Nicky Case's work is outstanding—they make lots of things that sit on this delightful line between "games" and "education". All of which deserves posting here, but in case you haven't seen much of...
Nicky Case's work is outstanding—they make lots of things that sit on this delightful line between "games" and "education". All of which deserves posting here, but in case you haven't seen much of it:
I first was introduced to their work by way of Coming Out Simulator, which is a little interactive story of figuring out how to come out to extremely conservative parents.
They also made The Evolution of Trust which is an interesting crash course on game theory.
How To Remember Anything is an introduction to spaced repetition, which has really changed how I learn things (I now use the Anki app daily), and I would have massively benefited from this in high school/etc.
I've been a fan for ages, but somehow never came across their memory one! Thanks for linking it; I might check out the app, too, since I've found this sort of thing hard to put into practice on my...
I've been a fan for ages, but somehow never came across their memory one! Thanks for linking it; I might check out the app, too, since I've found this sort of thing hard to put into practice on my own.
This uses interactive diagrams to show not only why FPTP is fundamentally flawed, but also, why ranked choice mechanisms (such as IRV) are also similarly flawed. But, it ends on an optimistic note...
This uses interactive diagrams to show not only why FPTP is fundamentally flawed, but also, why ranked choice mechanisms (such as IRV) are also similarly flawed. But, it ends on an optimistic note by showing that a better way is not only mathematically possible, but also quite easy and within our grasp with an endorsement from Dr. Arrow himself. You'll see that Arrow's Impossibility Theorem isn't as as doom and gloom as it may seem.
Did Ralph Nader Steal the 2000 election or did the Florida Supreme Court? God forbid we have some semblance of choice in American elections lest the establishment candidates do poorly. Won't...
Did Ralph Nader Steal the 2000 election or did the Florida Supreme Court? God forbid we have some semblance of choice in American elections lest the establishment candidates do poorly. Won't someone think of the poor political establishment and their candidates for once?
That said this was really interesting even if the author's understanding of American elections isn't the same as mine.
The only reason there are "establishment" candidates is because of this voting mechanism. FPTP mathematically ensures that there will be only two major parties no matter how anyone votes. Any...
The only reason there are "establishment" candidates is because of this voting mechanism. FPTP mathematically ensures that there will be only two major parties no matter how anyone votes. Any third choice can only play the role of the spoiler and both the original party and the splinter faction will, over time, realize it's in their best interest to reconcile as their mutual opponent wins over and over.
I have a problem with this scenario in the "How are the alternatives?" section: In this scenario, Henry Hexagon is politically near to Tracy Triangle. Therefore, as we move the voters closer to...
I have a problem with this scenario in the "How are the alternatives?" section:
For example, here's a sim of Instant Runoff Voting. In the beginning, Tracy Triangle is already winning, and you're going to move the voters even closer to her. Obviously, if a candidate is already winning an election, and becomes even more popular, they should still win afterwards, right?
In this scenario, Henry Hexagon is politically near to Tracy Triangle. Therefore, as we move the voters closer to Tracy, we also move them closer to Henry. However, Henry is also slightly nearer to the political centre, and therefore slightly closer to the centre of the voter group, no matter where you move them. That's why he attracts slightly more votes than Tracy and ends up winning the election.
If, between two similar candidates, the voters end up voting for the one whose policies they like more... why is that a problem? If they like Henry's policies slightly more than Tracy's, so they vote for Henry slightly more often than Tracy... why is that a bad outcome?
There seems to be some bias built in here, assuming that there are only two "winning" candidates and any other candidate must, by definition, be a spoiler candidate - without stopping to consider that maybe the other candidate is actually better overall in the electorate's eyes.
Imagine there was a left-wing candidate and a right-wing candidate, and both candidates were equally popular with their constituency and equally unpopular with the opposing constituency, and then a less-popular centrist candidate entered the field. In Instant Run-off Voting (or "preferential voting" as we call it here in Australia), it's very likely that everyone voting for the left-wing or right-wing candidates will put the unpopular centrist as their second choice, ahead of that other candidate from the other wing - and this might end up winning the centrist candidate the election. But... how is this a bad thing? Sure, noone got their first choice, but everyone got their second choice - and everyone avoided that other choice that they absolutely didn't want. Isn't that good?
I'm not seeing the problem with this scenario. But, then again, I've been raised in a country which uses preferential voting, and I know I'm biassed towards it - especially when I compare it to first-past-the-post voting.
That said, I know it's possible to "game" preferential voting. There are so-called "preference whisperers" in Australia who go around various minor parties and independent candidates, and get them to make deals with each other to "swap preferences" (you print how-to-vote cards for your supporters, telling them to put the other candidate as #2, while the other candidate does the same for you). On the basis of these deals, and our extra-strange voting system in the Senate - which is not straight preferential voting - we've had some very unexpected candidates get elected, including people who got less than 2% of the first-preference vote. So, it's not a perfect system. But it's still much better than first-past-the-post!
The article makes a valid point that Score Voting is just as "expressive" as preferential voting: you still get to say how much you like each candidate, compared to the other candidates. And it's also simpler than preferential voting. So that looks like a good option.
One of the linked articles in that section has a better example of where IRV breaks down. Consider this preference table: #voters their vote 6 C>A>B 2 B>A>C 3 B>C>A 4 A>B>C 2 A>C>B In this...
One of the linked articles in that section has a better example of where IRV breaks down.
Consider this preference table:
#voters
their vote
6
C>A>B
2
B>A>C
3
B>C>A
4
A>B>C
2
A>C>B
In this scenario, in the first round B is eliminated as they have only 5 #1 votes, to A and C's 6. 2 votes go to A and 3 votes go to C, per the 5 B voters' second preference, so C wins 9 to A's 8.
Now consider what happens when those two A>C>B voters change into C>A>B voters. Two people liking C more should make it so that C is more likely to win (and A less likely to win). But, this is the result:
#voters
their vote
8
C>A>B
2
B>A>C
3
B>C>A
4
A>B>C
Now, in this scenario, A is eliminated as they have 4 #1 votes compared to C's 8 and B's 5. Those 4 votes for A go to B. Now, B wins 9 votes to C's 8.
Now let's imagine those two voters sit at home instead:
#voters
their vote
6
C>A>B
2
B>A>C
3
B>C>A
Here C wins too.
This is undesirable because those 2 A>C>B -> C>A>B voters have to think strategically. If they vote their straight preference, the outcome they want won't happen. Those two would be better off if they didn't vote at all than to reveal their true preference. This is especially problematic because at least in FPTP you can reason about your strategic voting more or less, but in IRV, scenarios like the one I outlined can surprise you.
Yet another reason why I like score or approval voting is that the outcome is easily understood. You don't have scenarios like the one you describe where the public doesn't really understand how this person won. In approval voting you have "this many voters said OK to this candidate", or in score voting you have "this candidate earned this many stars". Easily understandable and computable with a simple summation.
I'm 100% with you that IRV>FPTP. I just wonder why IRV is the alternative that is always offered to us when there seem to be better options.
I think this analysis of instant run-off voting must have been written by people who have never actually used it. The analysis seems to have a built-in assumption that only candidates who get high...
I think this analysis of instant run-off voting must have been written by people who have never actually used it. The analysis seems to have a built-in assumption that only candidates who get high numbers of first-preference votes should be eligible to win. In the second scenario where C gets 8 first-preference votes, B gets 5 first-preference votes, and A gets 4 first-preference votes, it's seen as bad if B wins instead of C.
But those of us who use preferential voting know that our first preference may not win. We know that our second preference matters because our second preference might be the winner. We know that this form of voting is aimed at finding the best (or least-bad) compromise candidate, rather than the most popular candidate. We don't expect the candidate with the highest number of first-preference votes to win (even if that is the most common outcome). We know that candidates can "win on preferences", as we say here.
It's not that a non-#1 leader shouldn't win. Think about things from those two voters' perspective. They get the outcome they want by hiding their true preference. This is non-ideal in a voting...
It's not that a non-#1 leader shouldn't win. Think about things from those two voters' perspective. They get the outcome they want by hiding their true preference. This is non-ideal in a voting system. Arrow's Impossibility Theorem says that this will happen no matter what kind of ranked voting system we choose. In particular, this violates this goal of voting systems (from https://en.wikipedia.org/wiki/Arrow%27s_impossibility_theorem ):
Monotonicity, or positive association of social and individual values
If any individual modifies his or her preference order by promoting a certain option, then the societal preference order should respond only by promoting that same option or not changing, never by placing it lower than before. An individual should not be able to hurt an option by ranking it higher.
Non-ranked voting systems like AV or SV can also help find compromise candidates, but avoid these monotonicity issues completely.
But what is the outcome they want? Here's how it works in practice here in Australia. There are two main parties. (I'm simplifying slightly, because the one of those major parties is actually in a...
They get the outcome they want by hiding their true preference.
But what is the outcome they want?
Here's how it works in practice here in Australia.
There are two main parties. (I'm simplifying slightly, because the one of those major parties is actually in a coalition with a minor party, but let's pretend the coalition is a single party.) They represent right-wing and left-wing policies respectively (again: simplifying). There are also minor right-wing parties and minor left-wing parties, and independent candidates.
If someone wants to vote for a minor party or independent candidate in their electorate, they can do so. They do so knowing the reality that, unless a surprisingly high number of people want to vote for the minor party/independent, that candidate won't win. Only one of the candidates from the two major parties will win. However, there's a law in Australia that provides electoral funding to all candidates who achieve more than 4% of the first-preference vote in their electorate, in proportion to the number of votes they get (more votes = more funding).
On that basis, many people will vote for their preferred minor party/independent candidate as #1 and their preferred major party candidate as #2 (simplifying), so that their minor party/independent gets funding, and their preferred major party wins... and their minor party/independent candidate still has a small chance of winning.
So... what is that voter's desired outcome? Sure, their minor party/independent candidate might not win this election, but they get funding to keep fighting and keep building their support so that they can win the next election. The number of first-preference votes for the minor party/independent also shows everyone that the minor party/independent has a certain amount of support in the electorate - which can (and does!) influence the major parties' policies. And, the voter also got to choose which major party should form government.
Preferential voting in Australia allows for multiple outcomes. It's not as simple as whether your preferred candidate wins or loses.
Let's say that an Australian liked minor party candidate A and major party candidate B. They were going to vote for 1-A, 2-B per the strategy you describe. Let's say that day of the election, the...
Let's say that an Australian liked minor party candidate A and major party candidate B. They were going to vote for 1-A, 2-B per the strategy you describe. Let's say that day of the election, the voter changes their mind. Candidate B gives the speech of their political career and the voter decides to switch their voting order. They decide to vote 1-B, 2-A instead.
Now, it would be kinda broken if this made it so that B was now less likely to win and could even cost B the election, right? But this can happen in IRV. Can't happen in non-ranked voting systems.
1000% agreed there. FPTP is a travesty. A relic of when votes involved quill pens and oak boxes and were transported by horse. Its main benefit today is that it will mathematically produce a...
1000% agreed there. FPTP is a travesty. A relic of when votes involved quill pens and oak boxes and were transported by horse. Its main benefit today is that it will mathematically produce a two-party system which is convenient because now there's only two parties to bribe. I consider it the very core of the US's political dysfunction.
I'd rank them FPTP <<<<< IRV < { SV, AV }
Now, the question I have is why is IRV always put up as the alternative? The main argument against IRV and in favor of FPTP is that IRV is complicated and takes a while to compute the result. SV and AV have none of this, but most people have never heard of them. Both AV and SV have been used starting in the 13th century. IRV was invented fewer than 150 years ago in 1871. I can't help but have conspiratorial suspicions that they've offered us up the second worst option so that they can try to keep the worst option.
You'll get no disagreement from me! I've often made the point that the USA's political system could be improved overnight if they introduced something like preferential voting. Because it's used...
FPTP is a travesty. [...] I consider it the very core of the US's political dysfunction.
You'll get no disagreement from me! I've often made the point that the USA's political system could be improved overnight if they introduced something like preferential voting.
Now, the question I have is why is IRV always put up as the alternative?
Because it's used in actual elections? Are Score Voting or Approval Voting used in any real-world government elections?
But even Australia hasn't always had IRV. Looks like you had 4 elections with FPTP before you changed to IRV in 1919 (https://en.wikipedia.org/wiki/Electoral_system_of_Australia ). Why IRV? Why...
But even Australia hasn't always had IRV. Looks like you had 4 elections with FPTP before you changed to IRV in 1919 (https://en.wikipedia.org/wiki/Electoral_system_of_Australia ). Why IRV? Why not Score Voting or Approval Voting then? AV and SV had been around for half a millennia even back then, while IRV as an idea was at the time less than 50 years old.
Because it's used in actual elections?
FPTP is used in actual elections. Those AV and SV things in the 13th century were actual elections. This isn't a particularly compelling argument. It's used for elections because it's used for elections is rather circular logic.
Both Score Voting and Approval Voting are still currently used in some small polities, internally within parties and at the UN:
The answer is in that Wikipedia article: "The system was first used for election for the Queensland Parliament in 1892." Australians already knew about preferential voting in some colonies and...
The answer is in that Wikipedia article: "The system was first used for election for the Queensland Parliament in 1892." Australians already knew about preferential voting in some colonies and states when they introduced it at a federal level 20-ish years later. It was familiar, and it was already working in some places in Australia.
It's used for elections because it's used for elections is rather circular logic.
You misunderstand me. Your question was why Instant Run-off Voting is always offered as the alternative to first-past-the-post voting. My answer is that people keep suggesting Instant Run-off Voting as an alternative because it's widely used. They suggest it because they've seen it in action. You can't suggest something you don't know about!
While ridiculous that it might just be marketing that popularized an inferior design, it gives me hope that maybe, through a ton of effort, marketing could bring us an even better solution.
While ridiculous that it might just be marketing that popularized an inferior design, it gives me hope that maybe, through a ton of effort, marketing could bring us an even better solution.
Nicky Case's work is outstanding—they make lots of things that sit on this delightful line between "games" and "education". All of which deserves posting here, but in case you haven't seen much of it:
Naturally, there's more than just those three, but these are cool starts.
I've been a fan for ages, but somehow never came across their memory one! Thanks for linking it; I might check out the app, too, since I've found this sort of thing hard to put into practice on my own.
This uses interactive diagrams to show not only why FPTP is fundamentally flawed, but also, why ranked choice mechanisms (such as IRV) are also similarly flawed. But, it ends on an optimistic note by showing that a better way is not only mathematically possible, but also quite easy and within our grasp with an endorsement from Dr. Arrow himself. You'll see that Arrow's Impossibility Theorem isn't as as doom and gloom as it may seem.
Did Ralph Nader Steal the 2000 election or did the Florida Supreme Court? God forbid we have some semblance of choice in American elections lest the establishment candidates do poorly. Won't someone think of the poor political establishment and their candidates for once?
That said this was really interesting even if the author's understanding of American elections isn't the same as mine.
The only reason there are "establishment" candidates is because of this voting mechanism. FPTP mathematically ensures that there will be only two major parties no matter how anyone votes. Any third choice can only play the role of the spoiler and both the original party and the splinter faction will, over time, realize it's in their best interest to reconcile as their mutual opponent wins over and over.
This vid from CGP Grey is one of my favorites to explain the issue: https://www.youtube.com/watch?v=s7tWHJfhiyo
A vote for Jill Stein is a vote for exposing the charade of choice that is American first past the post elections.
I have a problem with this scenario in the "How are the alternatives?" section:
In this scenario, Henry Hexagon is politically near to Tracy Triangle. Therefore, as we move the voters closer to Tracy, we also move them closer to Henry. However, Henry is also slightly nearer to the political centre, and therefore slightly closer to the centre of the voter group, no matter where you move them. That's why he attracts slightly more votes than Tracy and ends up winning the election.
If, between two similar candidates, the voters end up voting for the one whose policies they like more... why is that a problem? If they like Henry's policies slightly more than Tracy's, so they vote for Henry slightly more often than Tracy... why is that a bad outcome?
There seems to be some bias built in here, assuming that there are only two "winning" candidates and any other candidate must, by definition, be a spoiler candidate - without stopping to consider that maybe the other candidate is actually better overall in the electorate's eyes.
Imagine there was a left-wing candidate and a right-wing candidate, and both candidates were equally popular with their constituency and equally unpopular with the opposing constituency, and then a less-popular centrist candidate entered the field. In Instant Run-off Voting (or "preferential voting" as we call it here in Australia), it's very likely that everyone voting for the left-wing or right-wing candidates will put the unpopular centrist as their second choice, ahead of that other candidate from the other wing - and this might end up winning the centrist candidate the election. But... how is this a bad thing? Sure, noone got their first choice, but everyone got their second choice - and everyone avoided that other choice that they absolutely didn't want. Isn't that good?
I'm not seeing the problem with this scenario. But, then again, I've been raised in a country which uses preferential voting, and I know I'm biassed towards it - especially when I compare it to first-past-the-post voting.
That said, I know it's possible to "game" preferential voting. There are so-called "preference whisperers" in Australia who go around various minor parties and independent candidates, and get them to make deals with each other to "swap preferences" (you print how-to-vote cards for your supporters, telling them to put the other candidate as #2, while the other candidate does the same for you). On the basis of these deals, and our extra-strange voting system in the Senate - which is not straight preferential voting - we've had some very unexpected candidates get elected, including people who got less than 2% of the first-preference vote. So, it's not a perfect system. But it's still much better than first-past-the-post!
The article makes a valid point that Score Voting is just as "expressive" as preferential voting: you still get to say how much you like each candidate, compared to the other candidates. And it's also simpler than preferential voting. So that looks like a good option.
One of the linked articles in that section has a better example of where IRV breaks down.
Consider this preference table:
In this scenario, in the first round B is eliminated as they have only 5 #1 votes, to A and C's 6. 2 votes go to A and 3 votes go to C, per the 5 B voters' second preference, so C wins 9 to A's 8.
Now consider what happens when those two A>C>B voters change into C>A>B voters. Two people liking C more should make it so that C is more likely to win (and A less likely to win). But, this is the result:
Now, in this scenario, A is eliminated as they have 4 #1 votes compared to C's 8 and B's 5. Those 4 votes for A go to B. Now, B wins 9 votes to C's 8.
Now let's imagine those two voters sit at home instead:
Here C wins too.
This is undesirable because those 2 A>C>B -> C>A>B voters have to think strategically. If they vote their straight preference, the outcome they want won't happen. Those two would be better off if they didn't vote at all than to reveal their true preference. This is especially problematic because at least in FPTP you can reason about your strategic voting more or less, but in IRV, scenarios like the one I outlined can surprise you.
Yet another reason why I like score or approval voting is that the outcome is easily understood. You don't have scenarios like the one you describe where the public doesn't really understand how this person won. In approval voting you have "this many voters said OK to this candidate", or in score voting you have "this candidate earned this many stars". Easily understandable and computable with a simple summation.
I'm 100% with you that IRV>FPTP. I just wonder why IRV is the alternative that is always offered to us when there seem to be better options.
I think this analysis of instant run-off voting must have been written by people who have never actually used it. The analysis seems to have a built-in assumption that only candidates who get high numbers of first-preference votes should be eligible to win. In the second scenario where C gets 8 first-preference votes, B gets 5 first-preference votes, and A gets 4 first-preference votes, it's seen as bad if B wins instead of C.
But those of us who use preferential voting know that our first preference may not win. We know that our second preference matters because our second preference might be the winner. We know that this form of voting is aimed at finding the best (or least-bad) compromise candidate, rather than the most popular candidate. We don't expect the candidate with the highest number of first-preference votes to win (even if that is the most common outcome). We know that candidates can "win on preferences", as we say here.
It's not that a non-#1 leader shouldn't win. Think about things from those two voters' perspective. They get the outcome they want by hiding their true preference. This is non-ideal in a voting system. Arrow's Impossibility Theorem says that this will happen no matter what kind of ranked voting system we choose. In particular, this violates this goal of voting systems (from https://en.wikipedia.org/wiki/Arrow%27s_impossibility_theorem ):
Non-ranked voting systems like AV or SV can also help find compromise candidates, but avoid these monotonicity issues completely.
But what is the outcome they want?
Here's how it works in practice here in Australia.
There are two main parties. (I'm simplifying slightly, because the one of those major parties is actually in a coalition with a minor party, but let's pretend the coalition is a single party.) They represent right-wing and left-wing policies respectively (again: simplifying). There are also minor right-wing parties and minor left-wing parties, and independent candidates.
If someone wants to vote for a minor party or independent candidate in their electorate, they can do so. They do so knowing the reality that, unless a surprisingly high number of people want to vote for the minor party/independent, that candidate won't win. Only one of the candidates from the two major parties will win. However, there's a law in Australia that provides electoral funding to all candidates who achieve more than 4% of the first-preference vote in their electorate, in proportion to the number of votes they get (more votes = more funding).
On that basis, many people will vote for their preferred minor party/independent candidate as #1 and their preferred major party candidate as #2 (simplifying), so that their minor party/independent gets funding, and their preferred major party wins... and their minor party/independent candidate still has a small chance of winning.
So... what is that voter's desired outcome? Sure, their minor party/independent candidate might not win this election, but they get funding to keep fighting and keep building their support so that they can win the next election. The number of first-preference votes for the minor party/independent also shows everyone that the minor party/independent has a certain amount of support in the electorate - which can (and does!) influence the major parties' policies. And, the voter also got to choose which major party should form government.
Preferential voting in Australia allows for multiple outcomes. It's not as simple as whether your preferred candidate wins or loses.
Let's say that an Australian liked minor party candidate A and major party candidate B. They were going to vote for 1-A, 2-B per the strategy you describe. Let's say that day of the election, the voter changes their mind. Candidate B gives the speech of their political career and the voter decides to switch their voting order. They decide to vote 1-B, 2-A instead.
Now, it would be kinda broken if this made it so that B was now less likely to win and could even cost B the election, right? But this can happen in IRV. Can't happen in non-ranked voting systems.
Okay. I'll concede there may be better voting systems than "instant run-off voting". But "instant run-off voting" is still better than other systems.
1000% agreed there. FPTP is a travesty. A relic of when votes involved quill pens and oak boxes and were transported by horse. Its main benefit today is that it will mathematically produce a two-party system which is convenient because now there's only two parties to bribe. I consider it the very core of the US's political dysfunction.
I'd rank them FPTP <<<<< IRV < { SV, AV }
Now, the question I have is why is IRV always put up as the alternative? The main argument against IRV and in favor of FPTP is that IRV is complicated and takes a while to compute the result. SV and AV have none of this, but most people have never heard of them. Both AV and SV have been used starting in the 13th century. IRV was invented fewer than 150 years ago in 1871. I can't help but have conspiratorial suspicions that they've offered us up the second worst option so that they can try to keep the worst option.
Check out this op-ed from David Cameron back when the UK was considering IRV: https://www.telegraph.co.uk/news/politics/av-referendum/8485118/David-Cameron-why-keeping-first-past-the-post-is-vital-for-democracy.html (he refers to IRV as AV, as it was billed the Alternative Vote).
All of his arguments against IRV and for FPTP wouldn't hold up at all if the alternative was Score or Approval voting.
And so the whole thing failed 2:1: https://en.wikipedia.org/wiki/Results_of_the_2011_United_Kingdom_Alternative_Vote_referendum
You'll get no disagreement from me! I've often made the point that the USA's political system could be improved overnight if they introduced something like preferential voting.
Because it's used in actual elections? Are Score Voting or Approval Voting used in any real-world government elections?
But even Australia hasn't always had IRV. Looks like you had 4 elections with FPTP before you changed to IRV in 1919 (https://en.wikipedia.org/wiki/Electoral_system_of_Australia ). Why IRV? Why not Score Voting or Approval Voting then? AV and SV had been around for half a millennia even back then, while IRV as an idea was at the time less than 50 years old.
FPTP is used in actual elections. Those AV and SV things in the 13th century were actual elections. This isn't a particularly compelling argument. It's used for elections because it's used for elections is rather circular logic.
Both Score Voting and Approval Voting are still currently used in some small polities, internally within parties and at the UN:
But not nearly to the extent IRV is used: https://en.wikipedia.org/wiki/Instant-runoff_voting#Global_use
The answer is in that Wikipedia article: "The system was first used for election for the Queensland Parliament in 1892." Australians already knew about preferential voting in some colonies and states when they introduced it at a federal level 20-ish years later. It was familiar, and it was already working in some places in Australia.
You misunderstand me. Your question was why Instant Run-off Voting is always offered as the alternative to first-past-the-post voting. My answer is that people keep suggesting Instant Run-off Voting as an alternative because it's widely used. They suggest it because they've seen it in action. You can't suggest something you don't know about!
People suggest and use things which are familiar. Those uses of Approval Voting and Score Voting you cite were esoteric and minimal. Meanwhile, "Instant-runoff voting is based on the Single Transferable Vote [proportional representation] electoral system, invented independently by Thomas Hare in 1857 and Carl Andrae in 1855." and "with the help of contemporaries such as John Stuart Mill and Catherine Helen Spence, Hare popularised the idea of proportional representation worldwide." The inventor of Instant Run-off Voting's predecessor deliberately promoted his new system as widely as he could, and had high-profile friends to help him with that. People therefore adopted and adapted the system they knew about, rather than a system used only by Cardinals in secret meetings, or by one tiny extinct kingdom. It's probably as simple as that.
While ridiculous that it might just be marketing that popularized an inferior design, it gives me hope that maybe, through a ton of effort, marketing could bring us an even better solution.