RCV is a bad option that’s presented as if it could fix anything.
RCV was first invented in the 1780s, and the inventor wrote about it as the bad idea that it was, but because he was a mathematician, he wrote about the dead ends in the search for something better than the simple First Past the Post system that was in use in America.
The inventor, by the way, was Nicolas de Caritat, Marquis of Condorcet. His life was fascinating, and his death tragic, but for the moment we’ll focus on his efforts to find a better voting system.
He created a criteria for a better voting system, now named in his honor. The Condorcet Winner is the candidate who can win against any other candidate in a one on one race. They’re sometimes called the pairwise winner.
The point being, RCV, or it’s older name of Instant Runoff, cannot reliably elect the Condorcet Winner.
This was why Condorcet abandoned the system.
It was revived by some guys a few decades after Condorcet’s death. They didn’t care that it was a flawed system, just that it was slightly better than the only other option available at the time.
But that was 200 years ago. We now have quite a few options that are not deeply flawed.
First is Approval. It’s a dead simple system that always finds the Condorcet Winner.
How Approval works is thus; you get a list of names on your ballot. Mark any and all that you approve of. You may mark more than one candidate for each position.
The candidate with the highest overall approval wins.
Then there’s STAR. It’s brand new as far as voting systems go, only created in 2014. But it’s also the best system designed to date.
Basically the voter rates each candidate on a scale of 0 to 5. Multiple candidates can have the same rating. To find the winner, you simply add up the ratings for each candidate, then you take the highest two and look at each ballot. The candidate with the higher rating on that ballot gets the vote. If neither of the top two is rated higher on a ballot, either being not rated or rated the same, then the ballot is counted as No Preference, and that number is reported as part of the final tally.
As you said, STAR is arguably better in some ways but Approval being dead simple to explain to people and technically already supported by existing voting machines is worth a whole lot on its own as far as being a good voting system.
Try explaining STAR or Approval to someone who is only familiar with FPTP, see which one they understand more quickly.
Because “Vote for everyone you’re OK with winning the office and it counts as a vote for any of them, whoever gets the most votes wins” or “It’s just like what we’ve been doing, but you can pick more than one person and your vote counts for all of them” explains Approval voting.
As opposed to having to do a cumulative total across all ballots to figure out if your ballot counts as a vote at all, before figuring out whether your vote actually counts as a vote for someone you voted 5 for or someone you voted 2 for.
That’s fair. I just think being able to sell people on it is a high value part of any solution, and any system where you can’t know how your vote will be counted until you count all the other votes is necessarily a harder sell. As is any system where they will have to do something radically different.
As to knowing who wins. Well, that’s always the rub. There is no system that lets you know who wins before all the counting is done.
Not knowing who wins, but knowing who your ballot will be counted as a vote for in the end. The answer is you can’t know and it might just be tossed as no vote depending on how everyone else votes.
Imagine we’ve moved to STAR and leading up to the first STAR election for President someone asks you how to be sure their ballot will in the end be counted as a vote for Jill Stein. The answer is that you can’t, because until every other ballot is counted it’s impossible to know if any ballots at all can be counted as ballots for Jill Stein in the end. Let alone trying to report on the count as it happens in a coherent way your grandparents might understand.
Mathematically it’s great, but it fails at being easy to explain, easy to implement, and easy to report on. Especially to people used to FPTP. It fails in the parts that it needs to most succeed at socially to be a viable option to see adoption.
I’m a pretty intelligent person. And I can understand this small article you’ve written about voting systems and it’s still bored the ever living shit out of me. Do you honestly expect the average voter to be capable of understanding all of what you just laid down here?
Ranked choice voting has the added benefit of everybody can understand what the fuck it is. It’s very simple and it’s much better than first past the post. Sure it’s not perfect but perfect is not going to happen not when people are so dumb that they vote against their own interests simply because a washed up TV star said things they wanted to hear.
Ranked choice voting has the added benefit of everybody can understand what the fuck it is. It’s very simple and it’s much better than first past the post.
Approval is better than Ranked Choice at this. And it lacks the weaknesses of RCV. Literally the only major change from FPTP is that you can vote for as many candidates as you’d like on your ballot. No spoiler effect because there’s nothing to spoil. Strategic voting isn’t remotely as bad either. RCV is in many cases a step up from FPTP (but not always and the more voters actually understand it the worse it gets) but it’s still a bad solution to the problem.
I’d support STAR over RCV because it’s mathematically better but my go to is Approval, mostly because of the specific issue with STAR you’ve pointed out - it’s more complicated to explain, vote under and report on than FPTP or Approval.
The thing is, Ranked Choice is broken in dozens of ways. It’s actually more broken than First Past the Post.
It gives bad results that do not match the lies that it’s advocates tell, because FairVote lies their asses off about the system.
Fuck, it fails the Monotonicity Criteria.
I’ll explain this one simply. The Monotonicity Criteria says more support for Candidate A should increase the chances that Candidate A wins. Under Ranked Choice, listing Candidate A first can cause Candidate C to win.
That and the insecurity around counting make Ranked Choice a fucking stupid idea. It was a bad system when invented and it’s not gotten any better.
RCV please
RCV is a bad option that’s presented as if it could fix anything.
RCV was first invented in the 1780s, and the inventor wrote about it as the bad idea that it was, but because he was a mathematician, he wrote about the dead ends in the search for something better than the simple First Past the Post system that was in use in America.
The inventor, by the way, was Nicolas de Caritat, Marquis of Condorcet. His life was fascinating, and his death tragic, but for the moment we’ll focus on his efforts to find a better voting system.
He created a criteria for a better voting system, now named in his honor. The Condorcet Winner is the candidate who can win against any other candidate in a one on one race. They’re sometimes called the pairwise winner.
The point being, RCV, or it’s older name of Instant Runoff, cannot reliably elect the Condorcet Winner.
This was why Condorcet abandoned the system.
It was revived by some guys a few decades after Condorcet’s death. They didn’t care that it was a flawed system, just that it was slightly better than the only other option available at the time.
But that was 200 years ago. We now have quite a few options that are not deeply flawed.
First is Approval. It’s a dead simple system that always finds the Condorcet Winner.
How Approval works is thus; you get a list of names on your ballot. Mark any and all that you approve of. You may mark more than one candidate for each position.
The candidate with the highest overall approval wins.
Then there’s STAR. It’s brand new as far as voting systems go, only created in 2014. But it’s also the best system designed to date.
Basically the voter rates each candidate on a scale of 0 to 5. Multiple candidates can have the same rating. To find the winner, you simply add up the ratings for each candidate, then you take the highest two and look at each ballot. The candidate with the higher rating on that ballot gets the vote. If neither of the top two is rated higher on a ballot, either being not rated or rated the same, then the ballot is counted as No Preference, and that number is reported as part of the final tally.
As you said, STAR is arguably better in some ways but Approval being dead simple to explain to people and technically already supported by existing voting machines is worth a whole lot on its own as far as being a good voting system.
Try explaining STAR or Approval to someone who is only familiar with FPTP, see which one they understand more quickly.
Because “Vote for everyone you’re OK with winning the office and it counts as a vote for any of them, whoever gets the most votes wins” or “It’s just like what we’ve been doing, but you can pick more than one person and your vote counts for all of them” explains Approval voting.
As opposed to having to do a cumulative total across all ballots to figure out if your ballot counts as a vote at all, before figuring out whether your vote actually counts as a vote for someone you voted 5 for or someone you voted 2 for.
There is an argument for simplicity, but if we’re going to change things, we might as well try for the experimentally best option.
That’s fair. I just think being able to sell people on it is a high value part of any solution, and any system where you can’t know how your vote will be counted until you count all the other votes is necessarily a harder sell. As is any system where they will have to do something radically different.
Two rounds of counting. It’s more of a feature. Think of it as an automatic recount.
As to knowing who wins. Well, that’s always the rub. There is no system that lets you know who wins before all the counting is done.
Not knowing who wins, but knowing who your ballot will be counted as a vote for in the end. The answer is you can’t know and it might just be tossed as no vote depending on how everyone else votes.
Imagine we’ve moved to STAR and leading up to the first STAR election for President someone asks you how to be sure their ballot will in the end be counted as a vote for Jill Stein. The answer is that you can’t, because until every other ballot is counted it’s impossible to know if any ballots at all can be counted as ballots for Jill Stein in the end. Let alone trying to report on the count as it happens in a coherent way your grandparents might understand.
Mathematically it’s great, but it fails at being easy to explain, easy to implement, and easy to report on. Especially to people used to FPTP. It fails in the parts that it needs to most succeed at socially to be a viable option to see adoption.
I’m a pretty intelligent person. And I can understand this small article you’ve written about voting systems and it’s still bored the ever living shit out of me. Do you honestly expect the average voter to be capable of understanding all of what you just laid down here?
Ranked choice voting has the added benefit of everybody can understand what the fuck it is. It’s very simple and it’s much better than first past the post. Sure it’s not perfect but perfect is not going to happen not when people are so dumb that they vote against their own interests simply because a washed up TV star said things they wanted to hear.
Approval is better than Ranked Choice at this. And it lacks the weaknesses of RCV. Literally the only major change from FPTP is that you can vote for as many candidates as you’d like on your ballot. No spoiler effect because there’s nothing to spoil. Strategic voting isn’t remotely as bad either. RCV is in many cases a step up from FPTP (but not always and the more voters actually understand it the worse it gets) but it’s still a bad solution to the problem.
I’d support STAR over RCV because it’s mathematically better but my go to is Approval, mostly because of the specific issue with STAR you’ve pointed out - it’s more complicated to explain, vote under and report on than FPTP or Approval.
The thing is, Ranked Choice is broken in dozens of ways. It’s actually more broken than First Past the Post.
It gives bad results that do not match the lies that it’s advocates tell, because FairVote lies their asses off about the system.
Fuck, it fails the Monotonicity Criteria.
I’ll explain this one simply. The Monotonicity Criteria says more support for Candidate A should increase the chances that Candidate A wins. Under Ranked Choice, listing Candidate A first can cause Candidate C to win.
That and the insecurity around counting make Ranked Choice a fucking stupid idea. It was a bad system when invented and it’s not gotten any better.