There has been a proposal for not using replies for voting.

This is cleaner and doesn’t expose everyone’s votes to anyone reading the thread.

I have a notion of how this would look, so I’m creating this discussion and the replies for each option as an example. Please discuss or upvote any of the options I wrote below, and we can all see how this would look.

Should registration of this instance be:

  • haxe11@sh.itjust.works
    link
    fedilink
    arrow-up
    1
    ·
    1 year ago

    Personally I can’t think of how voting can be done cleanly within just Lemmy yet. It seems like there should be a special type of post developed for it.

    I see a couple issues with your proposal. Actual discussion can either drown out the options to be voted, especially the ones that are being downvoted. Second, the snowball effect would be even worse than comment-based voting, as the earliest upvoted options would be on the top of the thread.

  • carbon_based@sh.itjust.works
    link
    fedilink
    arrow-up
    1
    ·
    edit-2
    1 year ago

    Perhaps add explainer:
    This poll is based on the preceeding discussion: [link]. Please make yourself familiar with the discussion before voting.
    There is a choice between several options.
    Upvote those options you would agree with, downvote those you would not agree with.

    (comment to abstain? I know this could be abused. …)

    Add an option: “Do not change the current mode of registration.

    … and i like to edit: Some information on how the final choice will be made may also be helpful (this could differ by subject): “The option on which the ratio of upvotes [(sum-of-votes – downvotes) / sum-of-votes] exceeds that of all others by at least 1.5 : 1 will be implemented; should none meet that target no change will be made.

    • tcely@sh.itjust.worksOP
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 year ago

      I like some of those suggestions. However, I don’t think down voting or abstaining should be supported. You either support an option by up voting or you take no action.

      Approval voting systems have well studied behaviors and we should not deviate from that without a compelling reason.

      Every approval / upvote is a distinct user endorsement for an option. The option with the most users endorsing it should be selected when that number exceeds 51% of the active users set we decide on.

      • carbon_based@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        1
        ·
        edit-2
        1 year ago

        That could be another way to gauge the outcome. I just suggested to get into account the actual disapproval also (if we were that advanced, i’d wish we could have weighted disapproval voting anyway, but i can’t hope that people understand the advantages of that). Why would my way of counting the approval ratio rather than approval-only be of any disadvantage? I mean, that way it would include the total number of participants as well, and an approval ratio would be a more robust decision-maker than just a fixed goal of absolute agreement. And why should i not want to improve on stuff? :-)

        But anyway, this was meant to suggest how to improve the ballot itself (after all, that’s the topic of the discussion here). So aside from how ever votes are to be counted, i suggest to include all that information in the ballot.

        • tcely@sh.itjust.worksOP
          link
          fedilink
          English
          arrow-up
          2
          ·
          1 year ago

          Why would my way of counting the approval ratio rather than approval-only be of any disadvantage?

          That’s just how voting systems work. Even seemingly insignificant changes to the algorithm can have outsized impacts on how well it performs.

          Plurality versus Approval is a specific example of this. Just changing “choose one” to “choose as many as you approve of” significantly impacts the amount of data that’s captured and often the outcome because of the effects on voters’ behaviors.

          I liked the suggestion of more information. When we have it all figured out those details should be included as you suggested.

          • carbon_based@sh.itjust.works
            link
            fedilink
            English
            arrow-up
            0
            arrow-down
            1
            ·
            1 year ago

            Thanks for the recognition!

            It is all true that different ways of evaluation will lead to different results. That can be understood though, and i think that i can wrap my mind around the maths of it (yet perhaps not everyone could, no offence intended). Some ways are more appropriate in certain context, some are inherently unfair, etc. – My suggestion should have been meant towards more accuracy …

            But maybe let’s forget about this entirely, and look here: Ranked Choice Voting https://sh.itjust.works/post/311690

            • tcely@sh.itjust.worksOP
              link
              fedilink
              English
              arrow-up
              1
              ·
              1 year ago

              RCV/IRV has been a scam for more than a century. Please look into how that was invented and how it actually performed. More than half the places that tried it actually removed it.

              I don’t know why people keep being fooled by complexity, but they do.

              • carbon_based@sh.itjust.works
                link
                fedilink
                English
                arrow-up
                0
                arrow-down
                1
                ·
                edit-2
                1 year ago

                … aha … now slowly i get some dissonance. You just advertised a system that doesn’t count dissenting votes, therefore making it impossible to find the real acceptance of some choice. I may look into how Ranked Choice was “invented” (hey, anyone could come up with such a thing). But please stop to belittle people just because you seem to not grasp some modest complexity.
                I can see for example, that not being able to rank options equally, will give an advantage to that which is listed first on the ballot. That is easy to see for me. My mind makes pictures.