Thanks Veritasium!
British Columbia tried to have a referendum for preferential voting, but they didn't explain it well AT ALL so most people just voted for "no change" rather than be confused by what might happen if we changed. I really wish they'd redo that vote but explain things better. Every single person I've talked to hates the current first past the post system in Canada.
Just because first-past-the-post's spoiler effect is awful doesn't mean there aren't worse alternatives. As Kenneth Arrow proved, any kind of preferential voting system will have some degree of arbitrariness. Ranked choice voting looks great at first, but it has a ton of major, non-obvious problems. (Look up "Instant Runoff Voting: Looks Good--But Look Again" by Stephen H. Unger if you're interested.)
And people being confused is a valid reason to not go with a preferential voting system; it makes voting tallies far more opaque to the public, which is a problem when people already have a low level of trust in their governments.
Score voting systems (including approval voting) and STAR voting are much better alternatives (and they _actually_ remove the spoiler effect, unlike ranked choice voting, which only seems to at first).
Ranked choice compared to first past the post adds complexity and cost and that is about it.
I wouldn't trust any British Columbia voters to make a right choice on anything. Once there was a vote for keeping HST so everyone pay less tax or switch back to PST+GST, they voted for paying more tax.
@@RobinClaassen To be kind to Unger, he's analyzing Ranked Choice in the context of abstract statistical anomaly. It's what Veritasium was saying at 7:30 - that a wing party dropping votes to the other wing can get the moderate eliminated first. Fun for math. Ridiculous for humans. Voters would be difficult to shift like that and a candidate would be crazy to try.
To be less kind to Unger, I also read some of his other articles. He sounds a lot like "The best system is the one that gets me what I want!" But I hope most Veritasium viewers would understand that this short-sightedness is not a good way to design your voting rules.
@@notme222 I do find it somewhat difficult to get a sense of how common some of the situations IUnger is talking about are likely to be (e.g. those situations in which all voters reversing their preference orders would result in the same candidate winning). I was more or less talking his word for it that they were plausible. In that post, he gives the impression of being unbiased. If you're saying that he gives the impression of having motivated reasoning in his other writings, that's a reason to question his assertions.
That said, some of his examples are more clearly likely to be common occurrences. For example, in any situation in which there are at least 3 candidates with high levels of support (each having some chance of winning), there will always be some voters who will be incentivized to downgrade their ranking of their most preferred candidate because they'll want to make sure that the first candidate with a high level of support to be eliminated isn't ideologically opposed to them (such ballots would be likely to have an also-ideologically-opposed candidate as their second pick
They'll instead want someone who they ideologically agree with to be eliminated first, and may therefore feel afraid to give their preferred candidate their first choice if they think that that might result in stopping that candidate from being eliminated first. It's essentially the same spoiler effect that we get with first-past-the-post systems.
So it shares that problem with first-past-the-post, while also having many problems that first-past-the-post doesn't have. Including the results being sensitive to the level at which they;re tallied (e.g. county, state/province, or national), and those tally results being more opaque to the public.
And given how much preference information voters are expressing on their ballots, it's frustrating how deaf to most of that information its vote-tallying algorithm is. It only reads one line at a time, ignoring the rest of the information, and never even reads most preference information expressed on most voters' ballots. When it does go past the first line of preferences, it arbitrarily only does so with some voters but not others.
While you can express more preferences on a ranked choice voting ballot than a approval voting ballot, the approval voting algorithm is actually the one that will take more of your preferences into account.
Its wild that he went so far but then never mentioned STAR or STLR. Its like running a marathon then stopping with a meter to go
Agreed. I was also waiting for STAR at the end ... and then it was just over.
What is that?
@@Franimus Just Google it buddy... it's not much work
@Franimus if you had 3 candidates A B and C, and then let the voters rank how good a candidate is overall on a 1-5 rating. The 2 candidates with the highest average go to a runoff, where everybody revotes for whatever candidate they want, whoever gets a higher approval wins
To be clear, the runoff's "revote" happens automatically because the ballots already show which of the two finalists each voter prefers, so voters only need to go to the polls once.
For example if you had ranked A (5 stars) B (3 stars) and C (0 stars) and then it turned out that A and C were finalists, your vote would be given to A in the runoff. If B and C had been finalists your vote would go to B.
It has the major advantage over classic RCV that it only ever requires 2 tabulation steps (the score and the run-off) and that both can be easily counted by hand. In theory it also eliminates the need for primaries, though it still works fine if you have them.
Here is a better title: "Why Winner-takes-it-all Democracy is Mathematically Impossible".
The first theorem only applies if there is a single winner in every district.
Having actual proportional representation does not have these flaws.
Not quite true. The underlying problems still exists even with proportional representation, if based on rankings. It's just less noticeable because the PR "simulates" cardinal information in aggregate.
I was thinking "Why democracy isn't an abstract, rational, mathematical problem, but if it was we'd have a problem"
JOIN THE FIGHT FOR SUPER EARTH AND MANAGED DEMOCRACY!
It’s not about that, humans are humans period. We have terrible traits as a species and those have been exploited for so long we are programmed to it.
Less clickbaity you say?
3:40 Love the subtle nod to CGP Grey who did a great video series on this topic like a decade ago!
Clearly FPTP is the worst system. We need to do better.
Just better, not perfect.
For real.
FPTP is fine (meaning not the worst) as long as the electoral district is small enough.
Yup. Just have a runoff between the two most popular options. It works in most places. It leads to good outcomes most of the times. Usually the more centrist of the two wins. Problem solved.
Or, like Sweden, you have party systems, so you vote for parties, that also works OK, but that would likely require constitutional changes jn the US. A two stage system just needs open primaries, on the same day, later in the cycle.
@@RegebroRepairslogically (relatively) center parties should win most often, but if you're saying center=better=good you're dead wrong
Approval voting solves literally everything discussed in this video, and requires virtually no change whatsoever in the way elections are done. Just "click or check" every person you like. Person with most clicks/checks wins. It's so simple and so obvious it is maddening we don't do it this way. In fact, a huge number of "spoiled ballots" are because someone checked more than one person, thus invalidating their ballot. A little disappointed Veratisium didn't spend more time explaining Approval Voting. How it requires virtually no adaptation to implement, and how it completely demolishes rank choice in every conceivable way. It has the finality of first past the post with the variance of options of ranked choice.
Lost me around 16:00. I don’t accept that one of the rules should be that an extremely divisive candidate that splits the vote between love and hate should either be completely on top or completely on the bottom. Intuitively, the middle IS where they belong.
If our criteria insist they have to win or lose, our criteria are wrong. Electing them or dropping them would make a LOT of people unhappy.
You have a point. When at paradox every given option is equally good, in order to decide one have to refer to other set criteria outside of the parodoxical ones. If {paradox} then: other formula.
I've seen lots of comments about the Bohr voters switching to Einstein. The reason he gave of *why* people switched distracts from the point. The real problem being highlighted is that the system incentivizes "strategic voting". Suppose that in polls leading up to the election, the results are 25% - 30% - 45%. If we assuming that people are voting according to their views, Curie would win (as explained in the video). So, the Bohr voters could (in theory) organize to have 6% of them rank Einstein 1st, which would result in Bohr winning. Obviously it's not good when voters can deliberately down-rank their preferred candidate to make them win.
polls could become illegal
This all comes down to if his critique of the RCV is a purely mathematical one or does it have any real life value. For one, the example he gave is completely illogical, even in purely mathematical terms, it will never happen. Bohr voters will never switch to Einstein. I am disappointed that a scientist would ignore logic just trying to make a point to fit his thesis. The explanation you gave, while perfectly logical, are so difficult to achieve it in real life, it might be almost impossible. Even without considering polling errors, the amount of effort needed to ask a CORRECT portion of your voters to hold their nose and strategically vote for another candidate is so huge, it's not realistic, purely theoretical. This, and the pivotal voter paradox, plus the terribly misleading title, makes this video one of the worst video this channel has put out. I mean, I still love this channel, but he can and should do better.
@@zhenfengliu8032Yes, there's a logic error thwre. Assuming that most of the voters jumped from Bohr to Einstein but still more of Curie's 2nd choice would move to him rather than to Einstein suggests Bohr is the moderate candidate, not Curie. And Bohr wins as a moderate, as more people oppose Einstein.
This comment chain needs more upvotes. Thank you for explaining better than the video or other commenters.
The first thing that came to my mind after seeing the title- "CGP Grey"
3:39
This feels like a mathematical proof that suggests problems in edge, corner or point cases make an entire system unusable. I’ve never interpreted “impossible” to mean “it only works 99% of the time”.
The recent 2022 Alaska special election had Palin spoil the election for Begich. There's not even any need to speculate about what voters would have done, since it's all in the voting data. Begich was preferred to both Palin and Peltola but Palin stole too many hardliner votes away from him.
It's more that real-life voters are not accurately modeled by the assumptions of the theorem. People act illogically!
@@DavidRoberts The theorem makes no assumptions on how people act. It's only concerned with showing that 5 random demands are mutually incompatible.
When it comes to first past the post, it's looking like the whole thing is edge.
Veritasium makes rage-bait to help people click and engage. It sucks, is probably very effective, and I hate it and him on every level for it
19:00 house of cards cutaway
I like how CGP is included when he also made videos about how the voting system sucks. Very cool
3:59 No way Gray is here😂
I mean he literaly made 10 videos abaut the same thing.
While it's perturbing that the current voting systems might be fundamentally flawed, I'm intrigued by the prospect of approval voting systems. I hope we soon see more real-world trials for the same.
Check out S.T.A.R. Voting.
Score Then Automatic Runoff. It came about in the last decade and shows incredible promise.
This video only speaks about Winner-takes-them-all voting systems. Most democratic countries don't use them. If the voting distributes about 4 parties like this: A: 35% B: 30% C: 20% D: 15% that's a totally fine result. You can fill the seats of the parliament proportionally. Then A and B can work together or A and C or even B, C and D can work together if they can agree on the same laws. Many democracies work that way and it's not mathematically impossible.
Besides that you could have even other ways of democracy. You could for example get rid of votes and instead pick random people for a parliament. Put people in groups of 6 and throw a w6 dice to choose one. When you have reduced the candidates to 1/6, do the same again with that group until you have enough people for the seats of your parliament.
Approval has a serious flaw. In the election between Ms Loved, Mrs Hated and Mr Indifferent, you obviously approve Loved and don't approve Hated; but if polls are close you don't know if you should approve Indifferent (which might beat Hated) or not (because he might beat Loved). This is called "later no harm" and can cause approval votes to behave as if they were FPTP.
The real best solution is something the video doesn't even mention, proportional representation.
Primer does an excellent video showing the impact and downside different voting systems have. He uses blobs for simulations and while yes they are programmed simulations they attempt to show how things can play out and the different strategies that voters could "use"... use being in quotes because honestly strategy only works if you have a large enough group to agree to it.
Also @3:48 nice CGP Grey ref!
Love this channel! 🤗
Before I watch, is it Arrow's Impossibility Theorem and various apportionment paradoxes?
EDIT: I was not disappointed - it was. Although I learned quite a few things about the history of the field. I had no idea that attempts to replace our silly plurality-rules format had gone so far back (and things similar to Arrow's had been noticed all the way back then). Moreover, I was greatly pleased to learn that Arrow's Impossibility is much less of a problem for IRV in actual practice. Of course, I wouldn't stop supporting ranked-choice voting/IRV just because of that (it's better than plurality and we can't let the perfect be the enemy of the good-enough).
I too, support rated-choice voting over ranked-choice, but with two caveats: First, I think we should establish ranked-choice first simply because it's simpler and people have heard of it. It is good enough, after all. Second, there is one form of rated-choice which I can *not* stand. That's approval voting. The problem with it is, that you preference for all tolerable options is considered identical. You wouldn't be able to spoil your least objectionable option, but you could spoil your favorite option. In ranked-choice, you make very clear which one is your favorite, and that one will get a leg up. If ranked-choice was available in 2016, I could've cast a vote I was proud of without spoiling things for the lesser-of-two-evils we had then (I still did, because I don't live in a swing state and the popular vote doesn't matter, but I digress). I couldn't have done that if all options voted for are rated equal; I would have had no way of communicating just how awful I considered my least objectional option to be.
For rated voting to work *well* the voter *must* be allowed to give a higher rating to some candidates than to others.
There is no single dimension that ranks all candidates, because by definition people will prefer different areas. Approval voting definitely makes more sense, since that means the most people agree on the candidate.
The problem with Approval voting is that it's actually just first past the post wearing a mustache. If we assume that A and B are the two "best" candidates, and everyone assumes one of the two will win, with B having a very small advantage. How should you vote if you like A more than B, but you approve of both? Or if you approve of neither? Either way you should say you approve of A and disapprove of B, because otherwise your vote is entirely worthless.
Now lets say that while A and B are both similar, C is very different and is also a decent candidate. Unlikely to win, but likely to draw about 35% approval. If both A and B are perfectly matched and both vote selfishly against each other, C will win. That's the spoiler effect. On the other hand, if supporters of A or B choose to approve of the other to make sure C doesn't win, then whichever candidate's base is more "selfish" will win.
In that case, the most fair thing to do would be to form a coalition between A and B, have an internal vote of their supporters to see whether they prefer A or B, and then have the loser drop out of the race so the winner can beat C for sure. And that's what we call party primaries in the US.
@@driftwisp2797 Comparing approval with FPTP is simply wrong, because the strategy in approval voting is never the same strategy in FPTP. There is never a reason not to approve your favorite candidate under approval, whereas this is common under FPTP. So any analysis or claims that approval falls back to FPTP is false, regardless of any convoluted scenario or logistics on top of it.
@@1ucasvb There is never a scenario in Approval where voting for your first choice is wrong, that's 100% true. Approval is "FPTP but it's always okay to vote for your first choice".
There are, however, many scenarios in which "approving" of your second choice is wrong. In fact, in almost any scenario where your first choice and your second choice both have any realistic chance of winning, voting for your second choice too is hurting the chances of your first choice winning.
If each person only casts a vote for their first choice, it has become first past the post. If people are casting multiple votes, it means they think only one person they're voting for has a reasonable chance of winning, so it's unlikely to ever have an impact on the election.
When it comes down to it, the main difference between Approval Voting and First Past the Post, outside of "any convoluted scenario" is that after the election is over you have a better idea of how well the candidates who had no chance of winning actually did, and you can see whether the candidates were generally unpopular (people throwing away their vote by not approving of either, like they do in FPTP) or generally popular (people throwing away their vote by voting for both). Great for statisticians, not changing the election.
Approval is ok. But tbh I'm not sure it's better than ranked choice.
The tick box version places a binary decision on the voter for each candidate, rather than a ranking. So individual voters have a hard time expressing who they REALLY want.
The rated system with numbers (e.g. -10 to +10) solves that problem, but creates a complexity issue that is likely to be politically gamed. How can a voter know how much approval to give someone? What if their +5 is enough to put that candidate over their +10 choice? It's not super complex, but it's enough that politicians will go with (+10 for me, -10 for everyone else). Which becomes problematic not mathematically, but electorally.
@@driftwisp2797I don't see how it is comparable to fptp given there's not ever a reason to not vote for your favorite candidate
Checking off/ giving % to any and all candidates you approve of is cool theoretically. But many people will realize they are seriously kneecapping their major picks for each minor pick included. Eventually we're back to 2 party system. The ranked voting in your example was polarized to the extreme, but candidates would realize the best way to succeed is to be accepted by most rather than polarizing
3:52 I would be a proud Grey party voter
Eh, CGP Grey has a mix of great videos (hexagons are bestagons is pretty convincing if unimportant, spaceship you was incredible in early Covid lockdown, and his various explanations of voting systems are great primers) with some pretty glaringly bad takes (the British royal family is actually cool and good, self-driving cars are the solution to traffic).
Hexagons are the bestagons!
@@anezay4987 Grey lost me when he proclaimed that death should be treated as a disease to be cured, not the natural order of things
@@RogueZealotwhat is wrong with that idea?
Same here
Thank you for sharing this video. The majority (59% in a recent poll) have never even heard of proportional voting systems, and the benefits (although not perfect) that they can provide. I highly encourage everyone that believes we can improve our democracy, to share the ideas of proportional voting to their family and peers.
What benefits? They are just empty claims when in reality it works the same no matter what you do. You end up with 2 opposing groups, the only difference is you are huffing your own farts thinking that you are brighter than others because you have your head so far up your own butt you aren't able to observe reality.
may i ask what poll this was and where
As a spanish comedian puts it: Democracy is the system that ensures we're not governed better than we deserve
That's cute you think you deserve something.
@@jessicastevens538 don't we all?
@@jessicastevens538 You do realize the joke is saying we deserve to be shafted, right?
@@RottenFishbone i think they lack the facilities to do that
Funny, but the point of the video was kind of the opposite. It's math - not morals or any real division - that prevents us from having good representation.
But it doesn't have to be.
My question for Rated Voting systems would be the impact of the “integrity” of the negative ratings. If a bunch of people voted 10 for their person and -10 for every opponent, but the opposition didn’t go to FULL -10 for their less-preferred ratings, would it encourage people to just approach the system and vote 10’s and negative 10’s? If the people who disagree with you are giving your person -10, wouldn’t a lot of people just try to game the math by only rating with the extremes? Is that accounted for in some way?
There are plenty of other attempts to remedy (though not solve) this problem. For example the winning party could be required to get a true majority (51%). If they receive less than 51% of the vote, they have to form a coalition. They must pick at least one other party to cross the 51% threshold. As a side effect this also incentivizes them to maintain a semi-decent relationship with competing parties instead of going completely off the rails during campaigning. This system has held up very convincingly in the real world. While it isn't a perfect solution, it's a very good improvement.
You describe a parliamentary system which typically has many parties. The United States almost always has only two parties of any consequence, voting for a 3rd party is basically a wasted vote BUT it will bleed off votes for one of the two remaining parties.
In practice this turns into a SINGLE PARTY having actual control of a particular branch of government.
Yes, that's actually how real democracies work. Systems like the one of the USA are fundamentally broken. What you described is completely normal for most countries. If you have e. g. A: 35% B: 30% C: 20% D: 15% as an voting result, A and C can just work together. There is no need at all for a single winner. Therefore the mathematically problem from the video completely vanish. It's not democracy that is impossible. It's the Winner takes it all system that the USA uses.
@@Duconi "What you described is completely normal for most countries."
As you have noticed, the United States is exceptional, that is to say, not like "most countries".
"A and C can just work together."
In which case they might as well be the same party. This is what happens in the United States. It is a natural outgrowth of "democracy" and eventually it will be just "A" a single party rule.
@@thomasmaughan4798 A and C can have different positions also maybe over time a new party E comes up that fits better to A and gets popular. With voting you can therefore also define priorities. One party maybe has environmental protection as focus another on wants to improve the health system. They can work together but voters can vote for what's most important to them. If you just have 2 parties you don't have a choice to prioritise. You can just pick one of the 2 parties and take whatever they offer. It's not possible to pick a party with different solutions or that wants more than one of the 2 parties offer.
@@Duconi " If you just have 2 parties you don't have a choice to prioritise."
Many parties exist in the United States. There's typically 8 or so choices for President of the United States. But since there can be only one President; any vote for anyone other than the winner is basically wasted.
This is true for Congress as well. Having someone NOT a Democrat or Republican is rare. It is very expensive for someone of other parties, Libertarian for instance, to run for office.
What tends to actually happen is that rather than create a new party, an existing party is evolved or adapted to new priorities. Democrats have been captured by the "far left" and in 2016 the Republicans were captured by.. well I don't remember exactly, almost libertarian.
Each party is thus *already* a coalition of sometimes very different priorities. Typically there's only one or two top priorities; for Democrats it is "woke" and abortion; for Republicans it is "liberty" and first and second amendments. This can also be viewed as "herd" versus "not-herd".
3:58 HEXAGON IS DEFINITELY THE BESTAGON
Just on my way to comment that😂
For those of you who are curious, the "post" is a metaphor from horse racing. The finish line in a horse race is an actual post in the ground. It goes back centuries. Originally there were two posts and the race was "start at post 1" then "round post 2" finally "end at post 1". Once we invented ovoid tracks with gates to hold the horses, there was only the one "post" remaining.
But the metaphor doesn't fit as there is no "finishing post" in FPTP. You might need 50% of the votes to win or less than 20% of the votes.
@@MrXyzzzzzzzzzzzzzzzz Well that's a beef with someone else. I'm just here to help with the etymology.
Didn't expect such a status quo ending.
"Things might be sht but gosh darn it, it's what we have and stop thinking. Also it built this city."
CGP Gray covered this quite a while ago from the perspective of voting and how to resolve it. It's long been one of my favorite mini series. This is a fascinating continuation to me!
As a math PhD student I will clear up everyone’s confusion. So a lot of people are confused about the pivotal voter being a dictator, which even I was prepared to go on a rant about how the non-dictator rule was a terribly chosen axiom. I feel like this could have been explained a little better, but here I am to explain it for you guys. Upon further inspection, the reason the pivotal voter creates a contradiction is not because of it being the median voter, but actually because if you notice that it creates a loop again with B > C > A > B in that case, which is actually a transitivity issue. The non-dictatorship rule is actually just a way to accompany multiple rules into one. The actual theorem states that for any transitive and complete preorder that the following three things cannot happen at once: Pareto Efficiency (if a is preferred to b for everyone, then a is preferred to b overall), Non-Dictatorship (for any overall ordering there does not exist a single voter which if that voter prefers a to b then a is preferred in the overall ordering), and Independence of Irrelevant Alternatives (if all voters prefer a to b, then the preference of c should not change the overall preference of a over b). The issue is not that the voter is the median voter or the tiebreaker, it is that his vote has a theoretical potential to cause a Condorcet loop. Therefore, in practice it can be very useful, but can become detrimental if the specific conditions are met, which would throw the democracy into collapse (unless there is a secondary voting method in place to counteract the Condorcet loop, which in that case, we should have used that method instead in the first place). If you guys want me to make a UA-cam video explaining the issue in finer detail please let me know.
No thanks.
This explanation is brilliant and perfectly clarifies the issue
Yes plsss🙏🙏
You just confused me even more. Pretend you're talking to an AVERAGE voter who doesn't understand math. Use names, not letters. Too many (non-colored) letters. Basically, what math is telling us is the idiot undecided voters who don't even know what their own preferences are decide our elections? I knew that already, lol.
Yes, would like a video explanation.
Just in time for that mid night veritasium video
Where are you from?
@@sashkaulanov2 30 am in Pakistan 🇵🇰
07:52 Idk why this is a problem. After all, you select your preferential candidate. While Bohr might have given an unpopular speech/policy proposal, he is still the most popular of the three. And if people are voting Einstein from Bohr, they think Curie is bad/incompetent. Voting comes down to population's favourite candidate at the end. The second choice exists to say who people prefer anyway. So it's good all around. I dont see the problem. But I agree with the commentary about how politicians try not to stand out too much so as to win 2nd or 3rd choice preference.
I was Latif’s Uber driver a couple of months ago and we had a nice chat and he told me all about Radiolab. So cool to see him in this video!
“Is that my guy? That’s my guy!” Was my exact response 😂
He's a gem! Love his work on Radiolab
0:18 seconds in and you have an image of the UK’s unelected House of Lords 😆
Biden should be president for another 4 years, he should have mentioned Pelosi RIGGED IT.
@suspicionofdeceit Inheritance. They're members of the UK nobility.
@@atroxiv Not all of them. Others are appointed for life as/for political favors.
When it was inherited seats, most were "cross-bench"; they stood with more than one party depending on what stance had. Now that's it's nearly all "life-peers" it's a joke. They get their life peerage from being part of the "inner circle" of the governing political party, or by being a large donor; it's full on corruption.
We've gone from a review house that did the best for the long term (usually anyway) because their title and family line of that title depend on it, to one where they only care about their short term profiting (either monetarily or socially).
And its used again. Thats a rather large mistake!
Nice nod to CGP Grey at 3:39
His videos on voting systems were great.
YT recommending it on elections day in Brazil. Very excited to vote after this
Transitivity seems to rely on an assumption that comparison is made based on any one dimension. For example like number, if number a > b and b > c, then surely a > c
But political preference doesn’t not work like that. It is more like an aggregation of comparisons along multiple dimensions. Therefore preferring a over b and b over c does not necessarily mean that a is preferred over c (given that there are many dimensions to consider )
Most would agree if your preference is not transitive, you are not rational. While people do behave like that, it is easy to prove that with cycles (i.e. without transitivity), impossibility will arise in some sort.
3:33 - Hexagons are bestagons! CGP Grey has also covered this topic some time ago
Switzerland has it right. They hold referendums regularly to decide on policy. That's closer to true democracy than voting once every four years.
I have reservations about the Swiss results. I think it makes difficult decisions nearly impossible. Just this year they voted for adding 13th pension to the year, despite the government's recommendation to vote against it.
@@BakhtiyarNeymanyou either want democracy or competent governance. Those two aspects rarely overlap. A good democracy fundamentally necessitates a weak executive. A weak executive will never get any thing done in terms of policy implementation.
@@BakhtiyarNeyman It's what the people want and therefore democracy. I would agree that democracy isn't perfect as people are incompetent, but then you would first have to come up with a new system that is fairer than democracy, which doesn't exist (yet).
Democracy is bad. It is 2 wolves and a sheep deciding what is for dinner. You and this video creator arent very bright.
@@entreprenr1476 no that isnt democracy...
Love the CGP Grey reference at 3:35. Hexagons ARE the bestagons.
I like the approval voting with points from 10 to -10 for approval and disapproval
Check out S.T.A.R. Voting.
Score Then Automatic Runoff. It came about in the last decade and shows incredible promise.
But it becomes disingenuous. For example, I would vote -10 for the candidate i don't like and 10 for the one I do like, even if I'm really choosing my candidate begrudgingly. It's overly complicated and doesn't really add information more than a ranked choice would. Just my opinion.
The main problem in representative democracy is that the representatives start to have their own distint interests that are often at odds with the people they are supposed to represent. The main barrier to improving any voting system is that the people who could make the change are necessarily the people who have just won under the existing system
A proper voting system pretty much devastates the incumbent advantage. The method I've recommended uses a 7- or 9-winner nonpartisan primary for a reason: less than 15% of the voters can throw you out of office. With 7 candidates equally spaced across the voters on the single-dimensional ideal axis, if 15% of the voters on one "side" move the hypothetical winner down on their ballot, the next most similar candidate wins. They can't tip the election from Trump to Bernie, but they can tip the election from e.g. Biden to Obama.
@@johnmoser3594 The solution is much simpler. You have a parliament with e. g. 100 seats. Just distribute the seats matching to the votes. Meaning, if the people voted for A: 35% B: 30% C: 20% D: 15% the get seats as following: A: 35 seats, B: 30 seats, C: 20 seats, D: 15 seats. Problem solved. So now this group of 100 people is discussing new laws. If for example Party A and C want the new law, they together will have 55% of seats, so the law passes (except it's a law that requires a 2/3 or 3/4 majority). There is no need for a winner-takes-it-all system.
@@Duconi Your approach leaves several questions. What do you mean "seats matching to the votes"? Do you mean by party, where we vote for parties in which the elite aristocracy decides who will represent us and what the policies will be? That's strongly disconnected from the voters. Have a look at the US Democratic party, with people like Bernie Sanders, Ocasio-Cortez, and Joe Manchin, on varying parts of the political spectrum; the Democratic Party would exclusively choose people like Joe Manchin, internally. This kind of system gave rise to the Nazi party because it moves almost all democratic power out of the hands of the voters.
Do you mean by single transferrable vote with 100 winners? That's viable, but it's a complex process. Do you have geographical concerns? If we're going with a blanket STV election across the entire nation, it's technically possible for all of the elected politicians in the US to come from Baltimore City.
How do you deal with single-winner elections such as for governor? You've only got one winner. Runoffs are completely broken.
The process I described addresses each part of this by making the elected candidates reflect voters instead of party-elite-chosen aristocracy, both with multi-winner elections and single-winner elections; and by giving some degree of geographical control, with the electoral process completely eliminating gerrymandering (tweaking the district lines is no longer a free action: to obtain, something of equal value must be given up in exchange, so you can manufacture a moderate liberal district to get rid of a far left-wing district but only if you replace a strongly conservative district with a moderately conservative district in exchange; and that only works for single-winner elections, otherwise you're basically just moving the political ideals between districts rather than blending them to a smooth and creamy consistency of uniform ideology).
The people who can't/don't think for themselves outnumber those that do, meaning whoever controls the media controls those same people.
which is why whomever controls the media does everything in their power to portray the other side as controlling the media... USING THE MEDIA.
True! People react to what is in their faces right now - bills to pay etc. They don't realize by actually holding politicians accountable that they may get a fairer shake. In their desperation they turned to Trump who wasn't a politician, but they apparently had never heard of a POS 'business man' before.
@@Kr0N05 People turn to populists to shake up the system, voters are smarter than you think.
@@Kr0N05 It's always the other guys who are dumb, right?
Which is why we need Robert Heinlein's service-before-franchise system. Among other things...
Great video which shows, that a mathematical approach to look at a problem, which concerns society, manages to almost miss every point, that matters. Like the undemocratic influence, which big coporations have over politics. And the fact that those companies are arguably more important then our governments and are organized in an completly undemocratic way. Or that many people think democracy just means the majority or their representatives decide, even though respecting human rights are just as important.
Unanimity… with two i’s.
Ramon Llull being pronounced Rayman Lull, too...
the closest approximation to English phonetics and phonotactics is something like:
ruh-Moan Yoo-ee (GA IPA: /rʌ'mown 'juwi/)
You and I are wearing the same socks 15:13
I feel connected in an oddly intimate way.
How did you steal Derek's socks
8:10 Wouldn‘t people rather switch from Bohr to Curie instead of instantly switching to Einstein, leading to Curie not being eliminated?
not always, enough real life examples. Also not about the specifics, but mostly about having 25, 30, 45 changing to 31, 30, 39 and what could happen.
A lot of people are angry, and will vote for ultra right or ultra left and even switch between them on a whim.
There's an argument that the political spectrum is not a line, but a circle - that far left and far right are closer to each other than they are to the moderates in the middle.
Not necessarily. They say "the political spectrum is U, not a straight line. The ends almost meet". Many people in eastern Germany vote either far left or far right, nothing in between. The common denominator that's relevant for them? Anti-establishment.
Yeah that was a poor example and poor explanation of a real effect. If we imagine three candidates A, B, and C in a spectrum where A and C all preference B and B is 50/50 split, if you're a C voter you _can_ increase the chance A wins if you don't vote strategically for B. If the votes fall 40/25/35 for A/B/C respectively, B gets eliminated and A wins. If C voters are strategic and change the result to 40/35/25 now C gets eliminated and B wins. A better result for C voters by ranking their favourite candidate lower, failing the no favourite betrayal criterion.
In the real world this seems to be surprisingly rare because true centrist parties are rare. You would more often find B voters split their votes 90/10 one way or the other. And even the C voters might be split 90/10.
@@TheWhitePianoKeyProductions Why would it be an issue for 31, 30, 39 to result in a Bohr win? If the preference of the public is to go extreme over center, then extreme should logically win. It'd mean the public generally believes Bohr > Einstein > Curie.
love the cgp grey reference, he's got a bunch of videos on the same topic from a few years back that are great :)
15:00 I have watched this example a couple of times and I still don't fully understand it.
It states that if everyone ranks a particular candidate either first or last, society as a whole must also rank that candidate first or last. But as far as I understand the condition of transitivity was never broken, since the group as a whole didn't agree initially nor in the altered vote that either A>B or B>A (and C>B or B>C by extension).
What the altered vote changes is that there is an unanimous agreement that C is preferred over A (C>A). Changing A>C to C>A doesn't break the condition of transitivity and therefore the change in result from A>B>C to C>B>A meets all five conditions.
If there's something incorrect about my thought process here I'd really like to know.
I agree, the goal of the system is to elect one winner not to sort out the exact order of preferences among the remaining choices. from the example he gives it is clear choice c should be the winner, the exact order of b versus a is interesting academically but not really relevant operationally.
It's condition 5 that is broken. B is the irrelevant alternative.
Take out B, and C wins by unanimity. Introduce B and suddenly A wins by the rules of transitivity.
@@OneLine122 I don't see your point there. Why should A win? It seemed to me that the A>B>C was just a thesis, which in my opinion is wrong. If you make the thesis C>B>A there is no problem.
If you take B out, its C>A ( b>"C>A" on one side and "C>A">b on the other).
That means it doesn't matter whether B is in the vote or not.
@@OneLine122 My point it that the condition of transitivity is never broken, since, as I mentioned before, the group as a whole didn't agree initially nor in the altered vote that either A>B or B>A (and C>B or B>C by extension).
I was thinking the same thing
It's like a return to the CGP Grey voting system videos from 13 years ago.
Because nothing has changed and fptp is still horrible
May be that's why there's a secret nod to hexagons 😂
Well he's much much better than grey but similar yes.
The "bestagon" with glasses was a nice touch!
Yeah except CGP Grey convinced everyone (including myself) to be an advocate for Ranked Choice Voting...which is a trash system.
This video is at least an attempt at undoing that damage. Ordinal systems are trash. Cardinal systems are good.