Alaska RCV tangent

Discussion in 'The Red Room' started by Order2Chaos, Sep 22, 2022.

  1. Order2Chaos

    Order2Chaos Ultimate... Immortal Administrator

    Joined:
    Apr 2, 2004
    Messages:
    25,186
    Location:
    here there be dragons
    Ratings:
    +21,380
    So once all the dust has settled, it actually does look like RCV elected the wrong person via center squeeze.

    https://www.reddit.com/r/EndFPTP/comments/x9oupk/2022_alaska_special_general_vote_breakdown/

    What this means: there was a candidate -- Begich -- who would have won any head-to-head election against his opponents. This is the Condorcet winner, and while it doesn't always exist, it's widely acknowledged that if it does exist, a voting system should pick it. This version of RCV did not do so.

    Additionally, if 2913 Palin voters that preferred Begich to Peltola betrayed Palin by strategically ranking Begich 1st he would have won instead of Peltola. This is a classic spoiler effect under RCV.

    If Peltola were to have gained the support of 5825 Palin voters, she would have lost to Begich. Yep, if she'd gotten more votes, she would have lost.

    Arguably even worse, if 5825 Palin > Begich > Peltola voters had stayed home, their 2nd choice instead of their 3rd choice would have won. We don't like voting systems that penalize participation.

    If 5828 Palin > Begich voters, 2915 Begich voters, and 2914 Peltola > Begich voters were removed from the election, Begich would have won. But if you counted just those removed votes, Begich also would have won. This is inconsistent. Two winning scenarios added shouldn't add up to a losing scenario.

    This is 5 different types of failure in a single election, culminating in a center squeeze. I may like the winner, but the fact is there was a spoiler effect that RCV proponents claim doesn't exist. It's there, but somewhat chaotic. Should Alaska go back to plurality? Hell no. But if reformers don't push for a different voting method, say, approval or score, it's likely people will demand a more comprehensible system if they're going to have to vote strategically anyway, and that's going to default to plurality.

    TLDR: there was a Condorcet winner; it was Begich; he wasn't chosen, and there are several awful counterfactual vote totals that will make people real suspicious of RCV and rightly so.
  2. Tererune

    Tererune Troll princess and Magical Girl

    Joined:
    Jul 5, 2014
    Messages:
    37,049
    Location:
    Beyond the Silver Rainbow
    Ratings:
    +26,386
    There is no need for such math in the red room. :ban:
    • Agree Agree x 1
  3. Damar

    Damar Liberal Elitist

    Joined:
    Mar 29, 2004
    Messages:
    1,659
    Location:
    FL
    Ratings:
    +2,913
    Maybe the Democrats exercised better control over the process by only running one candidate.
  4. Jenee

    Jenee Driver 8

    Joined:
    Jan 20, 2008
    Messages:
    25,340
    Location:
    On the train
    Ratings:
    +19,334
    Ranked choice voting will still be better for voters in the long run. It worked so well in Vermont, politicians attempted to abolish it. But, voters weren't having it and it's back as of May 2022
  5. Order2Chaos

    Order2Chaos Ultimate... Immortal Administrator

    Joined:
    Apr 2, 2004
    Messages:
    25,186
    Location:
    here there be dragons
    Ratings:
    +21,380
    I believe it was the 2007 mayoral election in Burlington that failed in a very similar fashion, which led directly to its abolition. We’ll see how long it lasts here. It’ll be harder in this case to repeal, but I think it will be in the next couple years.
  6. Order2Chaos

    Order2Chaos Ultimate... Immortal Administrator

    Joined:
    Apr 2, 2004
    Messages:
    25,186
    Location:
    here there be dragons
    Ratings:
    +21,380
    One candidate who would have lost if she’d gotten 5825 more first-rank votes from Palin. That’s not a good system. More to the point, the ostensible point of RCV is to allow more choices without spoilers. It utterly failed at that here, and it’s going to put people off non-plurality systems which is very bad.
  7. Ancalagon

    Ancalagon Scalawag Administrator Formerly Important

    Joined:
    Mar 29, 2004
    Messages:
    51,347
    Location:
    Downtown
    Ratings:
    +57,473
    Thanks for this. I’m gonna have to vote on either RCV or Approval in Nov. Was leaning towards RCV as I am more knowledgeable of it and it is what the council passed and I generally default to them over initiatives but am now gonna have to dig into it.

    Voters will first decide on Question 1, asking whether either of the two proposed voting systems should be adopted. Voters would then decide on Question 2 to choose between Proposition 1A (Initiative 134) for approval voting or Proposition 1B (City Ordinance 126625) for ranked-choice voting. Voters opposed to adopting a new voting system who vote ‘no’ on Question 1 can still vote for their preferred option in Question 2. If the first question is approved by a majority of voters, the option receiving the highest number of votes would be adopted.​

    https://news.ballotpedia.org/2022/0...ked-choice-voting-for-city-primary-elections/

    Does the fact that this is for the primary, not the general change things? All of Washington has an open ‘top two’ or jungle primary where there is one list for all voters and the top two regardless of party go on to the general. This will just use RCV or Approval voting narrow the field down to the two who go on to the general.
    Last edited: Sep 22, 2022
  8. Order2Chaos

    Order2Chaos Ultimate... Immortal Administrator

    Joined:
    Apr 2, 2004
    Messages:
    25,186
    Location:
    here there be dragons
    Ratings:
    +21,380
    Yeah if I had the choice I'd pick STAR over Score over approval over RCV over FPTP.

    That said, this is apparently only for primary elections, which distorts things in ways that are not entirely clear without more info. For instance, is it FPTP in the general? Is it an open primary? Ratfucking becomes much easier in open approval voting primaries and then you could get extreme candidates in a plurality general, which means a result as far from the median voter as you can get. When deciding, watch the general election interactions carefully!
  9. Ancalagon

    Ancalagon Scalawag Administrator Formerly Important

    Joined:
    Mar 29, 2004
    Messages:
    51,347
    Location:
    Downtown
    Ratings:
    +57,473
    Yeah, see my edits.

    All of WA has open ‘top two’ or ‘jungle’ primaries.

    The top two go to the general where it is FPTP.

    So this would only pick the top two who will face off in the general.
  10. Jenee

    Jenee Driver 8

    Joined:
    Jan 20, 2008
    Messages:
    25,340
    Location:
    On the train
    Ratings:
    +19,334
    Can you provide a link to this? I’m not finding it.
  11. Order2Chaos

    Order2Chaos Ultimate... Immortal Administrator

    Joined:
    Apr 2, 2004
    Messages:
    25,186
    Location:
    here there be dragons
    Ratings:
    +21,380
    Turns out there exists a Nash Equilibrium for pickiness in approval voting. https://quantimschmitz.com/2022/10/...-you-vote-for-in-an-approval-voting-election/

    Even more surprising, it’s “however picky everyone else is”. That’s not a terrible place for it to be. In particular it shows that “bullet voting” (always voting only for your favorite candidate, equivalent to FPTP) is less likely to get you what you want than being less picky. That’s a pretty big finding and makes me feel better on a theoretical level about approval voting. Still, it’s a numerical finding via Monte Carlo simulation. It’s not terribly intuitive, so I don’t know if real voters will actually converge to it.