4 ms·

It deeply frustrates me to see essays like this written without saying in big bold words at the top "Id take almost anything over fptp" People not deeply famili

by oehpr 6y ago

It deeply frustrates me to see essays like this written without saying in big bold words at the top "Id take almost anything over fptp" People not deeply familiar with the voting system world see this stuff and think to themselves "Wow. I guess there's no best voting system." Rather than what they should be thinking: "Oh wow, we have the worst voting system that forces bipartisan politics, we're burning alive from the inside AHHHHHHHH" So before we get into comments like "If the primary reason you fail a criterion is simply because your voting system just says more about a voters actual preferences, thus creating the potential for conflict, it is a weak failure when compared to a voting system that avoids that failure by simply lacking that information." We should always first, and last say: "I'd take nearly anything over FPTP" And anyone not doing this is not helping. As someone who has lived through 3 failed voting reforms in my province, I need to make it painfully clear to those not doing this. You are NOT helping.

godelski 6y ago
I wouldn't take IRV over FPTP. We've done it in the past and switched back. It also doesn't solve the problem. All it does is slightly increases the VSE but doesn't accomplish much else. IRV isn't great at allowing a new party to appear when major parties diverge from voter consensus. This isn't something you take several steps towards the right direction. If we take a step and it doesn't work, people are going to stop stepping in that direction and look for other optimization methods. You have to make a large enough step that the loss function actually decreases.
- oehpr 6y ago
  Out of curiosity how long did you have it for? Because voting reform can not fix the damage of a legacy of FPTP the moment it is instituted, it's going to take 4 or 5 elections before 1. the public figures out how it works 2. parties start appearing that arn't getting crushed by other network effects 3. that an understanding of what a strategic vote looks like in the new system is widespread and existing parties start playing the new game. And since election cycles are often something like... 4 years. You'd need a voting system in place for maybe 20 years before anything positive would start to happen, and that would apply to any voting system put forward. Like. These systems are for the far flung future of the country. Not... next election.
  - godelski 6y ago
    It was not national, but some cities and states had it for awhile. As to the time frame: remember that Australia has had it for 100+ years. They are dominated (>80%) by 2 parties (even though they have the MAJOR advantage of proportional representation through their parliament). One can say similar things about Ireland, though the domination isn't nearly the same and there are 3 strong parties (they use a variant of IRV called STV). But remember that neither the math nor the 100+ year experiment has shown that IRV won't collapse into a two party system. You'll find plenty of references through this thread and sources as to why if you search for "approval" (and "monotonicity criterion"). The thing is that the problem is rather complex. We don't just want it so you don't waste your vote, we don't want that to happen with ANY candidate (favorite betrayer criterion). We want other parties to rise up with no disadvantage. We want our voting to be expressive (STAR and score are substantially more expressive that IRV). We want high voter satisfaction efficiency (VSE) -- IRV doesn't do much better than FPTP. We want the system to be resistant to strategic voting. And we want the voting system to be dead simple. Frankly IRV doesn't fit these criteria. Approval fills every category but isn't extremely expressive. Score and STAR are at worst approval voting and at best much more expressive than ordinal (ranked) systems. But basically everything you're pointing out is well acknowledged and I won't challenge it. More what I'm suggesting is that Hasan and CGP Grey didn't do their research but drove the hype. There is no optimal system for voting, unfortunately, but there are pretty dang good ones. When the experts say cardinal, I'm going with that.
ScottBurson 6y ago
I don't know whether you'll see this belated reply, but anyway: I have had some sympathy with your position, as you can see in my sibling comment, BUT as I read more about IRV, I am actually coming to disagree. Look at this page, especially the chart near the bottom: https://electionscience.org/library/approval-voting-versus-irv/ https://electionscience.org/library/approval-voting-versus-i... What this says is that in simulation, given mostly tactical voters, IRV doesn't do any better than PV (FPTP). This surprised me, but I'm starting to see why, as the rest of that page explains. Here's another: https://www.rangevoting.org/TarrIrv.html https://www.rangevoting.org/TarrIrv.html The most telling argument here, I think, is that a century of IRV hasn't freed Australia from two-party domination. Another page with simuations: http://zesty.ca/voting/sim/ http://zesty.ca/voting/sim/ What voting reforms have been attempted in your province?

by oehpr 6y ago

godelski 6y ago

I wouldn't take IRV over FPTP. We've done it in the past and switched back. It also doesn't solve the problem. All it does is slightly increases the VSE but doesn't accomplish much else. IRV isn't great at allowing a new party to appear when major parties diverge from voter consensus. This isn't something you take several steps towards the right direction. If we take a step and it doesn't work, people are going to stop stepping in that direction and look for other optimization methods. You have to make a large enough step that the loss function actually decreases.

oehpr 6y ago
Out of curiosity how long did you have it for? Because voting reform can not fix the damage of a legacy of FPTP the moment it is instituted, it's going to take 4 or 5 elections before 1. the public figures out how it works 2. parties start appearing that arn't getting crushed by other network effects 3. that an understanding of what a strategic vote looks like in the new system is widespread and existing parties start playing the new game. And since election cycles are often something like... 4 years. You'd need a voting system in place for maybe 20 years before anything positive would start to happen, and that would apply to any voting system put forward. Like. These systems are for the far flung future of the country. Not... next election.
- godelski 6y ago
  It was not national, but some cities and states had it for awhile. As to the time frame: remember that Australia has had it for 100+ years. They are dominated (>80%) by 2 parties (even though they have the MAJOR advantage of proportional representation through their parliament). One can say similar things about Ireland, though the domination isn't nearly the same and there are 3 strong parties (they use a variant of IRV called STV). But remember that neither the math nor the 100+ year experiment has shown that IRV won't collapse into a two party system. You'll find plenty of references through this thread and sources as to why if you search for "approval" (and "monotonicity criterion"). The thing is that the problem is rather complex. We don't just want it so you don't waste your vote, we don't want that to happen with ANY candidate (favorite betrayer criterion). We want other parties to rise up with no disadvantage. We want our voting to be expressive (STAR and score are substantially more expressive that IRV). We want high voter satisfaction efficiency (VSE) -- IRV doesn't do much better than FPTP. We want the system to be resistant to strategic voting. And we want the voting system to be dead simple. Frankly IRV doesn't fit these criteria. Approval fills every category but isn't extremely expressive. Score and STAR are at worst approval voting and at best much more expressive than ordinal (ranked) systems. But basically everything you're pointing out is well acknowledged and I won't challenge it. More what I'm suggesting is that Hasan and CGP Grey didn't do their research but drove the hype. There is no optimal system for voting, unfortunately, but there are pretty dang good ones. When the experts say cardinal, I'm going with that.

ScottBurson 6y ago

I don't know whether you'll see this belated reply, but anyway: I have had some sympathy with your position, as you can see in my sibling comment, BUT as I read more about IRV, I am actually coming to disagree. Look at this page, especially the chart near the bottom: https://electionscience.org/library/approval-voting-versus-irv/ https://electionscience.org/library/approval-voting-versus-i... What this says is that in simulation, given mostly tactical voters, IRV doesn't do any better than PV (FPTP). This surprised me, but I'm starting to see why, as the rest of that page explains. Here's another: https://www.rangevoting.org/TarrIrv.html https://www.rangevoting.org/TarrIrv.html The most telling argument here, I think, is that a century of IRV hasn't freed Australia from two-party domination. Another page with simuations: http://zesty.ca/voting/sim/ http://zesty.ca/voting/sim/ What voting reforms have been attempted in your province?