This has the effect of making unpopular opinions invisible.
I don’t see how there’s a risk this turns into Reddit. The post you linked to said that people are more likely to upvote funny memes, rather than useful stuff, given some of pedanterrific’s comment that has basically already happened here. I don’t think humor is even that bad. And the question of whether we should not do so many upvotes doesn’t impact whether we should do more downvotes on already neutral posts, which is what I’m concerned with.
Having moderators solves the Reddit problem.
You also aren’t addressing the fact that in practice people are more likely to (down)(up)vote things which have already been (down)(up)voted, which leads to karma sinks.
This has the effect of making unpopular opinions invisible.
That doesn’t seem to happen in practice (yet) - downvoted posts are usually much more likely to have a tone or quality problem than be an unusual opinion.
You also aren’t addressing the fact that in practice people are more likely to (down)(up)vote things which have already been (down)(up)voted, which leads to karma sinks.
Possible, but that’s not a huge problem in itself; and the effect doesn’t seem very strong (it’s not rare to see posts eventually change sign).
This has the effect of making unpopular opinions invisible.
I don’t think this is in general a problem. Well argued positions will generally be upvoted. I for example have spent time arguing here against cryonics, the likelyhood of an intelligence explosion, and whether Kolmogorov complexity priors make sense. In all such cases I’ve been upvoted.
It is likely that a poorly argued argument for accepted views will not be downvoted as heavily as poorly argued arguments for contrarian views. But that’s a different claim than that unpopular opinions will be invisible.
This has the effect of making unpopular opinions invisible.
That only happens if people are misusing voting. And it does not happen in my experience.
There have been entire threads that were upvoted quite a bit but included disagreeing opinions. And we have a fair number of dissenters hereabouts. And some of us tend to weight comments slightly higher if they represent a minority position (or one I disagree with), to counter that sort of effect.
And the question of whether we should not do so many upvotes doesn’t impact whether we should do more downvotes on already neutral posts, which is what I’m concerned with.
Those were separate points. Jeff was talking about SO’s moderation system, I’m talking about ours. I pointed to Jeff’s post to help you understand why curation is important; it is separate information that up and down votes are our method of curation.
You also aren’t addressing the fact that in practice people are more likely to (down)(up)vote things which have already been (down)(up)voted, which leads to karma sinks.
I did not see anything about that in your comment. What is a ‘karma sink’ and why is it bad? I agree that if people have this tendency, that’s bad, though I do not know how I would even observe this putative tendency in the wild.
The Reddit thing is reasoning by anology at best because the argument is that curation is key to stop useless things from becoming popular. You’re also completely shifting the Reddit thing from what you initially said it was supposed to show, from “curation key to stop LessWrong to becoming Reddit” to “it shows why curation is important”, where the reason it was important was “because it stops Reddit” but yet you conceded that the reason it’s important is not “because it stops Reddit”, there’s an implicit contradiction. All Jeff’s article says is that memes aren’t productive. Rude things aren’t inherently unproductive. And 1⁄3 negative karma means that we would get rid of many productive things, so more downvoting is still an awful idea.
Below stuff I agree with, except I disagree with Emile about the intensity of karma sinks, I think the effect is pretty strong. Personally, I’m experiencing some, and people are going around hating on everything I do, which sucks. Karma sinks are where people are more likely to downvote things that have been downvoted. It’s like the Communism example I gave earlier in a few different ways.
I don’t think you’re correctly using the phrase “karma sinks.” Or at least, you’re not using it the way I see it is typically used on LW. Karma sinks refer to comments users make in order for other commentators to purposefully downvote if they upvoted another specific comment. This is so we can do things like straw polls without the polling user gaining tens or hundreds of karma.
That might be unclear, so I’ll give an example. I recently made a comment asking whether people would be interested in a New Jersey meetup. I could have also asked those interested to upvote my comment so I could have a rough estimate of how many people are interested. As per community norms, I would also make a second comment for users to downvote if and only if they upvoted the first comment. This second post (the karma sink) would help assure my net gain is 0.
Anyway, that doesn’t invalidate your point. But for clearer communication, you may want to use a phrase like “positive feedback loop.” That seems to better describe what you’re talking about, where downvotes make it more likely that other users will also downvote, which in turn leads to even more downvotes, and so on.
They are, in part because they predictably make people behave less rationally.
And 1⁄3 negative karma means that we would get rid of many productive things
I don’t agree with your implicit estimate that less than 1⁄3 of the comments here are net negative contributions. Still, noise would indicate that your conclusion is correct—I’d expect even now some large amount of productive things are hidden from view.
so more downvoting is still an awful idea.
That does not follow (that you thought it followed is indicated by your use of “so”). If you wanted that to follow, you would have to also establish that people here actually do downvote at least 1⁄3 of the comments here, which is clearly false—even I don’t downvote that much.
I assert that people are extremely stingy with their downvotes.
I assert that people are extremely stingy with their downvotes.
I anecdotally second this assertion: I hardly ever downvote. I only downvote comments that I think are egregiously bad, and then not when they’re already well buried, which they usually are by the time I get there.
You’re also completely shifting the Reddit thing from what you initially said it was supposed to show, from “curation key to stop LessWrong to becoming Reddit” to “it shows why curation is important”, where the reason it was important was “because it stops Reddit” but yet you conceded that the reason it’s important is not “because it stops Reddit”, there’s an implicit contradiction.
I’m confused. Where did I assert that “because it stops Reddit” (in shorthand) is not a reason that curation is important?
Karma sinks are where people are more likely to downvote things that have been downvoted.
More likely than what?
I would expect that people are more likely to downvote things that have been downvoted, than things that haven’t been, since things that have been downvoted are generally more downvote-worthy.
It is possible that he means that more likely than if the comment were already in the positives. It might be interesting to compare the voting patterns in this regard of people who do or do not use the anti-kibbitzer.
It is possible that he means that more likely than if the comment were already in the positives.
Yes, that one makes sense, but I think one still needs to fix the referent of “the comment” a little more explicitly. Comments in the positives are generally better comments than those in the negatives. This assertion has to be about counterfactuals, or changes in karma of the same comment over time.
A piece of evidence against that tendency: highly contentious issues where tempers are high tend to get downvoted very quickly, then upvoted to well above 0 over a short period of time. It seems like it would be very hard to get out of the negatives if this tendency existed and were worth noting.
This has the effect of making unpopular opinions invisible.
I don’t see how there’s a risk this turns into Reddit. The post you linked to said that people are more likely to upvote funny memes, rather than useful stuff, given some of pedanterrific’s comment that has basically already happened here. I don’t think humor is even that bad. And the question of whether we should not do so many upvotes doesn’t impact whether we should do more downvotes on already neutral posts, which is what I’m concerned with.
Having moderators solves the Reddit problem.
You also aren’t addressing the fact that in practice people are more likely to (down)(up)vote things which have already been (down)(up)voted, which leads to karma sinks.
That doesn’t seem to happen in practice (yet) - downvoted posts are usually much more likely to have a tone or quality problem than be an unusual opinion.
Possible, but that’s not a huge problem in itself; and the effect doesn’t seem very strong (it’s not rare to see posts eventually change sign).
I don’t think this is in general a problem. Well argued positions will generally be upvoted. I for example have spent time arguing here against cryonics, the likelyhood of an intelligence explosion, and whether Kolmogorov complexity priors make sense. In all such cases I’ve been upvoted.
It is likely that a poorly argued argument for accepted views will not be downvoted as heavily as poorly argued arguments for contrarian views. But that’s a different claim than that unpopular opinions will be invisible.
That only happens if people are misusing voting. And it does not happen in my experience.
There have been entire threads that were upvoted quite a bit but included disagreeing opinions. And we have a fair number of dissenters hereabouts. And some of us tend to weight comments slightly higher if they represent a minority position (or one I disagree with), to counter that sort of effect.
Those were separate points. Jeff was talking about SO’s moderation system, I’m talking about ours. I pointed to Jeff’s post to help you understand why curation is important; it is separate information that up and down votes are our method of curation.
I did not see anything about that in your comment. What is a ‘karma sink’ and why is it bad? I agree that if people have this tendency, that’s bad, though I do not know how I would even observe this putative tendency in the wild.
The Reddit thing is reasoning by anology at best because the argument is that curation is key to stop useless things from becoming popular. You’re also completely shifting the Reddit thing from what you initially said it was supposed to show, from “curation key to stop LessWrong to becoming Reddit” to “it shows why curation is important”, where the reason it was important was “because it stops Reddit” but yet you conceded that the reason it’s important is not “because it stops Reddit”, there’s an implicit contradiction. All Jeff’s article says is that memes aren’t productive. Rude things aren’t inherently unproductive. And 1⁄3 negative karma means that we would get rid of many productive things, so more downvoting is still an awful idea.
Below stuff I agree with, except I disagree with Emile about the intensity of karma sinks, I think the effect is pretty strong. Personally, I’m experiencing some, and people are going around hating on everything I do, which sucks. Karma sinks are where people are more likely to downvote things that have been downvoted. It’s like the Communism example I gave earlier in a few different ways.
I don’t think you’re correctly using the phrase “karma sinks.” Or at least, you’re not using it the way I see it is typically used on LW. Karma sinks refer to comments users make in order for other commentators to purposefully downvote if they upvoted another specific comment. This is so we can do things like straw polls without the polling user gaining tens or hundreds of karma.
That might be unclear, so I’ll give an example. I recently made a comment asking whether people would be interested in a New Jersey meetup. I could have also asked those interested to upvote my comment so I could have a rough estimate of how many people are interested. As per community norms, I would also make a second comment for users to downvote if and only if they upvoted the first comment. This second post (the karma sink) would help assure my net gain is 0.
Anyway, that doesn’t invalidate your point. But for clearer communication, you may want to use a phrase like “positive feedback loop.” That seems to better describe what you’re talking about, where downvotes make it more likely that other users will also downvote, which in turn leads to even more downvotes, and so on.
I think you mean “positive feedback loop” of downvotes.
Ah, silly me. Correcting some else while making a mistake. You’re correct, thanks.
Another possibility might be “information cascade”, which refers to this specific phenomenon.
They are, in part because they predictably make people behave less rationally.
I don’t agree with your implicit estimate that less than 1⁄3 of the comments here are net negative contributions. Still, noise would indicate that your conclusion is correct—I’d expect even now some large amount of productive things are hidden from view.
That does not follow (that you thought it followed is indicated by your use of “so”). If you wanted that to follow, you would have to also establish that people here actually do downvote at least 1⁄3 of the comments here, which is clearly false—even I don’t downvote that much.
I assert that people are extremely stingy with their downvotes.
I anecdotally second this assertion: I hardly ever downvote. I only downvote comments that I think are egregiously bad, and then not when they’re already well buried, which they usually are by the time I get there.
I’m confused. Where did I assert that “because it stops Reddit” (in shorthand) is not a reason that curation is important?
More likely than what?
I would expect that people are more likely to downvote things that have been downvoted, than things that haven’t been, since things that have been downvoted are generally more downvote-worthy.
It is possible that he means that more likely than if the comment were already in the positives. It might be interesting to compare the voting patterns in this regard of people who do or do not use the anti-kibbitzer.
Yes, that one makes sense, but I think one still needs to fix the referent of “the comment” a little more explicitly. Comments in the positives are generally better comments than those in the negatives. This assertion has to be about counterfactuals, or changes in karma of the same comment over time.
A piece of evidence against that tendency: highly contentious issues where tempers are high tend to get downvoted very quickly, then upvoted to well above 0 over a short period of time. It seems like it would be very hard to get out of the negatives if this tendency existed and were worth noting.
Typo? That sounds contradictory.
Yes, edited to fix that while you were replying.