3 ms·

Can you compare this against ranked-choice voting with instant run-offs?

godelski 5y ago
The first link there does actually compare against ranked choice. It is just called IRV. There's a larger comparison that includes more types of ranked choice (also called "ordinal") here[0]. Though I particularly like this animation[1]. I'll also add that there is a lot of nuance to voting and VSE isn't everything. There's a lot of factors to consider including: how do strategies work? Is it scalable? Can it easily be counted and/or verified statistically? How easy is it for voters? How effective are (strong and weak) spoiler candidates? And many more questions. For the most part cardinal methods (like Approval) fair better on all these accounts. There is a slight tradeoff to VSE. Note STAR in [0] has a slightly lower VSE than RP/Schulz (ranked), but a tighter grouping (better vs strategies). But also STAR is more complex for voters and counting than Approval is (I'd still argue STAR is easier for voters and counters than ordinal systems). So these are the types of tradeoffs you're making [0] https://electionscience.github.io/vse-sim/VSE/ https://electionscience.github.io/vse-sim/VSE/ [1] https://www.youtube.com/watch?v=-4FXLQoLDBA https://www.youtube.com/watch?v=-4FXLQoLDBA
- clairity 5y ago
  i'd actually love to see a voting system developed that's closer to a conjoint, which is a statistical method designed to reveal preferences, rather than relying on stated preferences (which are notoriously unreliable). it's more complicated, but we have plenty of technology nowadays to make it practical. one method of implementing this (from thinking about it for all of 2 minutes), is developing a list of the top 10-20 issues that voters care about in a given election and providing tradeoff choices for each, then matching them to the candidate closest to their revealed preference. of course, the biggest issue with method is that it moves the problem up one level, from voting to choice design. [0]: https://wikipedia.org/wiki/Conjoint_analysis https://wikipedia.org/wiki/Conjoint_analysis
  - godelski 5y ago
    I honestly think it is very difficult to determine what voters' actual preferences are. I mean think of your standard political compass tests. Or even Meyer Briggs. They can change day to day or mood to mood and lots of questions tend to be leading. While I love the idea, I think logistically it would be a nightmare and greatly reduce transparency. It would also be far more vulnerable to so called experts manipulating the system if they aren't acting in good faith.
    - clairity 5y ago
      it's exceedingly difficult to fairly aggregate voter preferences, but all current voting systems seem to reduce all that complexity down to a singular choice in the overwhelming majority of cases. that's a shitload of information discarded, to put it crassly. the neat thing about conjoint is that it multiplexes the choice-making (when designed well), so that the choice burden scales reasonably while also better representing collective choice. nothing is going to be perfect, but conjoint could potentially be better than all current systems, probably by an order of magnitude or more. it's relatively complicated, so yes, 'experts' (i disdain this term, frankly) could in theory manipulate the system, but transparency of voting data (just like now) and survey design (the more likely bias vector) would make that difficult. transparency would allow competing independent journalists, academics/researchers, political analysts, etc. to weigh in and balance out biases (not perfectly of course), as well as verify results.
  - stocknoob 5y ago
    Referendums exist to get a state-wide policy change. In the last election, historically conservative states voted for liberal policies like legalized marijuana, $15 min wage, etc. For politicians, for me, it's more about (perceived) character vs. their stated policies.
    - clairity 5y ago
      you could certainly include character traits among choice decisions, but that's more fraught by subjectivity. that said, that's the magic of conjoint done well: turning subjective hidden preferences into (more) objective choice revelation. currently, there is no option to choose a candidate supporting marijuana legalization, $15 minimum wage, pro-life, and pro-gun policies (as an example), since we practically only have two relatively fixed packages of pre-ordained, all-or-nothing policy positions to choose from. revealed preferences would pressure parties into fielding candidates that better match the non-partisan preferences of constituents. or better yet, it would lead to the waning of the two-party system altogether, toward a more direct representative democracy.
troydavis 5y ago
Other people gave good answers about how this compares to RCV/IRV (thread: https://news.ycombinator.com/item?id=29266519#29269246 https://news.ycombinator.com/item?id=29266519#29269246), and both of the links in my first footnote above (https://news.ycombinator.com/item?id=29266519#29266537 https://news.ycombinator.com/item?id=29266519#29266537) cover it. That's all academic, though. A great real example is the 2021 Seattle City Attorney primary, which had only 3 candidates. The votes were split: 36%, 32%, 30%. These 2 candidates emerged from the primary: https://www.nytimes.com/2021/10/30/us/seattle-city-attorney-election.html https://www.nytimes.com/2021/10/30/us/seattle-city-attorney-.... The candidate that most accurately represented the opinion of the electorate, Holmes, was eliminated. That would have still happened in a ranked ballot. The voters who voted for Thomas-Kennedy or Davison in our current system would have still put Thomas-Kennedy or Davison #1 in a ranked ballot. I tried to distill this in https://twitter.com/troyd/status/1461389190206156802 https://twitter.com/troyd/status/1461389190206156802. And a 3-way race like that is actually the simplest possible case. Effects like that are much more likely with more candidates and they happen everywhere in the elimination cycle, they’re just less obvious to casual observers. That's what you see in https://www.youtube.com/watch?v=7btAd1HYvjU&t=1329 https://www.youtube.com/watch?v=7btAd1HYvjU&t=1329 (note that IRV == RCV in that video). (Obligatory disclaimer: I don't care who wins. I do care how accurately the winners reflect the inputs, namely, voter opinion.) Separately, Washington law requires cities hold a primary that produces 2 candidates for the general election, so cities can't legally eliminate primaries. That said, whether primaries are a net positive or negative is a really complicated discussion. Beyond election mechanics, the disadvantage of primaries is that only a subset of voters in the general will turn out in the primary. The advantage is that voters don't spend much time on ballots and it focuses that time. Seattle had 15 candidates for Mayor (many races are 10+). If someone spends 10 minutes on their ballot and the ballot has 10 races, that's 1 minute per race - 6 seconds per candidate. No easy answer here. However, for practical purposes we don't need to answer that hard question; this one change delivers an incredibly representative election.

The first link there does actually compare against ranked choice. It is just called IRV. There's a larger comparison that includes more types of ranked choice (also called "ordinal") here[0]. Though I particularly like this animation[1]. I'll also add that there is a lot of nuance to voting and VSE isn't everything. There's a lot of factors to consider including: how do strategies work? Is it scalable? Can it easily be counted and/or verified statistically? How easy is it for voters? How effective are (strong and weak) spoiler candidates? And many more questions. For the most part cardinal methods (like Approval) fair better on all these accounts. There is a slight tradeoff to VSE. Note STAR in [0] has a slightly lower VSE than RP/Schulz (ranked), but a tighter grouping (better vs strategies). But also STAR is more complex for voters and counting than Approval is (I'd still argue STAR is easier for voters and counters than ordinal systems). So these are the types of tradeoffs you're making [0] https://electionscience.github.io/vse-sim/VSE/ https://electionscience.github.io/vse-sim/VSE/ [1] https://www.youtube.com/watch?v=-4FXLQoLDBA https://www.youtube.com/watch?v=-4FXLQoLDBA

clairity 5y ago
i'd actually love to see a voting system developed that's closer to a conjoint, which is a statistical method designed to reveal preferences, rather than relying on stated preferences (which are notoriously unreliable). it's more complicated, but we have plenty of technology nowadays to make it practical. one method of implementing this (from thinking about it for all of 2 minutes), is developing a list of the top 10-20 issues that voters care about in a given election and providing tradeoff choices for each, then matching them to the candidate closest to their revealed preference. of course, the biggest issue with method is that it moves the problem up one level, from voting to choice design. [0]: https://wikipedia.org/wiki/Conjoint_analysis https://wikipedia.org/wiki/Conjoint_analysis
- godelski 5y ago
  I honestly think it is very difficult to determine what voters' actual preferences are. I mean think of your standard political compass tests. Or even Meyer Briggs. They can change day to day or mood to mood and lots of questions tend to be leading. While I love the idea, I think logistically it would be a nightmare and greatly reduce transparency. It would also be far more vulnerable to so called experts manipulating the system if they aren't acting in good faith.
  - clairity 5y ago
    it's exceedingly difficult to fairly aggregate voter preferences, but all current voting systems seem to reduce all that complexity down to a singular choice in the overwhelming majority of cases. that's a shitload of information discarded, to put it crassly. the neat thing about conjoint is that it multiplexes the choice-making (when designed well), so that the choice burden scales reasonably while also better representing collective choice. nothing is going to be perfect, but conjoint could potentially be better than all current systems, probably by an order of magnitude or more. it's relatively complicated, so yes, 'experts' (i disdain this term, frankly) could in theory manipulate the system, but transparency of voting data (just like now) and survey design (the more likely bias vector) would make that difficult. transparency would allow competing independent journalists, academics/researchers, political analysts, etc. to weigh in and balance out biases (not perfectly of course), as well as verify results.
- stocknoob 5y ago
  Referendums exist to get a state-wide policy change. In the last election, historically conservative states voted for liberal policies like legalized marijuana, $15 min wage, etc. For politicians, for me, it's more about (perceived) character vs. their stated policies.
  - clairity 5y ago
    you could certainly include character traits among choice decisions, but that's more fraught by subjectivity. that said, that's the magic of conjoint done well: turning subjective hidden preferences into (more) objective choice revelation. currently, there is no option to choose a candidate supporting marijuana legalization, $15 minimum wage, pro-life, and pro-gun policies (as an example), since we practically only have two relatively fixed packages of pre-ordained, all-or-nothing policy positions to choose from. revealed preferences would pressure parties into fielding candidates that better match the non-partisan preferences of constituents. or better yet, it would lead to the waning of the two-party system altogether, toward a more direct representative democracy.

troydavis 5y ago

Other people gave good answers about how this compares to RCV/IRV (thread: https://news.ycombinator.com/item?id=29266519#29269246 https://news.ycombinator.com/item?id=29266519#29269246), and both of the links in my first footnote above (https://news.ycombinator.com/item?id=29266519#29266537 https://news.ycombinator.com/item?id=29266519#29266537) cover it. That's all academic, though. A great real example is the 2021 Seattle City Attorney primary, which had only 3 candidates. The votes were split: 36%, 32%, 30%. These 2 candidates emerged from the primary: https://www.nytimes.com/2021/10/30/us/seattle-city-attorney-election.html https://www.nytimes.com/2021/10/30/us/seattle-city-attorney-.... The candidate that most accurately represented the opinion of the electorate, Holmes, was eliminated. That would have still happened in a ranked ballot. The voters who voted for Thomas-Kennedy or Davison in our current system would have still put Thomas-Kennedy or Davison #1 in a ranked ballot. I tried to distill this in https://twitter.com/troyd/status/1461389190206156802 https://twitter.com/troyd/status/1461389190206156802. And a 3-way race like that is actually the simplest possible case. Effects like that are much more likely with more candidates and they happen everywhere in the elimination cycle, they’re just less obvious to casual observers. That's what you see in https://www.youtube.com/watch?v=7btAd1HYvjU&t=1329 https://www.youtube.com/watch?v=7btAd1HYvjU&t=1329 (note that IRV == RCV in that video). (Obligatory disclaimer: I don't care who wins. I do care how accurately the winners reflect the inputs, namely, voter opinion.) Separately, Washington law requires cities hold a primary that produces 2 candidates for the general election, so cities can't legally eliminate primaries. That said, whether primaries are a net positive or negative is a really complicated discussion. Beyond election mechanics, the disadvantage of primaries is that only a subset of voters in the general will turn out in the primary. The advantage is that voters don't spend much time on ballots and it focuses that time. Seattle had 15 candidates for Mayor (many races are 10+). If someone spends 10 minutes on their ballot and the ballot has 10 races, that's 1 minute per race - 6 seconds per candidate. No easy answer here. However, for practical purposes we don't need to answer that hard question; this one change delivers an incredibly representative election.