Here is one interesting seat (Brisbane) that shows clearly how Preferential Voting (Ranked Choice or Instant Runoff) works in a real life example.
CANDIDATE (PARTY) VOTES PERCENT
EVANS, Trevor (MP) (Liberal National Party of Queensland) 33,400 38.35
JARRETT, Madonna (Australian Labor Party) 24,061 27.63
BATES, Stephen (Queensland Greens) 23,389 26.86
HOLD, Trevor (Pauline Hanson's One Nation) 1,780 2.04
KNUDSON, Justin (United Australia Party) 1,593 1.83
KENNEDY, Tiana (Animal Justice Party) 1,530 1.76
BULL, Anthony (Liberal Democrats) 1,333 1.53
Formal 87,086 98.14%
Informal 1,647 1.86%
Total 88,733 100.00%
The count is about 72% completed. And because the ALP (Jarrett) and Greens (Bates) more-or-less swap all preferences, one of these two must win the seat, and the sitting member (Evans) will lose. So the race is between the ALP and the Greens to NOT come third - you have to finish in the top two to get the preference flow you need. If you finish third you are eliminated (after all the minor parties are distributed), and your second preferences are then distributed (Greens to ALP, or vice versa, we'll see).
The votes of the (mostly) conservative minor parties will flow to the Liberal (Evans), but they won't help him anywhere near enough to prevent a decisive win by either the ALP or Greens.
"The votes of the (mostly) conservative minor parties will flow to the Liberal (Evans), but they won't help him anywhere near enough to prevent a decisive win by either the ALP or Greens."
But what if they did? This was Kenneth Arrow's whole point.
It's not a bug, it's a feature ... you might suspect I'm not a Kenneth Arrow disciple!
Every vote has equal weight - if after distribution of the second preferences of the four minor parties, the Liberal has 52% of the votes, while Labor and the Greens have 48% between them, then the most broadly preferred candidate triumphs.
Ranked Choice works well either way, even when the candidate coming third initially leapfrogs over second place with preferences, and ultimately wins it with further preferences. It means they were proportionally everyone's second choice.
I have no problem with either outcome. The only way to have fairer representation is to have five-member districts, and use proportional representation, so the Liberals, Greens, and ALP would definitely win one each, and the three would fight for a second, and the minors would fight it out for the final fifth place. Like in our Senate.
"If after distribution of the second preferences of the four minor parties, the Liberal has 52% of the votes, while Labor and the Greens have 48% between them, then the most broadly preferred candidate triumphs."
This is almost the same as First Past The Post. The votes are split among many candidates so the winner is an accident. If you are okay with this, then why would you oppose First Past The Post?
Assume a First Past The Post election where a Liberal gets 40% and a Social Democrat gets 30% and a Socialist gets 30%, so the Liberal wins. Is it bad that the Liberal wins? If yes, why would it then be okay to get a similar kind of vote splitting with RC/IR?
"This is almost the same as First Past The Post. The votes are split among many candidates so the winner is an accident. If you are okay with this, then why would you oppose First Past The Post?"
This is a highly inaccurate interpretation, and it leads me to suspect that you're not fully aware of how the distribution of preferences precisely works under our RC/IR system.
It is totally and utterly unlike FPTP in every possible way - please let me know if you require an explainer on the distribution of preferences.
"Assume a First Past The Post election where a Liberal gets 40% and a Social Democrat gets 30% and a Socialist gets 30%, so the Liberal wins. Is it bad that the Liberal wins?"
If you're used to the democratic strength of Preferential Voting, then yes - it's very very bad indeed. Because it does not (necessarily) reflect the broad view of the electorate, in terms of how they order the candidates.
And as in Brisbane (which has just been won by the Greens by-the-way), the result depends on who finished third after the distribution of the second preferences of the minor parties. As it should.
I'm glad the Greens won in Brisbane. I think the Greens can play a positive role in helping Australia imagine what its future looks like, in a world that moves away from fossil fuels.
We can assume that Social Democrats and Socialists have a lot in common, so my example was perhaps not perfect, and likewise, you mention that Labor and Greens were similar in many ways and could informally be thought of as a coalition, so it isn't much of a problem when their votes are effectively combined. But in the graph they offer on that other web site they give an example where the blue square and the red hexagon are very far apart, but you still get the same effect of the two of them accidentally operating as a coalition under RC/IR.
But I think for now, it would be accurate to say this deserves more study. It would be good if more governments using RC/IR released all details of each round, so political scientists could develop a more fine-grained understanding of when, in real life, two parties that are far apart are accidentally given coalition dynamics just by the odd math of RC/IR.
But I'll also repeat, as a practical matter, I think any system other than FPTP can work fine, or FPTP with a run-off election can work fine. There are many problems with the designs of our democracies, and in some sense the system of voting is the least important issue (though I have noticed that on this Substack and elsewhere, it is the subject that generates the most conversation).
These "three-corner" contests show the breadth and fairness of the Ranked Choice system.
In a FPTP (plurality wins) system Evans would win despite being the preferred candidate for less than 40% of the whole electorate. But of course if we did have a FPTP system, the ALP and Greens would never run against each other, since it would always lead to mutually assured destruction.
So the Preferential Voting system is especially good in that it encourages third parties to have a shot, and for everyone to turn up and vote, because their second preference really matters.
In effect, the "most broadly preferred" candidate (or if you will, the "least disliked" candidate) will ultimately be elected with a majority (ie, with more votes than everyone else combined). It seems to me that that is fairly democratic.
The Greens are favoured in the Brisbane seat because they will receive the bulk of Animal Justice (vegetarian activist) Party second references, and therefore moving a bit ahead of the ALP in the race to come second.
Switching subjects, and modes of conversation, it is interesting to see the Greens make progress in a country whose exports still rely heavily on fossil fuels. I'd be happy to see Australia play more of a leadership role in a revolution of sustainability.
Indeed - and it has been that way for decades. I've voted Green since I could (in the 1980s), but mostly because I'm a conservationist, rather than a global warming devotee. And despite our fossil-fuel based wealth, there are a LOT of us who will vote Green, but very rarely in concentrated-enough numbers to win any seats, however this time round it has occurred in four prosperous, educated, yuppie, big-city electorates.
It's a bit like the US - polling shows solid majority support for abortion rights, gender rights, gun control, cannabis legalisation, and much else, but this doesn't translate into political representation or political action.
I'm nervous about tearing down our fossil-fuel based electricity generation before an alternative is in place ... wind turbines, solar panels, and big batteries don't do the job.
"These "three-corner" contests show the breadth and fairness of the Ranked Choice system. In a FPTP (plurality wins) system Evans would win despite being the preferred candidate for less than 40% of the whole electorate."
Right. Well, no one on this Substack is going to defend FPTP.
Let's start with theory, and in another comment I'll talk about reality.
"These "three-corner" contests show the breadth and fairness of the Ranked Choice system. In a FPTP (plurality wins) system Evans would win despite being the preferred candidate for less than 40% of the whole electorate."
I don't think Kenneth Arrow was saying that Rank Choice always fails, but rather, he was saying it can always fail.
Even First Past The Post does well sometimes. If you have two candidates, one of whom gets 90% of the vote and one of whom gets 10% of the vote, and the one who gets 90% of the vote is declared the winner, then FPTP has done a good job of allowing voters to express their preferences. However, FPTP often fails. Not always, but often.
Kenneth Arrow was then able to show that any system where the voters are trying to pick the best candidate will fail. Most rank systems are less obvious in their failures than FPTP, but the edge cases and failures are still there.
This link offers a nice visualization of all the voting systems and it specifically mentions Australia:
This page also explains where the RC/IR Australian system can fail. Again, the argument isn't that it fails all the time, but rather, the argument is that every system of rank voting has glitches that malfunction under some situations.
But what does "malfunction under some situations" actually mean? Most essays written about voting systems focus on the question "Which system of voting allows voters to best express their preference?" I'd say that is the wrong question. The only question that I'm interested in is, which system of voting produces the best result for everyone? I do my best to explain this here:
"So, not only is Instant Runoff's glitch as undemocratic as First Past The Post's glitch, it's possibly worse – because while FPTP's counting method is simple and transparent, Instant Runoff is anything but. And a lack of transparency is an even deadlier sin nowadays, when our trust in government is already so low."
This borders on word salad - and as I said previously, the "glitch" is a feature, not a bug. It's not just your number of first preference votes that count, but how many others have you as their second preference!
Though the experts like to cite colourful examples, it's extremely rare indeed for a candidate to ultimately lose under RC/IR because "too many" people voted for them (as either first or second preference).
Plus if you're educated about RC/IR, or grow up with it, it's simple and transparent. The author is clutching at straws, methinks!
"Kenneth Arrow was then able to show that any system where the voters are trying to pick the best candidate will fail."
This is his first big error, in my view. Voters are never (individually or collectively) trying to "pick the best candidate" - they are all trying to help their preferred person to win ... even if they are a racist, homophobic bigot, and very stupid.
As I've said previously, Proportional Representation (requiring a minimum of five seats per district) is the most representative system, but it has real issues of governance.
The next best compromise is Preferential (RC/IR) Voting, which allows single-member districts. and while it encourages third parties to both run and influence the outcome, it doesn't lead to an unmanageable rabble in parliament.
"The next best compromise is Preferential (RC/IR) Voting, which allows single-member districts"
You haven't really made an argument, other than "I personally really like single-member districts with RC/IR." I get that designing a political system is always an exercise in aesthetics, but you haven't described what lead you to your preferences.
I thought I did describe it pretty well, but to reiterate:
Firstly, I prefer single-member electorates of 100,000 people over say five-member electorates of 500,000 - apart from the unwieldy size, multi-member electorates dilute representational responsibility (obviously), since electors don't have "a member", but a group.
There is no science to it ... it's just my preference (single member electorates in the lower house) - but I make no apology for such a preference.
And if you're going to have single-member electorates - in a healthy democracy with four or five or six active political parties - then you need a workable solution to find a winner when no-one gets more than 35% of first preferences, other than the ridiculous FPTP system.
I'm used to Preferential Voting because to me it works mathematically, and Australian voters are very used to it too - and they know that their second preference (and even sometimes their 3rd or 4th) can be quite important.
Plus it fosters minor parties to make the effort to come into existence.
I don't have a closed mind to variations (Scores etc), but it seems to me as soon as you don't have ranking, you get too dangerously close to FPTP, since you stop the count too soon.
It could be like giving your favourite person a "1", and then everybody else a "2" - it perverts the ranking system.
"And if you're going to have single-member electorates " -- it's a big "if" statement. RC/IR works well fairly well in this situation. Approval voting and score voting would probably work about as well. FPTP with a run-off election also works about as well. I think single-member districts are such a problem, so much so that the voting method doesn't matter much. It is fascinating to see that the overall swings from left to right, averaged over decades, seems to run in a similar manner in Australia and the USA, despite the different voting systems. The growth of the regulatory state after 1960, then the turn towards a weak version of "deregulation" in the 1970s or 1980s -- these changes have happened in both countries, despite the political differences.
I'd be interested to hear where you think the different political system has lead to different results in Australia? Is there any major part of Australia's law or programs or culture that you feel is because of the different political system?
"This page also explains where the RC/IR Australian system can fail."
I read the whole post, and while RC/IR was described, I couldn't make the leap to call any outcome a 'failure'. Like in our Brisbane case - there is no ultimate result that I would consider a failure; in three-corner contests the first requirement is to finish in the top two, and the second is to have good friends to swap preferences with.
Labor and Green have achieved the second goal, but ultimately one of them must finish third, and be eliminated.
The Liberals easily achieved the first goal, with 38%, but they have no big friends, and there isn't enough conservative support to get him to 50%, so he loses. The system works, and I don't think it's complicated or opaque - Brisbane is basically not a majority conservative seat, and the result will reflect that.
And I don't think there could have been a "Monotonicity Violation" in Brisbane under any circumstances - the swapping of preferences between Labor and the Greens usually exceeds 90%. From the point of view of RC/IR, they effectively are a coalition.
Fair enough, but the closeness of Labor and the Greens is not a deal-breaker, making RC/IR a problematic method otherwise.
Even if there are four viable candidates in the mix (with say around 30%, 30%, 20%, 20%), with no tight preference swaps, then it's going to depend critically where all the voters for those four candidates choose to direct their second preferences. That is the whole point of the PV method.
Despite Kenneth Arrow, the reality is that where people send their second preference will (a) most likely determine who wins a tight race, and (b) ultimately reflect the broad preferences of that electorate. The most broadly preferred person wins - even if they are the first choice of only 25% or so of the voters. Where is the problem?
I think we can agree that RC/IR works better than FPTP in most common multi-candidate situations. But it does still have the weird cases where the vote flips, in a manner that resembles the vote-splitting problems we see with FPTP.
Here is a good overview of how Australian elections work:
https://theconversation.com/how-does-australias-voting-system-work-177737
Here is one interesting seat (Brisbane) that shows clearly how Preferential Voting (Ranked Choice or Instant Runoff) works in a real life example.
CANDIDATE (PARTY) VOTES PERCENT
EVANS, Trevor (MP) (Liberal National Party of Queensland) 33,400 38.35
JARRETT, Madonna (Australian Labor Party) 24,061 27.63
BATES, Stephen (Queensland Greens) 23,389 26.86
HOLD, Trevor (Pauline Hanson's One Nation) 1,780 2.04
KNUDSON, Justin (United Australia Party) 1,593 1.83
KENNEDY, Tiana (Animal Justice Party) 1,530 1.76
BULL, Anthony (Liberal Democrats) 1,333 1.53
Formal 87,086 98.14%
Informal 1,647 1.86%
Total 88,733 100.00%
The count is about 72% completed. And because the ALP (Jarrett) and Greens (Bates) more-or-less swap all preferences, one of these two must win the seat, and the sitting member (Evans) will lose. So the race is between the ALP and the Greens to NOT come third - you have to finish in the top two to get the preference flow you need. If you finish third you are eliminated (after all the minor parties are distributed), and your second preferences are then distributed (Greens to ALP, or vice versa, we'll see).
The votes of the (mostly) conservative minor parties will flow to the Liberal (Evans), but they won't help him anywhere near enough to prevent a decisive win by either the ALP or Greens.
"The votes of the (mostly) conservative minor parties will flow to the Liberal (Evans), but they won't help him anywhere near enough to prevent a decisive win by either the ALP or Greens."
But what if they did? This was Kenneth Arrow's whole point.
It's not a bug, it's a feature ... you might suspect I'm not a Kenneth Arrow disciple!
Every vote has equal weight - if after distribution of the second preferences of the four minor parties, the Liberal has 52% of the votes, while Labor and the Greens have 48% between them, then the most broadly preferred candidate triumphs.
Ranked Choice works well either way, even when the candidate coming third initially leapfrogs over second place with preferences, and ultimately wins it with further preferences. It means they were proportionally everyone's second choice.
I have no problem with either outcome. The only way to have fairer representation is to have five-member districts, and use proportional representation, so the Liberals, Greens, and ALP would definitely win one each, and the three would fight for a second, and the minors would fight it out for the final fifth place. Like in our Senate.
"If after distribution of the second preferences of the four minor parties, the Liberal has 52% of the votes, while Labor and the Greens have 48% between them, then the most broadly preferred candidate triumphs."
This is almost the same as First Past The Post. The votes are split among many candidates so the winner is an accident. If you are okay with this, then why would you oppose First Past The Post?
Assume a First Past The Post election where a Liberal gets 40% and a Social Democrat gets 30% and a Socialist gets 30%, so the Liberal wins. Is it bad that the Liberal wins? If yes, why would it then be okay to get a similar kind of vote splitting with RC/IR?
"This is almost the same as First Past The Post. The votes are split among many candidates so the winner is an accident. If you are okay with this, then why would you oppose First Past The Post?"
This is a highly inaccurate interpretation, and it leads me to suspect that you're not fully aware of how the distribution of preferences precisely works under our RC/IR system.
It is totally and utterly unlike FPTP in every possible way - please let me know if you require an explainer on the distribution of preferences.
"Assume a First Past The Post election where a Liberal gets 40% and a Social Democrat gets 30% and a Socialist gets 30%, so the Liberal wins. Is it bad that the Liberal wins?"
If you're used to the democratic strength of Preferential Voting, then yes - it's very very bad indeed. Because it does not (necessarily) reflect the broad view of the electorate, in terms of how they order the candidates.
And as in Brisbane (which has just been won by the Greens by-the-way), the result depends on who finished third after the distribution of the second preferences of the minor parties. As it should.
I'm glad the Greens won in Brisbane. I think the Greens can play a positive role in helping Australia imagine what its future looks like, in a world that moves away from fossil fuels.
But again, look here:
https://ncase.me/ballot/
We can assume that Social Democrats and Socialists have a lot in common, so my example was perhaps not perfect, and likewise, you mention that Labor and Greens were similar in many ways and could informally be thought of as a coalition, so it isn't much of a problem when their votes are effectively combined. But in the graph they offer on that other web site they give an example where the blue square and the red hexagon are very far apart, but you still get the same effect of the two of them accidentally operating as a coalition under RC/IR.
But I think for now, it would be accurate to say this deserves more study. It would be good if more governments using RC/IR released all details of each round, so political scientists could develop a more fine-grained understanding of when, in real life, two parties that are far apart are accidentally given coalition dynamics just by the odd math of RC/IR.
But I'll also repeat, as a practical matter, I think any system other than FPTP can work fine, or FPTP with a run-off election can work fine. There are many problems with the designs of our democracies, and in some sense the system of voting is the least important issue (though I have noticed that on this Substack and elsewhere, it is the subject that generates the most conversation).
These "three-corner" contests show the breadth and fairness of the Ranked Choice system.
In a FPTP (plurality wins) system Evans would win despite being the preferred candidate for less than 40% of the whole electorate. But of course if we did have a FPTP system, the ALP and Greens would never run against each other, since it would always lead to mutually assured destruction.
So the Preferential Voting system is especially good in that it encourages third parties to have a shot, and for everyone to turn up and vote, because their second preference really matters.
In effect, the "most broadly preferred" candidate (or if you will, the "least disliked" candidate) will ultimately be elected with a majority (ie, with more votes than everyone else combined). It seems to me that that is fairly democratic.
The Greens are favoured in the Brisbane seat because they will receive the bulk of Animal Justice (vegetarian activist) Party second references, and therefore moving a bit ahead of the ALP in the race to come second.
"The Greens are favoured in the Brisbane seat"
Switching subjects, and modes of conversation, it is interesting to see the Greens make progress in a country whose exports still rely heavily on fossil fuels. I'd be happy to see Australia play more of a leadership role in a revolution of sustainability.
Indeed - and it has been that way for decades. I've voted Green since I could (in the 1980s), but mostly because I'm a conservationist, rather than a global warming devotee. And despite our fossil-fuel based wealth, there are a LOT of us who will vote Green, but very rarely in concentrated-enough numbers to win any seats, however this time round it has occurred in four prosperous, educated, yuppie, big-city electorates.
It's a bit like the US - polling shows solid majority support for abortion rights, gender rights, gun control, cannabis legalisation, and much else, but this doesn't translate into political representation or political action.
I'm nervous about tearing down our fossil-fuel based electricity generation before an alternative is in place ... wind turbines, solar panels, and big batteries don't do the job.
"These "three-corner" contests show the breadth and fairness of the Ranked Choice system. In a FPTP (plurality wins) system Evans would win despite being the preferred candidate for less than 40% of the whole electorate."
Right. Well, no one on this Substack is going to defend FPTP.
Let's start with theory, and in another comment I'll talk about reality.
"These "three-corner" contests show the breadth and fairness of the Ranked Choice system. In a FPTP (plurality wins) system Evans would win despite being the preferred candidate for less than 40% of the whole electorate."
I don't think Kenneth Arrow was saying that Rank Choice always fails, but rather, he was saying it can always fail.
Even First Past The Post does well sometimes. If you have two candidates, one of whom gets 90% of the vote and one of whom gets 10% of the vote, and the one who gets 90% of the vote is declared the winner, then FPTP has done a good job of allowing voters to express their preferences. However, FPTP often fails. Not always, but often.
Kenneth Arrow was then able to show that any system where the voters are trying to pick the best candidate will fail. Most rank systems are less obvious in their failures than FPTP, but the edge cases and failures are still there.
This link offers a nice visualization of all the voting systems and it specifically mentions Australia:
https://ncase.me/ballot/
This page also explains where the RC/IR Australian system can fail. Again, the argument isn't that it fails all the time, but rather, the argument is that every system of rank voting has glitches that malfunction under some situations.
But what does "malfunction under some situations" actually mean? Most essays written about voting systems focus on the question "Which system of voting allows voters to best express their preference?" I'd say that is the wrong question. The only question that I'm interested in is, which system of voting produces the best result for everyone? I do my best to explain this here:
https://demodexio.substack.com/p/should-a-system-of-voting-aim-to?s=w
By the way, this link:
https://ncase.me/ballot/
Summarizes the situation:
"So, not only is Instant Runoff's glitch as undemocratic as First Past The Post's glitch, it's possibly worse – because while FPTP's counting method is simple and transparent, Instant Runoff is anything but. And a lack of transparency is an even deadlier sin nowadays, when our trust in government is already so low."
This borders on word salad - and as I said previously, the "glitch" is a feature, not a bug. It's not just your number of first preference votes that count, but how many others have you as their second preference!
Though the experts like to cite colourful examples, it's extremely rare indeed for a candidate to ultimately lose under RC/IR because "too many" people voted for them (as either first or second preference).
Plus if you're educated about RC/IR, or grow up with it, it's simple and transparent. The author is clutching at straws, methinks!
"Though the experts like to cite colourful examples, it's extremely rare indeed "
There are few studies on this, but they article cites the possibility of these kinds of glitches happening 14% of the time, which is a lot.
"Kenneth Arrow was then able to show that any system where the voters are trying to pick the best candidate will fail."
This is his first big error, in my view. Voters are never (individually or collectively) trying to "pick the best candidate" - they are all trying to help their preferred person to win ... even if they are a racist, homophobic bigot, and very stupid.
As I've said previously, Proportional Representation (requiring a minimum of five seats per district) is the most representative system, but it has real issues of governance.
The next best compromise is Preferential (RC/IR) Voting, which allows single-member districts. and while it encourages third parties to both run and influence the outcome, it doesn't lead to an unmanageable rabble in parliament.
"The next best compromise is Preferential (RC/IR) Voting, which allows single-member districts"
You haven't really made an argument, other than "I personally really like single-member districts with RC/IR." I get that designing a political system is always an exercise in aesthetics, but you haven't described what lead you to your preferences.
I thought I did describe it pretty well, but to reiterate:
Firstly, I prefer single-member electorates of 100,000 people over say five-member electorates of 500,000 - apart from the unwieldy size, multi-member electorates dilute representational responsibility (obviously), since electors don't have "a member", but a group.
There is no science to it ... it's just my preference (single member electorates in the lower house) - but I make no apology for such a preference.
And if you're going to have single-member electorates - in a healthy democracy with four or five or six active political parties - then you need a workable solution to find a winner when no-one gets more than 35% of first preferences, other than the ridiculous FPTP system.
I'm used to Preferential Voting because to me it works mathematically, and Australian voters are very used to it too - and they know that their second preference (and even sometimes their 3rd or 4th) can be quite important.
Plus it fosters minor parties to make the effort to come into existence.
I don't have a closed mind to variations (Scores etc), but it seems to me as soon as you don't have ranking, you get too dangerously close to FPTP, since you stop the count too soon.
It could be like giving your favourite person a "1", and then everybody else a "2" - it perverts the ranking system.
"And if you're going to have single-member electorates " -- it's a big "if" statement. RC/IR works well fairly well in this situation. Approval voting and score voting would probably work about as well. FPTP with a run-off election also works about as well. I think single-member districts are such a problem, so much so that the voting method doesn't matter much. It is fascinating to see that the overall swings from left to right, averaged over decades, seems to run in a similar manner in Australia and the USA, despite the different voting systems. The growth of the regulatory state after 1960, then the turn towards a weak version of "deregulation" in the 1970s or 1980s -- these changes have happened in both countries, despite the political differences.
I'd be interested to hear where you think the different political system has lead to different results in Australia? Is there any major part of Australia's law or programs or culture that you feel is because of the different political system?
"This page also explains where the RC/IR Australian system can fail."
I read the whole post, and while RC/IR was described, I couldn't make the leap to call any outcome a 'failure'. Like in our Brisbane case - there is no ultimate result that I would consider a failure; in three-corner contests the first requirement is to finish in the top two, and the second is to have good friends to swap preferences with.
Labor and Green have achieved the second goal, but ultimately one of them must finish third, and be eliminated.
The Liberals easily achieved the first goal, with 38%, but they have no big friends, and there isn't enough conservative support to get him to 50%, so he loses. The system works, and I don't think it's complicated or opaque - Brisbane is basically not a majority conservative seat, and the result will reflect that.
And I don't think there could have been a "Monotonicity Violation" in Brisbane under any circumstances - the swapping of preferences between Labor and the Greens usually exceeds 90%. From the point of view of RC/IR, they effectively are a coalition.
"The swapping of preferences between Labor and the Greens usually exceeds 90%. From the point of view of RC/IR, they effectively are a coalition."
Right, so in this situation RC/IR can be made to work. But this is a specific circumstance that allows RC/IR to work well for this situation.
Fair enough, but the closeness of Labor and the Greens is not a deal-breaker, making RC/IR a problematic method otherwise.
Even if there are four viable candidates in the mix (with say around 30%, 30%, 20%, 20%), with no tight preference swaps, then it's going to depend critically where all the voters for those four candidates choose to direct their second preferences. That is the whole point of the PV method.
Despite Kenneth Arrow, the reality is that where people send their second preference will (a) most likely determine who wins a tight race, and (b) ultimately reflect the broad preferences of that electorate. The most broadly preferred person wins - even if they are the first choice of only 25% or so of the voters. Where is the problem?
I think we can agree that RC/IR works better than FPTP in most common multi-candidate situations. But it does still have the weird cases where the vote flips, in a manner that resembles the vote-splitting problems we see with FPTP.