Animated Voting Methods
HTML-код
- Опубликовано: 4 окт 2024
- In 2006, Ka-Ping Yee introduced a way to examine single-winner election methods via computer graphics (see: zesty.ca/voting.... Each colored circle represents a candidate in a 3- or 4-candidate election in a two-dimensional political space. The color of the background represents which candidate would win under each method if a randomized electorate, centered at that point, were to vote.
Yee's diagrams show some serious pathologies with the plurality and instant runoff methods, but it is unclear from his descriptions whether these were frequent occurrences, or chosen specifically to make those methods appear worse than they are.
This video follows on Yee's work by animating the positions of the candidates in the two-dimensional political space, and adds Score Voting, Star Voting aka Score Runoff Voting and a one-voter "ideal winner" model. Where plurality and IRV tend to squeeze out candidates in the center, Score tends to give an advantage to candidates who are positioned in between other candidates. Score Runoff Voting (aka Star Voting) consistently performs closest to the ideal model of the systems visualized.
Simulation Notes:
Each pixel represents an election of 4,096 voters in a random radial Gaussian distribution around the "center of public opinion" at that point. Voters vote using the following rules:
Plurality: vote for the nearest candidate
IRV: rank the candidates in order of increasing distance
Score: score the closest candidate 5, the furthest candidate 0, and the others scaled along that spectrum
SRV: score the closest candidate a 5, the furthest candidate 0, and the others scaled from 1-4 based on distance between nearest and furthest
One voter: pick the candidate closest to the center of public opinion
The source code for generating these animations is available at github.com/nar...
this deserves waaaay more views
If you want to tweak this video:
1) change your shirt. I want to be able to educate people, including kids.
2) slow down the animation; call out more specific cases and explain why IRV is wrong in those cases
3) explain what you mean by correct winner and why that is right
4) Y no Condorcet? Knowing how it works, it seems like the only reason it was left out was that it would get this simple test perfect with no flaws anywhere and make STAR look bad (which it isn't). Picking out a metric for comparison and then not using the perfect winner by that metric doesn't engender confidence in anyone who knows about that winner. So, please point out some of STAR's benefits that would bring it over the top vs something that very slightly outcompetes it by this metric.
I agree with your 1-3. As for #4, I suspect the reason is that there are very many voting methods that are being studied, and for a short and effective video you need to focus tightly. "Pure" Condorcet is not usable since there are elections that don't produce a Condorcet winner. There are several Condorcet-based methods that have different fall-back mechanisms. These are worth studying and some seem to do very well in many situations, but it's difficult to imagine any of them being widely adopted in public elections because they are a little more difficult to explain.
I really like your question #3 because his "correct winner" method is not-at-all obviously-correct. I don't see why one the candidate favored by one voter at the center-most position among all voters best addresses the needs of the whole electorate.
Terrific job in describing the Ka-Ping Yee graphs. I might create a longer version of this video where you go into more detail as to what is going on. But in 6 minutes you made the point very effectively.
I'm convinced STAR voting is one of the best systems for single-seat elections. What I would like to see is a similar scoring system that can be used for multi-seat proportional representation.
We had some late night musings at equal.vote/pr -- but not sure it's ready for prime time. Recommend checking in with the Election Science Google Group.
Thanks!!
Check out the Fair Representation Act.
This is a nice video for looking at the comparison between instant runoff and plurality, but those Yee diagrams are more or less generically meaningless for score-based systems. You indicated how scores are being distributed as a function of distance, but altering that function can have dramatic effects on the Yee diagram.
I really hope star voting gets Implemented In the West and Europe.
US does need STAR, yes, since its federal structure relies on so many single-winner positions.
Many European parliaments have proportional representation so STAR isn't appropriate there, but we do have a lot of positions that currently aren't democratically elected at all, like kings, queens, and EC commissioners, and it'd be great if STAR voting was used for them.
We need in India also!🥲. But its very unlikely to be implemented as the current powers will be silent even if they knew the advantage to the people.
Great video! The animations are incredibly informative! And holy crap, IRV sucks so much.
By the way, from my experience explaining the benefits of SRV/Star to others, I'd like to add that "center of public opinion" doesn't mean "politically centrist" positions. A lot of people get this wrong and criticize the arguments made by us by saying the system will elect "lukewarm, weak politicians". (Pretty messed up conclusion, too!)
If the center of public opinion is on the extreme left of the 2D space, then that clearly is not a centrist position. A centrist position would be a the center of the 2D diagram, the center of the 2D political spectrum.
The idea is that we're electing candidates in the center of the population's overall opinion. This "central opinion" is the actual the center of the majority we should be striving for in voting systems, the overall consensus of society. SRV/Star selects it beautifully.
Keep up the good work, guys!
(By the way, is Star Voting the preferred name now?)
Agreed -- the one bit that was interesting in animating Score Voting is that, because we assume that voters will use the full range, the "lukewarm, weak politician" in the center of the other candidates actually does gain an advantage in that system. IRV favors extremists, Score favors "centrists" and Star Voting favors accuracy :-).
Exactly! STAR really is brilliant in finding that sweet spot, and these animations show it beautifully.
They don't show one thing: whether the choice is right also for people who don't vote. Taking that into account would make the visualization a cube.
AstralStorm There’s no way to determine that with just the information provided by a vote. That problem could be solved by making voting mandatory.
I wish there were different words for "centrist on an absolute scale", "centrist relative to the voters", and "centrist relative to the other candidates".
very amazing illustration on how STAR Voting outshines all the other models - thanks for sharing!
The jagged edges seem to signify a discontinuity of voting behavior that might indicate opportunity for strategic voting if I'm understanding what going on correctly. I need to find a further explanation of the distortion of score voting.
Yes, they do. But the idea is that they are much less common than in other voting systems in realistic scenarios. Arrow and Gibbard proved that ALL voting systems have opportunity for strategic voting. There's no way to avoid it completely.
I don't think they proved that for the cases in question. www.rangevoting.org/GibbSat.html
I don't think I've fully grasped the nature of the issues. Distortions of the method may not enable tactical voting.
@@aicram62 I'm familiar and share election science often. I'm currently looking at some distortion when a reduced score like 5 points with 5 candidates the spread can result in a problem and star doesn't resolve but hypothetically approval results in better outcomes for that situation with some cut off assumptions.
@@aicram62 thanks I'll try to get in there.
Yes, but the distortions are small enough and the ballot is expressive enough that practical strategic voting in real political elections, particularly large ones in the US, is effectively impossible.
This video is great and 5 minutes and 17 seconds too long for public memetic consumption. Would like a 60 second no frills version of the salient points to share en masse
That's on the to do list. In the meantime this is a great short video to share. ruclips.net/video/3-mOeUXAkV0/видео.html
What a beautiful animated explanation! So much better than the original!
As others have mentioned, I'd like to see the option of Approval voting (aka 1-star-only STAR voting) simulated alongside. I think it has value for its simplicity, and so if it performed almost as well in most cases, that might make it a worthwhile choice.
The same way Score Voting better maximizes utility as you allow for more expression (e.g., 100 "stars" produces better results than 1 "star"), so does STAR (aka Score Runoff Voting). That said, Approval Runoff Voting beats out traditional Approval Voting the same way STAR beats out traditional Score.
I’d be interested to see how these voting systems hold up when there are more than 2 dimensions of political space
Yes, this. In reality, if we improve voting to discourage strategic voting and, reward sincere voting, the public happiness will go up as people feel their honest votes weighed against their communities. Over perhaps 10 years or so, it will kill two-parry dominance and the divisiveness that it brings. The political spectrum will expand dramatically, and problem solving will be valued above all else.
So I would like to see successive iterations to the models over many voting generations, as the electorates mature and there are no longer ANY dominant parties.
This video needs more views.
Unfortunately the casual profanity makes it less shareable to normies.
Heh, nice shirt!
Also I hope they keep SRV instead of STAR. Without anything in front or behind it Star sounds weak. That's why you see Northstar, 5 Star, Starpointe and so on. Score Runoff just flows off the tongue nicely as does SRV.
Glad you dig the threads! We'll have them for sale eventually :-).
Star Voting came about because folks new to the system (lay voters) consistently got it wrong, and confused it with ranked choice. That said, we are keeping Score Runoff Voting and SRV around as well.
Feel free to call it 5 STAR Voting as well. The name change came about in part because 5 STAR Voting describes the 5 star ballot or 5 star rating. Aside from folks new to the system, we also had a major problem with folks experienced with the system (our own steering committee) routinely bungling the acronym. RCV, IRV, STV, SRV, etc.
I find the constant movement of the visuals to be very confusing. I can't follow it at all. Can't you give some examples in a less distracting way?
That's good feedback. Thank you. In the meantime you might like our articles better. There are a number of them that compare voting methods. starvoting.us/articles
Thank you for your work
Nice! Would love to see how Approval Voting and Ranked Robin perform.
Anybody concerned about politics should understand this video. There's no better way to improve the lives of millions than updating our old voting system to STAR. Polarization has consumed every area of life over the last few decades and we're all sick of it. Our problems lie in emotional thinking, and the solution is with MATH!
Ranked choice is a type of voting, but instant runoff is a method of tabulation. Ranked ballots are probably the best way to vote, but there are lots of better ways to determine a winner from that data.
Those graphics are amazing? Can you maybe add a video about how Star compares to condorcet methods?
Smith//Score is a condorcet method better than STAR, but STAR is easier to explain.
Condorcet methods (or at least the better ones) produce results that are best in class and quite similar to STAR, and they don't exhibit center-squeeze or center-expansion bias. In Yee diagrams (which use 2D models and which don't include Condorcet cycles) Condorcet voting systems with honest voters always produce the optimum diagram. rangevoting.org/IrvExtreme.html
So recently I've seen the light in terms of increase of strategic expression in star over score. I'm not following the center expansion of splitting in score voting or seeing where that would result in a regretful choice.
I'm still thinking there should be a better way to get strategic expressing it's just not simply from base score voting.
It's an open question whether we should want to motivate voters to differentiate similar candidates. I'm more concerned with candidate behavior than voter expression.
I see no problem with score maximization (rather than motivated differentiation). In fact, I'd rather give no advantage nor backfire potential to tactical voters. Approval limitations may be best for voters and candidates (democracy) in total.
The center expansion in Score voting isn't that bad and Score doesn't have vote-splitting. The issue with Score Voting is as you said, there are strategic incentives to exaggerate your ballot that are fixed by STAR voting. STAR actively incentivizes voters to show their honest preference order and honest level of support for the candidates.
But keep in mind that Score voting still delivers pretty accurate results (better than RCV and Approval) even if voters are strategic. Score also doesn't have a strong center expansion or center squeeze bias. It's a good system, and if combined with a top two general election, it's basically up there with STAR.
I want to see my system in there, but find it hard to find the time to add it to the code.
A way to look at a class of systems is to see that they try to eliminate some of the worst candidates first and then they have an automatic runoff. STAR says, let's eliminate the N-2 worst candidates first, and it uses Score to do that. An alternative would eliminate the one worst candidate in the first round and then compare the remaining N-1 candidates in a final round. IRV tries to eliminate the single worst candidate in each round; it just uses a wrong procedure to decide who is "worst".
I think you are talking about coombs method.
At coombs method, you use first past the post to check if there is a winner (more than 50% of "first place" votes) and if there is not you use anti plurality (who has most amount of "last place" votes) to decide who will be removed before the next round.
Anti-plurality system keep non famous candidates because they are unknown and so people won't have enought information to say they suck as hell and so say they are the worst. You have many guys saying hillary is the worst thing ever and other saying trump is the worst thing ever.
At coombs method, all ranks that aren't the last one, are considered the same for the sake of deciding who will be removed, if you have two guys candidate A with 1 last rank vote, and 99 second to last vote, and candidate B has 98 first rank votes and 2 last rank votes. Candidate B will be the one removed before the next round.
@@exedeath No, not Coombs.
I've been thinking on this some more. I'm just wondering what approval then ranked choice would look like. If people only ranked those they approved of.
My current worry is that limited to 5 scores doing double duty will distort scores for the purpose of ranking. And distort ranking because of the limited number of distinctions.
Some IRV jurisdictions only allow you to express an opinion about your top-3 favorites. This is certainly better than that.
@@eyescreamcake True. But that is a voting machinery problem that we RCV activists have been working to fix for awhile. And it should be fixed by 2020.
The 0-5 range was selected based on research about cognitive load. People are limited to about 5-7 different things they can hold in their mind at once and compare to each other. Having 6 different ratings covers that pretty well and “5 stars” is super marketable.
STAR and Score suffer from similar (not identical) candidate splitting. The AR motivates vote differentiation and would likely encourage negative campaigns between similar candidates.
Approval has a much more exaggerated harsh motivation for differentiation. Rather than slightly like/dislike, Approval voters are much less likely to (but can) completely betray similar but lesser favorites.
Are Score and Approval exactly the same with respect to Yee diagrams?
No, Approval voting has less resolution so it's not the same as Score voting.
@@sarafwolf true, but in practice and in aggregate the results are similar.
At least in the real-human mock elections I've seen, AV and Score have indistinguishable results in the extremes (least and most popular candidates) with mid-popular candidates acquiring higher scores than approvals. Which I think is what we'd expect from a binary threshold vs "higher resolution" in between. Score, IMO is just easier to explain to the individual voter.
Please compare your proposed STAR (where scores of 5 are actually counted as 5 votes in the first scoring step) to a STAR-system, where the scores are divided by the number of voter scores.
What about Approval Voting??
Approval is score voting with one star. Was not really looked at specifically but would be a better candidate to implement than star/score/RCV for simplicity and I believe has very low "regret" possiblity
Edit: assuming you probably know more about that then I do actually but just wanted to point out approval is a version of score voting
To get a really good simulation of approval voting, you'd need to define a radius of acceptance for each voter - candidates inside get a vote, candidates outside do not.
What causes Score's distortion? Utilitarianism vs Majoritarianism? People scaling their scores so min is the worst candidate running and max is the best?
"People scaling their scores so min is the worst candidate running and max is the best?" - yes.
So if you used "true" scores (e.g. if you used the distances as scores), that would result in a Voronoi diagram?
Yes, "true" scores would look quite close to Voronoi. But people do not vote like that, so full range ballots seem more realistic for a comparison.
It would be good to show the effects of different strategies, though.
I ask because IIRC Condorcet methods always result in the ideal Yee picture, but IMO the Condorcet winner isn't always the (utilitarian) best winner, which Score should be able to find sometimes. So Score should look different from the ideal sometimes, and be better as a result. Unless this gets into the weakness of this simulation, which would make sense because cycles are impossible here.
Can you compare systems where multiple seats are given to the voters?
I feel like this is a hybrid of IRV and Scored. Why not just say that? You are still ranking to a degree but it is your level of satisfaction that is being ranked and then a run off.
They do say that on their page. "STAR Voting is the new and improved hybrid of RCV and Score Voting."
Even though I'm an IRV supporter, I Liked this video because it was well made and interesting, and encouraged me to take a closer look at STAR voting. Thanks for putting it up.
BTW, had I been in Oregon for this vote, I would have voted to try it out. It deserves a look. I think it would have succeeded at first, but then failed over time due to strategic voting, but still, it deserves a try-out somewhere.
jerel42 check out www.equal.vote/strategic-star . Most strategies proffered by skeptics can be shown ineffective, even counterproductive in STAR.
@@equalvote It doesn't really matter whether a voting strategy is counteproductive. Like gamblers, people will believe in voting strategies and use them, and distort the results in random ways. The same applies in IRV, where to use an effective strategy you must know how all the other voters will vote, and they must not predict what you will do. But people are likely to try to do strategic voting even though they don't have the information they'd need to get results.
to me its important to expose a list of platform issues to the voters and let them decide on direct issues instead of relying on candidates who often fail to follow through on their putative platform as outlined during their campaign ... is the topic worthy of a video deep dive ?
I like the video. The takeaway is just how bad IRV can be -- it would be worse than simply staying with plurality. I would like to eventually see comparisons with 3-2-1 and majority judgment.
LOL. If you look at the variation graph IRV would be slightly better than plurality.
Could you add Approval Voting to the analysis? I like STAR Voting but it feels susceptible to errors based on people voting strategically rather than honestly. Approval is not immune either but does seem less susceptible. These simulation models also clearly focus on policy but people often vote for the candidate more directly than their policies.
This model is only based on single member districts correct? Do some work better in Multi Member Districts?
I've never been a fan of range voting but the graphic representation is very convincing. Unfortunately I wasn't able to follow any mathematics of what was happening, just the conclusion. Hopefully a slightly more detailed breakdown could help with the explanation.
I found it interesting that runoff made Plurality even worse but made Score even better. Is there an explanation for that?
What concerns me about score voting is not the math obviously, but the human element. I see it devolving into a kind of self-imposed plurality voting, if people are afraid that giving a less popular preferred candidate too high a score could lead to the more popular but tolerable candidate losing to the least preferred option, so they just give 5 stars to the red team or blue team and be done with it.
If voters are strategic, Score devolves into approval voting. I expect that STAR does close to the same thing, although the runoff component would make optimal strategy slightly more complicated.
Could you run the same simulation given different voting distribution, say, gamma distribution? This would model a situation where there are voting blocs.
Additionally, what happens if you quantize the distributions or reduce numbers of those votes, which would test participation criteria.
(I can modify the sources myself to do that too, just spotted the link.)
What happens if we apply the method(s) to multiple winner situations? Both a "vote score" cutoff and hard number of winners?
Indeed, the source is available in the info block above. In general, this visualization is intended to test best-case performance with a simple distribution. More advanced simulations such as VSE (electology.github.io/vse-sim/VSE/) include much more complex voter distribution models.
I'm interested in voting blocks, sldo how parties play into this.
This method of analysis doesn't take into account someone liking polarizing candidates
This isn't clear at all. Some actual examples - especially when it comes to IRV - would be very helpful.
Model Approval Voting' please.
My biggest concern about all weighted/scoring systems is strategic voting. I often recommend weighted/scoring systems for non-public elections, but for public elections people shamelessly vote emotionally, they often vote strategically, and they often just follow the instructions from one campaign or another. In general scoring systems are great for non-emotional elections, and look great in mathematical modeling, but don't perform well in highly emotional, real world, messy, public elections.
For public elections, STAR clearly attempts to mitigate the damage of (inevitable) strategic voting, such as bullet voting, by using the runoff at the end. I think this is really clever, fun, and interesting. I can see why people would get excited by it.
But are you claiming it is no longer vulnerable to fairly simple strategic voting? I haven't studied it enough to know the answer to this question yet, but I would guess it could be easily defeated by real world strategic voting.
I think the only practical solution to strategic voting is to make it clear to the average voter that it is too difficult for them to successfully calculate a strategy. And I think that with the 4 or 5 candidates you might expect to see if every race was IRV (or STAR, etc.) plus the uncertainty inherent in even the best polling, strategic voting would be functionally impossible for almost all, if not all, people.
In general, strategic voting merely compensates for whatever flaws are present in a system. With plurality or RCV, strategic voting can help overcome the center-squeeze effect, etc. The real problem is getting voters to vote strategically and do it well.
There are systems that break down with strategic voting, such as Borda and Coombs, but they tend not to be used for that reason.
Can you model the following two voting methods?
Method1: Each voter is allowed 3 approval tokens and 1 disapproval token. All tokens must be played against unique candidates.
Method 1a: Each voter must spend all their tokens, but can 'stack' approvals toward one candidate if desired.
Method 2: Each voter is allowed 100 pennies. All 100 pennies must be spent. Voter is allowed to apportion their pennies unrestricted.
Let's say 5 people were running for mayor: Abel, Baker, Cain, Donetilli, and Everton. I like Donetilli best. Wouldn't I be maximizing his chances to win by voting Donetilli = 5, and everyone else = 0? Or, how about Donetilli 5, and everyone else 1?
jerel42 the two highest scorers are finalists, so while you might help Donetelli become a finalist, you don’t actually help him win the runoff any more by bullet voting than by giving differentiated scores. It’s also a risky strategy: if Donetelli is not preferred by a majority, then your bullet vote will have promoted a weak candidate that will then lose the automatic runoff, and you’ve given up the ability to express a runoff preference between any other candidates. Not a good strategy.
@@equalvote I don't know...........this feels a lot to me like Bucklin, which eventually devolved to bullet voting. Voters often have one favorite candidate and if asked to bullet vote by that campaign, will do so. My gut is that this would happen with STAR voting, also. I'd still vote for it over Plurality Voting, though. Plurality Voting is a truly horrible system.
@@jerel42 STAR and Bucklin are very different. In Bucklin, if no candidate gets a majority on first choice votes, second choice votes are added to the total, and then third place votes and so on. This creates a huge incentive to bullet vote, because in successive rounds, other voters will help your favorite win, while you won't help theirs.
STAR, on the other hand, is a single count, and there are two finalist spots. Bullet voting doesn't help your candidate AT ALL in the automatic runoff versus giving other candidates token support. Yes, you may artificially promote your candidate to the runoff by bullet voting, but that's dangerous -- if your favorite doesn't actually have majority support, bullet voting will promote your weak choice to the runoff, where he or she will get clobbered by the other finalist. And if your favorite doesn't make the runoff at all, you've given up any ability to choose between any of the other candidate pairs.
The bullet voting argument is a non-starter with STAR.
@@equalvote If all you care about is helping your favorite candidate win, then you should bullet-vote in any system. If he wins, then great! If he loses, then too bad.
If you have a second choice, then you have a trade-off between helping your second choice win if your favorite loses, versus helping your second choice win against your favorite. IRV does not help your second choice at all against your favorite, but if your second choice loses before your favorite does, then it's too late to help him. That's a trade-off. Other systems leave you with other trade-offs. If you think the second choice is almost completely as good as the favorite, then the problem is minimized.
Equal Vote Coalition What is the disadvantage of promoting your favorite weak candidate to the runoff where they will at least have a chance?
4:15 ("Center expansion") Score looks to be what we want (with no "correction"). Yellow and Orange are closer to the center of the electorate. "Dark blue" (in the corner) should not win! Thus STAR's correction is an error.
So If I got this right your have mathematically proven funktion here, that solve the voting system problem. What are the (honest) best arguments against your favorite voting system.All you have to do now is convince the 2 dominant parties to give up there power forever; good luck ;) :)
Well, that is the paradox of being an electoral reform activist. That's not unique to STAR voting. It is truly hard, *but not impossible*, to get electoral reform. RCV activists have got it adopted in many cities and now a whole state, women did win the right to vote, Black folk did win the right to vote, etc. So it can be done, it is just truly not a struggle for the faint of heart. :)
Anybody can come up with their own criteria for who ought to win elections, and come up with simulations showing that various voting systems fail to get the results they want.
There is nothing particularly valid about that. It's all personal opinion.
IRV is the only voting system which gets the right result, because the result that IRV gives us is the one we should want. Or maybe you might have some other opinion. But it's just opinion.
SO MUCH THIS. I felt flim flammed watching this honestly. And here is another issue. Voter confidence. If a reasonably competent voter cannot process the results themselves, the populace will always distrust the results. RCV and variations have been used by organizations under RONR for well over a century and can be done simply with piles of paper that most people easily grok.
>"IRV is the only voting system which gets the right result, because the result that IRV gives us is the one we should want."
That's such a ridiculous, tautological statement. Score/STAR/Approval/Condorcet voting supporters at the very least make an effort to explain *why* their methods are appealing and what goals they are trying to achieve: they maximize some notion of consensus that exists in society by aggregating the opinions of everyone.
Your statement is like "well, who cares what we want with elections? whatever IRV selects is the correct thing we actually should want, because that's what I believe in". It's meaningless. It's as if the system wasn't supposed to serve us, instead of us serving the system.
IRV actually is deeply anti-democratic and anti-social, as the ballot is as stubborn, petty, uncollaborative and factionalist as it can be. The ballot literally means "I want my favorite and ONLY my favorite. You'll have to take it away from me by force! And then, I'll want my next favorite and only them." (and so forth). And this is not my opinion. This is literally what the ballot means and what the elimination procedure of IRV is doing with the preferences. It is exactly the same kind of ultra-aggressive over-imposing behavior that is deemed "problematic" in other systems when it causes strategical manipulation, but under IRV everyone is assumed to be doing it at all times. This is why IRV is called "strategically resistant", but it is in fact just "strategically saturated". If everyone is being as aggressive and imposing as possible at all times, there's no room to be *more* aggressive and imposing.
And to highlight the problems with IRV even further, the logic of the system only makes any sense if you assume the population is completely divided into exactly two large factions which are maximally opposing of one another. Every candidate is just "noise" within one of these two factions, and the mathematical goal of IRV's algorithm is to ignore that structure and ensure the bigger of the two factions wins. It completely internalizes polarization into a two-faction system, rendering any extra options/parties merely symbolic proxies for one of the two factions, which will now be seen as "coalitions". This is also not opinion, it is mathematically what the system is designed to achieve, and empirically what it has achieved everywhere it has been used in the last century. Even pre-election polls reflect this two-faction narrative: look up "two-party preferred vote".
None of these are goals to strive for in a democracy, let alone a civil society where elected representatives are supposed to act on the behalf of everyone. IRV is a terrible system if the goal is to improve the political landscape and its representation, and challenge the status quo.
@@1ucasvb Yes, it is tautological, as are all voting system choices. They all come down in the end to "This is the version I like because I like it."
I can explain some of why I like IRV, but of course it turns out that the reason I like my criteria is that I like them.
IRV is one-voter/one-vote. Just, you get one vote at a time. You can vote for as many candidates as you like, but your second choice doesn't take effect until your first one has lost. So if your best candidate is one that you think will probably lose, still you get to vote for him and him only until he HAS lost, and then you get to vote for your second choice.
Sometimes people want to vote strategicly. Instead of voting for who they want, they vote for something else that they think will get a better result than their honest choice would get. For example, with FPTP their best strategic result is to guess who the top two candidate are who are most likely to win, and vote for the one of those two they dislike less. Any other vote is at best wasted. With IRV strategic voting mostly does not work. You have to know in great detail how everybody else will vote to make it work. If you try to vote strategically, a small difference in votes for the fifth and sixth candidate can mess you up. Better to just vote for who you want, in the order of how much you want them.
"The ballot literally means "I want my favorite and ONLY my favorite."
No, it doesn't. That's FPTP. If you want your favorite and ONLY your favorite, then you should vote for only one and nobody else.
I don't follow your reasoning at all. It makes no sense for IRV. It sounds like you have IRV mixed up with some other voting system entirely.
WOW
how about choosing the best deccision maker so that we factor complex issues that the masses dont understand?
ok
This is interesting, but it assumes that the winner should be closest to the center of the distribution. Often that is wrong. So the results are wrong.
That doesn't make any sense. Why would it be good for the winner to *not* be the best representative of the electorate?
@@eyescreamcake I expect that sometimes it's better for the winner to be close to the center of the electorate, and sometimes it isn't.
"Falling between two stools."
The donkey equally distant between two piles of hay, who doesn't go in either direction.
The national leader faced with war, who can't decide whether to fight or surrender and does neither.
Sometimes it's better to be in the middle, and sometimes it isn't.
Sometimes it's better to satisfy 51% of the voters than to satisfy none of them.
Since it can go either way, we shouldn't judge voting systems by how often they choose the guy in the middle instead of either of the ones who have support. Or if we do, the voting system needs some way to judge whether the guy in the middle "many people's second choice" is better than either of the front-runners. So it will choose him when it ought to, and not when it shouldn't. We'd need some way to guess whether its method to determine that is right.
The author makes claims that (while true) seem worded to mislead. For example in RCV. "the second choices of only SOME of the voters are taken into account". This seems to propose a condition of unfairness and is presented as an argument against RCV. It matches a lot of half-truth-half-speech that dishonest people use to argue a point only for the sake of self-interest. In fact, as SOME of the voter's 2nd choices are accounted, there is no reason for the OTHER voters to abandon their 1st (still-in-the-race) choice. Yeah... this author is imparting only baseless opinion. Pretty graphs dance meaninglessly before you like so much hand-waving. Dependence on the pretty-picture effect is a disingenuous trick. The same kind of communication method I see in snake-oil salesmen and the establishment. Sorry, the warning bells are going off on this guy. All opinion. Absolutely No data.
Believe what you wish. You have the permission of me and my panties. He does a poor job of selling his argument to anyone who does not already want to buy his goods.
You are defending a bad counting algorithm. You make it sound like the voters have agency when you say "there is no reason for the OTHER voters to abandon their 1st (still-in-the-race) choice." The race is over once the counting begins -- none of the voters in IRV have the opportunity to abandon their choices or not. As a result, voters who vote honestly can get badly punished in IRV, when, for example, their second choice gets knocked out before their first choice and then their third choice goes on to beat their first choice. Because IRV has this very bad feature, supporters' statements such as "With ranked choice voting, you can honestly rank candidates in order of choice without having to worry about how others will vote and who is more or less likely to win." (FairVote) is blatantly false. This scenario actually happened in the real world in Burlington, and it led to the repeal of the system: www.equal.vote/burlington
Only choice and decision are over once the counting begins. The race continues with every iteration of count ->evaluate->compare->eliminate->recount. There is no need to abandon or not. Do you suggest someone will complain "Hey my selection holds enough of the votes on the 1st iteration to gain more votes in the next count. How dare they not let me abandon my 1st choice." You seem to argue there is some unfair action because only people who's vote falls to the lowest counting tier get to have their vote transferred away to a 2nd choice, or 3rd. And, how does one vote dishonestly? They play 3-dimensional chess with the voting system, thinking they are going to control the outcome of a population of human beings. They make their selection using Princess Bride's, Vincini logic. They, in effect simply become noise in the system. The concept of voting is meant as a means to vote FOR someone, not against someone. It seems like the Burlington vote actually worked fine. Sure some may not like the results, but it turned out in a pretty logical fashion. Progressive, Dem, and Republican left standing.... Dem tossed... votes went to progressive. You can try to compare results using head-to-head match-ups, but your human variables are going to make any such exercise ridiculous. Would people still vote as they did using another system? Any effort to repeal the system seems more a function of the Dem-Publican cash-machine objecting to a 3rd-party, Progressive victory.
Again, you're simply trying to defend a bad algorithm. The voters vote, then those votes are tallied using the voting method, which is just a counting algorithm. We all recognize that the plurality method is a bad counting algorithm because it is subject to vote-splitting and produces non-representative results. What this video shows quite clearly is that IRV is also subject to vote-splitting and produces non-representative results. If you are willing to defend the Burlington result, where IRV failed to elect the beats-all winner, who was also the only candidate in the race with any kind of a majority win, and where the result of that election resulted in the repeal of the system, it'll be tough to have a reasonable discussion.
Thanks M T. Really appreciate you leveling up the discussion. Here's another video that's shorter and uses smaller words that might help you get the concept a little more clearly: ruclips.net/video/aiQ9Z5sME00/видео.html
what of younfool the voters by goving them a short term desire but not the long term (hard to understand) benifit. thendumb voter scenerio is almost always a factor because life is more complicated than elections.