For the 2020 Review, people have continued using the 7-point preliminary voting option throughout the Review Phase. I've personally found it fairly satisfying to use. I'm considering whether to add back the "fine-tuned quadratic voting" option from previous years in the Final Voting Phase.

I currently am mostly happy to leave it as-is (i.e. just use the 7-button voting system the entire time). The extra complexity cost of a separate voting system doesn't seem worth it. But I'm interested in hearing from anyone who wanted to make a strong bid (or, even a weak desire) for a return to the original system.

Some facts about how the current system is playing out:

  • 175 users have voted at all. 112 have cast at least 2 votes. 73 have cast 5+ votes.
  • The system doesn't let you know how many "points" you've spent, and most users have spent a lot less than their allotted 500 points (which are getting spent 'under the hood')
  • There are not very many downvotes.

My personal take is:

  • It's fine for many users to just be voting on a few posts they have an opinion on, rather than "maxing out their ballot."
  • I feel like when I look at the raw votes, they feel like they capture people actually voting how good they thought each post was, which makes them a more useful signal of how people felt about it (by contrast, last year the votes felt more like an arbitrary sorting signal, at least to me)
  • If I were starting over, I'd probably only have one "negative" vote option (which I'd set to -4) rather than 3 options.
    • Not enough people are using the negative options to justify the 3 extra buttons.
    • Having the "downvote" option be set to -4 (which costs 10 points), also removes the incentive to downvote everything you don't like to -1 (which is often a cheaper way to boost the relative values of posts you do like, but which makes "downvotes" a more confusing signal).
  • Meanwhile, if I were starting over, I might prefer to have an extra positive option. Right now most of my votes are "4s", but I'd kinda like to be able to signal two gradations of "medium post", and reserve 9s for "actually massively important."
  • (The previous two ideas probably aren't worth the complexity cost of switching-mid-stream this year, but my current guess is that I should implement them for next year)

Curious what other people think.

New to LessWrong?

New Comment
2 comments, sorted by Click to highlight new comments since: Today at 12:45 AM

Weak desire for quadratic voting. This is chiefly because it never shows up anywhere else, and there are very few areas of life where I care enough to vote and have the surplus capacity to actually engage with a new voting system.

If I don't endorse it in these conditions, then I effectively don't endorse new voting systems anywhere, which feels weird.

To determine which kind of voting makes sense, you should decide whether you want to have agreement to many posts with weak consensus or whether you want to reward polarizing posts. In the former kind of system, "the incentive to downvote everything you don't like to -1" is not a bug, it's a feature. I'd also prefer a system with 3 extra buttons to an asymmetric one.