Navigation

    Voting Theory Forum

    • Register
    • Login
    • Search
    • Recent
    • Categories
    • Tags
    • Popular
    • Users
    • Groups

    Ranked Robin - which preference matrix is correct?

    Single-winner
    4
    5
    296
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • masiarek
      masiarek last edited by

      8:Ava>Cedric>Deegan>Bianca>Eli
      6:Ava=Bianca=Cedric>Eli>Deegan
      6:Eli>Ava>Bianca=Cedric=Deegan
      6:Deegan>Bianca=Cedric>Eli>Ava
      4:Bianca>Ava>Eli>Deegan>Cedric
      3:Eli>Deegan>Bianca=Cedric>Ava
      2:Deegan=Eli>Bianca=Cedric>Ava

      1. preference matrix at https://electowiki.org/wiki/Ranked_Robin:
        3359b14a-286b-4105-befd-b81ff2f1c124-image.png

      2. preference matrix at - https://www.cs.angelo.edu/~rlegrand/rbvote/calc.html
        31b1b5b7-94da-4dd6-af4a-5cadd9cdb27b-image.png

      C 1 Reply Last reply Reply Quote 0
      • C
        cfrank @masiarek last edited by cfrank

        This post is deleted!
        1 Reply Last reply Reply Quote 0
        • J
          Jack Waugh last edited by Jack Waugh

          Maybe the differences across the values at i, j and j, i are the same and are all that matter? @Sass?

          I suspect that the ~rlegrand tool is counting equal rankings by giving a half point to each candidate in the pair.

          Approval-ordered Llull (letter grades) [10], Score // Llull [9], Score, STAR, Approval, other rated Condorcet [8]; equal-ranked Condorcet [4]; strictly-ranked Condorcet [3]; everything else [0].

          1 Reply Last reply Reply Quote 0
          • Sass
            Sass last edited by Sass

            I just manually checked mine (the first one) and it's correct. I suspect the second one counts equal preferences as half points for each candidate in the pairwise comparison. I haven't seen that before, but there is a metric that a voting method could use that would be screwed up by doing it: total number of pairwise preferences over all candidates. That would be sum of the entire row of a candidate. Candidates who are ranked equally to other candidates on more ballots would benefit. However, that's an unreliable metric generally and is really only useful for tiebreaking. So basically my conclusion is that both approaches are correct and should give you the same results, at least for Ranked Robin (not including the 3rd degree tiebeaker).

            1 Reply Last reply Reply Quote 1
            • J
              Jack Waugh last edited by

              Maybe it would make sense to regard a canonical preference matrix as having only an upper triangle. However, for software, I think it is convenient to represent it using the whole matrix.

              Approval-ordered Llull (letter grades) [10], Score // Llull [9], Score, STAR, Approval, other rated Condorcet [8]; equal-ranked Condorcet [4]; strictly-ranked Condorcet [3]; everything else [0].

              1 Reply Last reply Reply Quote 0
              • First post
                Last post