• 0 Posts
  • 70 Comments
Joined 3 years ago
cake
Cake day: January 4th, 2024

help-circle


  • It doesn’t matter if your uniformed vibe check passes, because IRV has serious problems.

    It’s like looking at a house built out of rotted timber found in the woods and thinking, sure, the outside looks good, it must be good. That other house with it’s fancy modern design sure doesn’t look a falling down log cabin, so it must not be a good house.

    IRV has been shown to reinforce two party dominance via application of the spoiler effect. It has had multiple real world elections where the pair-wise loser was elected because of a quirk that’s unique to IRV.

    STAR on the other hand has been designed by actual experts, I know people don’t seem to trust those anymore, but then people are also dying from measles again. I guess we’re just choosing the absolute worst option on principle these days.


  • How is rating each candidate on a scale of 0-5 harder than ranking them in some arbitrary order?

    Rankings lead to worse results than ratings, and are just less satisfying to use. What if you like two candidates the same? Better chose one over the other and then hope that other people in your situation ranked the two in the same order, otherwise one of the two will act as a spoiler for the other. Arrow’s Impossibility Theorem clearly says that all ranked voting methods will have a spoiler candidate, so you’ll run face first into that.

    But rated systems do not have a spoiler effect, if you’re using STAR, you rate each candidate independently of each other. The system doesn’t care if you like A more than B, or like them the same, it just cares what the total rating for each candidate is. With the added step of looking at the top two candidates and seeing which one was rated higher on each ballot.






  • Except it doesn’t at all.

    All Ordinal voting systems trend towards two party dominance. Because all Ordinal voting systems have some degree of spoiler effect.

    The Marquis de Condorcet wrote about the mathematical flaws of a system identical to Ranked Choice back in 1785. His chief complaint was that it could not reliably elect the pair-wise winner, that is, the person who would win a one on one election against every other candidate. Condorcet found that the more candidates that you add, the less likely you’ll see the pair-wise winner be elected.

    There is hope, Cardinal voting systems exist, and tend to work better the more candidates you have (to a point).

    STAR was designed back in 2014. It’s good.