Discussion of HN at 4chan's /g/

Re:Discussion of HN at 4chan's /g/ by berekuk on Sunday July 29, 2012 @11:16AM (#4308673)

I don't think upvoting-on-agreement problem can be fixed technically. It's a cultural problem, and it should be fixed by internalizing the better model (see below), and then spreading it loudly enough and discouraging the default human reaction ("upvoting those who think like me"), because it ruins the signal-to-noise ratio in most of online communities.

The best rule on whether to upvote I've seen is this:

http://wiki.lesswrong.com/wiki/FAQ#What_are_the_moderation_r...

"Please do not vote comments up or down based mainly on whether you agree with them, unless the comment consists of a proposal and you wish to register your opinion regarding the proposal.

"Instead, vote on comments to the degree to which they improve the accuracy of your map. For example, a comment that you agree with that says nothing new should be voted down or left alone. A comment you initially disagree with that causes you to update some in favor of the author's opinion should be voted up."

[ Parent | on HN ]

Re:Discussion of HN at 4chan's /g/ by techiferous on Sunday July 29, 2012 @11:53AM (#4308774)

"I don't think upvoting-on-agreement problem can be fixed technically."

I think there is a technical fix: two orthogonal upvotes. One upvote/downvote pair for agreement, another for whether the post/comment added value. It's a pretty simple fix and I think it would work, but it could have other unwanted side-effects.

[ Parent | on HN ]

Re:Discussion of HN at 4chan's /g/ by stcredzero on Sunday July 29, 2012 @12:15PM (#4308834)

People would just hit both down arrows.

[ Parent | on HN ]

Re:Discussion of HN at 4chan's /g/ by techiferous on Sunday July 29, 2012 @12:24PM (#4308857)

Very good point.

This could still perhaps be solved technically, although it would take a lot of effort and is probably not worth it.

You can detect cheaters. People who give up/downvotes on the "value add" dimension that are not well correlated with the community's opinion of value add would be detected as cheaters, or at least detected as users who do not share the same values as the community.

Much harder to do, but possible, is to detect cheating in agreement by analyzing trends. Users who agreed with story A, B, and C may have a 90% chance of agreeing with story D. A user who bucks these trends enough can be detected as cheating or randomly voting.

[ Parent | on HN ]

Re:Discussion of HN at 4chan's /g/ by DanBC on Sunday July 29, 2012 @05:08PM (#4309750)

A few people would, but many others would be able to say "This is a quality comment, but stupid idea".

[ Parent | on HN ]

Re:Discussion of HN at 4chan's /g/ by stcredzero on Tuesday July 31, 2012 @06:23PM (#4319954)

The site could display the "down-correlation" number, which will be high if someone consistently hits both down arrows at once.

[ Parent | on HN ]

Re:Discussion of HN at 4chan's /g/ by fluxon on Sunday July 29, 2012 @04:13PM (#4309563)

I'd like the quoted text in the berekuk comment enabled as a tooltip on the uparrow, as a mouseover.

I'll contribute $10 for somebody to add this feature. Where's the "tip" jar?

[ Parent | on HN ]