• Natanael@infosec.pub
    link
    fedilink
    arrow-up
    2
    arrow-down
    1
    ·
    edit-2
    4 hours ago

    EDIT: I’m describing single winner elections.

    Every round, you count one new step of rankings, and eliminate the lowest candidate.

    Round one starts with simply counting all 1st rank votes first.

    Everybody’s vote counts once, to one person. It goes to that voter’s highest still-in-the-running candidate.

    In the first round, if some candidate already gets 50%, they win directly. If not, you eliminate the person with the lowest votes.
    (some variants pre-filter all candidates with too few total votes to have a chance, especially people who didn’t even get ranked by most, or only low ranks)

    What happens now is that people whose 1st vote was for the eliminated person gets redirected to their next highest person still in the running for the next round on the ballot which they cast. So in round 2 that’s your rank 2 on your ballot. Each remaining candidate now has their 1st rank votes from round 1 plus also their 2nd rank votes from all ballots which had been cast to the now-eliminated candidate.

    Note that in each consecutive round, all votes which was already counted to a person who still in the running in that round remains untouched. So all who picked a 1st which is still in are still getting counted to that person. You don’t look at your 2nd rank yet if your 1st is still active.

    So now in round 2, after the redirection to second choices, if that puts somebody over 50% they win. If not you eliminate yet another person, the lowest one every time. Every vote counted to that eliminated person, either 1st’s or 2nd’s, now gets redistributed again to that voter’s next highest person still in the running on their cast ballot - if your 1st is eliminated now, it goes to the 2nd if that person still is in, if not it goes to your 3rd. If you were already on your 2nd and that person is eliminated, it also goes to your 3rd. If somebody’s 1st and 2nd are eliminated AND they don’t have a 3rd, their vote gets ignored in the rest of the tallying process.

    In round 3 and on you just keep going, until somebody gets 50% or all but one is eliminated. You may force a revote in many places if that person is beneath some threshold.

    • Rivalarrival@lemmy.today
      link
      fedilink
      English
      arrow-up
      2
      ·
      12 hours ago

      Please rewatch the video from about 4:40. Your comment doesn’t address the “White Tiger”, 65% issue.

      Ideally, white tiger, purple tiger, and green gorrilla would be elected, as this is the outcome most proportional to voter intent.

      But under the logic you describe, purple tiger, with the least votes, would be eliminated after the first round. 2/3 of the populace have indicated they want tigers, but we’re about to empanel one tiger and two gorillas.

      In actuality, we are not removing the least popular candidate in each round. We are only removing the least popular option if the most popular option does not exceed the mathematical threshold for locking a seat. If they do exceed that threshold, we are removing the top candidate - not the bottom - before we count the next round. The top candidate gets the first seat, and the remaining candidates are competing for the second and third seats.

      • Natanael@infosec.pub
        link
        fedilink
        arrow-up
        1
        ·
        edit-2
        4 hours ago

        I was describing single winner, you’re describing multiple winner

        Those are implemented very differently.

        You can guarantee that popular choices gets represented (assuming enough seats for all popular choices) if you do pre-filtering rounds where you give weighted scoring to candidates based on their accumulated ballot rank positions. This would quickly sort out many last rank choices and few-first choices. Then you can proceed with the rounds of eliminations.

        Practically speaking, perfect fairness is mathematically proven impossible. You have to make tradeoffs and pick a system that will likely get a good enough result

        https://en.wikipedia.org/wiki/Arrow's_impossibility_theorem

        https://www.sciencedaily.com/releases/2026/08/260801042812.htm

    • Sharkticon@lemmy.zip
      link
      fedilink
      arrow-up
      2
      ·
      13 hours ago

      No I’m sorry I don’t think you watched the same video I did or you didn’t read my comment very carefully. You’re basically describing ranked choice voting which is very simple yeah. That was not my question.