Strong Votes [Update: Deployed]

Up­date: This has now been pushed to the live site. For now, on Desk­top, strong up­votes re­quire click-and-hold. Mo­bile users just tap mul­ti­ple times. This is most likely a tem­po­rary solu­tion as we get some feed­back about how the re­spec­tive modes work un­der re­al­is­tic con­di­tions.

Over on our dev-site, less­es­t­, we have a fairly ma­jor fea­ture-branch: Re­cal­ibrated vot­ing power, and Strong Upvotes/​Down­votes.

tl;dr – Nor­mal votes on the test server range in power from 1 to 3 (de­pend­ing on your karma). You have the op­tion of hold­ing down the up/​down­vote but­ton for a strong vote, which ranges in power from 1 to 15.

We’re look­ing for feed­back on the UI, and some im­ple­men­ta­tion de­tails.

Flow vs Stock

This post by Jame­son Quinn notes that there’s two com­mon rea­sons to up­vote or down­vote things. This is similar to my own schema:

  • Con­ver­sa­tional Flow – When you like (or dis­like) the effect a com­ment has on a con­ver­sa­tion, and you want to give the au­thor a metaphor­i­cal smile of ap­pre­ci­a­tion, or awk­ward silence/​stare.

    • “Ah, good point” (+)

    • “Hmm, this gives me some­thing to think about” (+)

    • “This com­ment cited sources, which is rare. I want to re­ward that.” (+)

    • “This was clever/​funny.” (+)

    • “I think this post con­tains an er­ror.” (–)

    • “This com­ment is tech­ni­cally fine but an­noy­ing to read” (–)

    • “I don’t think the au­thor is be­ing very char­i­ta­ble here” (–)

    • Some com­bi­na­tion of the above (up­vote or down­vote, de­pend­ing)

  • Sig­nify­ing Im­por­tance – When you think other peo­ple should go out of their way to read some­thing (or, definitely should not). Ideally, posts and com­ments that con­tribute to the longterm stock of value that LessWrong is ac­cu­mu­lat­ing.

    • “I learned some­thing new and use­ful” (++)

    • “The ar­gu­men­ta­tion or thought pro­cess illus­trated by this post helped me learn to think bet­ter.” (++)

    • “This post con­tains many fac­tual er­rors” (––)

    • “This com­ment is literal spam” (––)

    • “The rea­son­ing here is deeply bad.” (––)

Peo­ple in­stinc­tively use up­vot­ing to cover both Flow and Im­por­tance, and this of­ten re­sults in peo­ple up­vot­ing things be­cause they were a good thing to say in a con­ver­sa­tion. But then later, if you want to find the most use­ful com­ments in a dis­cus­sion, you end up sift­ing through a bunch not-ac­tu­ally-use­ful stuff.

Peo­ple also of­ten un­re­flec­tively up­vote things they like, with­out pay­ing much at­ten­tion to whether the ar­gu­ments are good, or whether it’s good for the longterm health of the site. This means peo­ple who think hard about their up­votes get counted just as much as peo­ple ca­su­ally click­ing.

So the idea here is that by de­fault, click­ing re­sults in a Nor­mal Upvote. But, if you hold the but­ton down for a cou­ple sec­onds, you’ll get a Strong Upvote. (And same for down­votes).

Can you tech­ni­cally Strong Upvote ev­ery­thing? Well, we can’t stop you. But we’re hop­ing a com­bi­na­tion of mostly-good-faith + triv­ial in­con­ve­niences will re­sult in peo­ple us­ing Strong Upvotes when they feel it’s ac­tu­ally im­por­tant.

I have some more thoughts on “what good effects pre­cisely are we aiming for here”, which I’ll flesh out in the com­ments and/​or the fi­nal blog­post when we ac­tu­ally de­ploy this change to pro­duc­tion.

Vote-Power by Karma

Quick overview of the ac­tual num­bers here (vote-power/​karma)

Nor­mal votes

  • 3 – 25,000 karma

  • 2 – 1,000 karma

  • 1 – 0 karma

Strong Votes

  • 16 – 500,000 (i.e. Thou­sand year old vam­pire—the level above Eliezer)

  • 15 – 250,000

  • 14 – 175,000

  • 13 – 100,000

  • 12 – 75,000

  • 11 – 50,000

  • 10 – 25,000

  • 9 – 10,000

  • 8 – 5,000

  • 7 – 2,500

  • 6 – 1,000

  • 5 – 500

  • 4 – 250

  • 3 – 100

  • 2 – 10

  • 1 – 0

(We con­sid­ered us­ing an­other log scale, but log5 didn’t quite give us the gran­u­lar­ity we wanted, and smaller log scales pro­duced weird num­bers that just didn’t re­ally cor­re­spond to the effect we wanted. So we just picked some num­bers that felt right.)


We’re still hash­ing out the ex­act UI here – in par­tic­u­lar, the UI for helping users dis­cover the fea­ture. (Posts ba­si­cally have lit­tle-to-know dis­cov­er­abil­ity, com­ments have a lit­tle hover-over mes­sage).

Check out less­es­t­ and note your feed­back here. (If you cre­ated your user ac­count re­cently, you may need to cre­ate an al­ter­nate ac­count on the less­es­t­wrong de­vel­op­ment database)

No nominations.
No reviews.