There has been a great deal of discussion here about the proper methods of voting on comments and on how karma should be assigned.  I believe it's finally reached the point where a post is warranted that covers some of the issues involved.  (This may be just because I find myself frequently in disagreement with others about it.)

The Automatic Upvote

First, there is the question of whether one should be able to upvote one's own comment.  This actually breaks apart into two related concerns:

(1) One is able to upvote one's own comments, and

(2) One gains a point of karma just for posting a comment.

These need not be tied.  We could have (2) without (1) by awarding a point of karma for commenting, without changing the comment's score.  We could also have (1) without (2) by simply not counting self-upvotes for karma.

I am in favor of (2).  The main argument against (2) is that it rewards quantity over quality.  The main argument for (2) is that it offers an automatic incentive to post comments; that is, it rewards commenting over silence.  As we're community-building, I think the latter incentive is more important than the former.  But I'm not sure this is worth arguing further - it serves as a distraction from the benefits of (1).

I am also in favor of (1).  As a default, all comments have a base rating of 0.  Since one is allowed to vote on one's own comments, and upvoting is the default for one's own comments, this makes comments effectively start at a rating of 1.  The argument against this is that it makes more sense for comments to start with a rating of 0, so that someone else liking a comment gives it a positive rating, while someone disliking it gives it a negative rating.  I disagree with this assessment.

If I post a comment, it's because it's the best comment I could think of to add to the discussion.  I will usually not bother saying something if I don't think it's the sort of thing that I would upvote.  When I see someone else's comment that I don't think is very good, I downvote it.  Since they already upvoted it, I'm in effect disagreeing that this was something worth saying.  The score now reflects this - a score of 0 shows that one person thought it was a worthwhile comment, and one person did not.

Furthermore, if I was not able to vote on my own comments, I would be much more reluctant to upvote.  Since I would not be able to upvote my comment, upvoting someone else's comment would suggest that I think their comment is better than my own. But by hypothesis, I thought my comment was nearly the best thing that could be said on the subject; thus, upvotes will be rare.

And so I say that we implement a compromise - (1) and not (2).

What should upvote/downvote mean?

I think it is established pretty well that upvote means "High quality comment" or "I would like to see more comments like this one", while downvote means "Low quality comment" or "I would like to see fewer comments like this one".  However, this definition still retains a good bit of ambiguity.

It is too easy to think of upvote and downvote as 'agree' and 'disagree'.  Even guarding myself against this behavior I find the cursor drifting to downvote as soon as I think, "Well that's obviously wrong".  But that's clearly not what the concept is there for.  Comments voted up appear higher on the page (on certain views), which allows casual readers to see the best comments and discussions on any particular post.  If we use upvote and downvote to mean 'agree' and 'disagree', then this is effectively an echo chamber, where the only comments to float to the top are the ones that jive with the groupthink.

Instead, upvote and downvote should reflect overall quality of a comment.  There are several criteria I tend to use to judge a good comment (this list is not all-inclusive):

  1. Did the comment add any information, or did it just add to the noise? (+)
  2. Does the comment include references or links to relevant information? (+)
  3. Does the comment reflect a genuinely held point-of-view that adds to the discussion? (+)
  4. Is the comment part of a discussion that might lead somewhere interesting? (+)
  5. Is the comment obvious spam / trolling? (-)
  6. Is the comment off-topic? (-)

Since we feel the need to voice whether we agree or disagree with comments, but 'I agree' and 'I disagree' comments are noisy, it's been suggested that there should be separate buttons to indicate agreement and disagreement.  Thus, someone posting a well-argued on-topic defense of theism can get the upvote and 'disagree', while someone posting an off-topic 'physicalism is true' can get the downvote and 'agree'.  Presumably, we'd only count upvotes and downvotes for karma, but we could use 'agree' and 'disagree' for "most controversial" or other views/metrics.

Whether votes should require an explanation

It has been suggested that votes, or downvotes specifically, should require an explanation.  I disagree with both sentiments.  First, requiring explanations for downvotes but not upvotes would bias the voting positively, which would have the effect of rewarding quantity over quality and decrease the impact of downvotes.

But requiring explanations for votes is in general a bad idea.  This site is already a burden to keep up with; for those of us that do a lot of voting, writing an explanation for each one would be too much time and effort.  Requiring an explanation for every vote would doubtless result in a lot less voting.  Also, explaining votes is almost always off-topic, so adds to the noise here without really contributing to the discussion.

Note Yvain's more personal rationale:

I'm not prepared to write an essay explaining exactly was wrong with each of them, especially if the original commenter wasn't prepared to take three seconds to write a halfway decent response.

Adding to the burden of those already performing the service of voting unduly penalizes those who are doing good, to the end of appeasing those who are contributing to the noise here.

Relevant Comments

For reference, some links to relevant posts and sections of comments.  I tried to be inclusive, since there have been a lot of discussions about these issues - more relevant ones hopefully near the top. (Please comment if you know of any other relevant discussions)

1 2 3 what upvote and downvote should mean and whether there should be agree/disagree buttons

4 whether karma should be the sum of individual post scores, or (perhaps) an average

5 super-votes

6 The utility of comment karma

7 whether one should unselect the self-upvote

8 9 whether downvotes should require explanation

10 whether Eliezer Yudkowsky gets fewer upvotes than others

11 whether karma can be used to gauge rationality

12 whether people downvote for disagreeing with groupthink

13 whether karma promotes a closed-garden effect

14 whether administrators should delete comments entirely

15 Lesswrong Antikibitzer: tool for hiding comment authors and vote counts

ETA: I might concede that this post is possibly off-topic for Less Wrong - but the blog/community site about "Less Wrong" does not exist yet, so this seems like the best place to post it.

ETA2: Public records of upvotes/downvotes might solve some of these problems; discuss.


New Comment
47 comments, sorted by Click to highlight new comments since: Today at 8:33 PM

I don't think there's enough downvoting going on - in particular, comments of such low quality that I would not wish them to insult the eyes of new users judging us, are not successfully voted down to -4 and hidden. It seems we're wandering into a norm where 0 is insult enough, -1 is terrible, -4 is hardly imaginable.

Those of you who are not familiar with the literature on online communities should bear in mind that online communities die primarily as a result of failing to solve the problem of quality control, and that refusing to accept the unfortunate necessity of quality control is a primary reason. So there are broken windows and they attract hoodlums, and the higher-quality recruits encountering the community for the first time decide to go elsewhere. And this has happened over and over again since before the days of the Eternal September.

Here, the quality control is downvoting, but people are refusing to use it. It has turned into something awful, horrible, unspeakable, a punch in the nose that requires a full-blown court drama. No community can defend its quality standards in such a fashion.

Downvoting really should not be that awful. And so I hope that starting all comments out at 0 will encourage more downvoting, which will make a score of -1 seem less awful, which will encourage even more downvoting, and so LW will not go the way of so many other online communities that tried to be nice and refused to defend their quality standards.

IAWYC, and I'll start downvoting comments more. I've been hesitating because I'm reluctant to hurt peoples' feelings, to make participation a net downer instead of a net mood-lifter (because people are supposed to be more saddened by losses than they are boosted by gains), and to discourage participation from new users who might, if they hang around, long-term add to the community. But now that you point it out, yes, quality dilution is a bigger risk.

Is it worth trying to reduce the negative side-effects of downvoting that were making me hesitate, e.g. by placing signs everywhere that you should expect your comments to have an average score around zero and shouldn't be discouraged or feel you ought not participate if your comments are voted down, or by causing people to gain karma for comments even when those comments are not up-voted? I'm honestly not sure.

Part of the community's hesitation to down-vote may be caused by your posts on how we shouldn't be so afraid of cooperating and expressing approval. It might be worth clarifying how to be a warm, accepting community and have quality standards. I'm sure there are ways this can work, but I don't have a detailed picture of what they look like and how to help toward them.

Is it worth trying to reduce the negative side-effects of downvoting that were making me hesitate,

Absolutely - that's one reason a lot of people are looking for a good 'about' page or FAQ that clearly explains things like voting and karma. But we need to decide definitively what voting is for before we can explain it to everyone.

Definitely experiment. Here's another: start at +6, vanish at +2. That way, a vote would look small compared to the total score and no one would have to push through 0.

The programming language community blog Lambda the Ultimate had similar problems some years ago and seems to have emerged successful. Guess having a benevolent dictator editor is important. Here's the relevant discussion from 2006. My own two cents there, that I still stand by today:

The only way to improve the quality of discussions is to filter out the bozos. Don't moderate postings; moderate people.

For me, read-only access to "old LtU" would be more pleasant than read-write access to "new LtU". In other words: I don't think bans are "extreme", my vote is for a lot more bans, and I wouldn't mind being banned myself if this improves the signal-to-noise ratio.

Edit: it seems someone took my words to heart and started downvoting my whole comment history without explanation. Nice! Sadly I can't tell if it's happening to anyone else.

It happens to me regularly - my karma drops by roughly the number of comments I've posted since the last time it happened. It's annoying, but since my overall karma still grows, it's clearly noise.

Also, it's possible the same person is downvoting everybody in this thread; I had a drop of about 10 points today. (I skipped posting any comments this weekend, so my overall comment rate has gotten kind of low lately.)

Perhaps someone is just trying to "win" at karma?

I just lost 20 points in half an hour.

A similar thing happened to me a while back.

If somebody went to your userpage and then sequentially voted down all your comments in a short period of time based on nothing related to quality control (which is probably what happened here if the downvotes were by one person), the system should be able to spot that sort of abuse and ban the user or take other action.

If that was automatic, it would be a problem for me. I often go back through the recent comments listing and do a flurry of quality-control voting; otherwise, comments on older posts slip through the cracks.

I worry that some people may hold conflicting views on what the comments are for.

  • First, comments may act as little notes, contributing a piece of information or insight, relevant to the context in which it's made.
  • Second, comments may serve communication, letting a specific person know something, correcting a mistake, voicing an opinion, adding a detail to resolve an earlier misunderstanding.

Relevant communication must not be punished. There is no way to support a conversation without these communication comments that are not intended to hold generally appreciated pearls of wisdom in them.

I think this better reflects conflicting views on what voting is for. I don't vote something down as punishment; I vote it down because I don't think it's interesting enough to belong at the top of the page.

If posting a certain kind of comments results in expected decrease in karma, people will avoid doing that, so there is an inseparable component of punishment. What you say would only work if there were two separate comment ratings, one for other readers, and one for the writer. Essentially, the hard/soft voting distinction developed to play this role.

I suppose that's true if people would really take a decrease in karma as a reason not to post a comment.

If people are doing that, they probably should be punished for it.

I'll see about adding the ability to vote down comments that weren't posted for not being posted.

My (somewhat limited) knowledge of Python suggests to me that this would not be an easy task; I daresay it might be nearly impossible.

But I agree - we should implement something like that if we can.

I suppose that's true if people would really take a decrease in karma as a reason not to post a comment.
If people are doing that, they probably should be punished for it.

You are evil.

You are evil.

I have to know - in what sense did you mean this? "not morally good"? "deviously scheming to align the world with my preferences"? "hates paladins"?

For now, I'll just take it as a compliment.

I said userpage, not recent comments. If you don't often view all comments of just a single user and then issue 20 or more downvotes in quick succession, I wasn't suggesting anything that would affect you. There are lots of abuses that can't easily be caught, but abuse like I've outlined, which based on comments here has occurred a few times recently, should be easy to catch.

Woah, I also lost 20 in half an hour. Doesn't look like a coincidence.

I've not only noticed this happening to me (which is to be expected), but to a number of other people, most especially those who have seemingly earned the dislike of the local Powers That Be.

I'm considering stepping in to correct such injustices, but thus far my desire to vote up/down based solely on individual merit has prevented me. If it keeps up, let us know, and I'll start changing my strategy.

I'm considering stepping in to correct such injustices, but thus far my desire to vote up/down based solely on individual merit has prevented me. If it keeps up, let us know, and I'll start changing my strategy.

Don't. It probably happens to people who like the local Powers That Be, too. My karma dropped 17 points in 10 minutes some days ago, for no reason I was aware of. Karma doesn't matter all that much, and not acting like power-struggling children about karma does matter. Even if there is systematic injustice in some particular direction, there's better information-flow in pointing out the distortion and letting people consciously realize that karma is unreliable in such-and-such a way, than in adding further information-poor voting to try to compensate.

The only way to improve the quality of discussions is to filter out the bozos. Don't moderate postings; moderate people.

I disagree. Most people who make high-quality comments also make a large number of low-quality comments. Filtering out bad comments is essential if we want the good ones to float to the top.

Sure, but usually it's only a very few folk who make actively damaging comments - comments that for no reward absorb more time and energy from the community than just the time to read them because of the responses they attract.

I'd like to filter out actively damaging comments and the large number of low-quality comments. Banning only people will not help with that.

Median-based, stars-out-of-five voting might encourage downvoting, because you can make it clear with your vote that you're saying "not all that great" rather than "terrible, should be burned".

Still, there's a separate exercise in persuading people that thinking "this comment is just trash" doesn't necessarily mean you're hopelessly in the grip of confirmation bias. This probably deserves a top-level post of its own.

The hesitation to make a vote that is too strong leads to the lack of voting. So, in the light of my observation about hard/soft voting, I suggest making 4 buttons that explicate the distinction: (--, -, +, ++), where the central buttons correspond to the soft votes, and outside buttons to more aggressive hard votes.

How would the separate agree/disagree button help?

The agree/disagree button gives people an easy way of expressing disagreement. If they have no such easy way, many will downvote instead.

When that happens, the quality control aspect of the voting system is severely impaired. All posts for positions that most LW readers disagree with would have a de facto handicap, since they would get many negative votes that they should not have gotten based on quality.

I somewhat suspect though that even with a disagree button, voting down will sometimes be used to express disagreement, just because its effect is more visible and of more consequence. In this respect, I think that a agree/disagree system will be effective to the extent that it has real noticeable effects (just as karma has a real effect in terms of granting privileges and conveying virtual status). If all it does is change a number attached to the comment, I think people will still succumb to abusing downvoting, albeit much less frequently than presently.

Quality of argument and correctness aren't necessarily linked. You might easily wish to encourage postings of similar quality and yet say that you don't agree with the position taken. At present, feedback has only one axis along which to vary, and it's not clear what that axis is supposed to represent.

What are the quality standards?

If python code improvements were free, I'd favor some easy-to-use feature that let voters optionally explain to the comment-author the reason for their up- or down-vote, anonymously. With the explanations visible to the author but not to the rest of the world (or not to the rest of the world unless the world clicked?), and so not cluttering the thread. Explaining the up-votes would let people spread goodwill, and point out specific strengths that the commenter could perhaps extend; explaining the down-votes would let people help new users learn to write useful comments, and would keep criticism less discouraging. (It is easier to respond well to "you should do such-and-such" than to blank rejection.)

But, yes, mandatory explanations would be terrible, as would a norm in which explanations were expected to accompany down-votes.

Agreed - even just functioning private messages would be a fix for this.

I completely agree with you on the question of self-upvotes.

In fact, there's yet another option: do away with the automatic self-upvote, so that users may actively vote on their own comments just like anyone else's (with the same impact on karma). This doesn't sound like a major change, but who knows -- the results may be surprising.

On the question of explanations, I'm less sure (cf. my reply to Yvain's comment that you quoted). If I put a lot of thought into a comment, and it gets downvoted, I'm going to be perplexed enough to want an explanation. In particular, I won't appreciate an implicit suggestion that I didn't put a lot of thought into the comment.

The problem with being able to self-upvote but not have it be automatic is that the people who have the lowest standards would vote their own comments up the most. This would effectively handicap people who hold themselves to higher standards.

I think for this reason that it should either be automatic or not exist at all.

Ah, but many of the comments made here clearly aren't intended to be profoundly important or even particularly useful.

It would be useful for people to be able to make inconsequential posts and then not vote them up. Similarly for the ability to emphasize particularly important points.

If instead we make it impossible to vote our own comments and posts either up or down, then the automatic upvote should absolutely be removed.

If I thought my own comment was downvote-worthy, I probably wouldn't have posted it

When downvoted you can hope for an explanation, and you can hate it when people don't give one, but forcing one?

I've offered three bright-line tests for when you can feel entitled to an explanation of what is wrong with your comment.

Notice how your use of the word comment, as though all comments and explanations are equal, strips out the quantitative aspect. I don't think that you can expect people to put more effort into explaining a down vote than the writer of the original comment put into writing it. If you spend five minutes writing a comment that contributes a tangle of confusions to the discussion you are not entitled to have a down-voter spend half an hour on a comment that untangles it all for you.

One the other hand, if you spend some extra time on your post, distinguishing subtle nuances of words, and tagging them, eg free1=gratis, free2=libre, and then make your your point with the tagged words, eg the GNU GPL focuses on free2 and free1 is collateral damage, then you have born much of the burden of untangling the ordinary, boring confusions. It is much less labour for a respondent to explain why he disagrees with you, and he should say.

If I thought my own comment was downvote-worthy, I probably wouldn't have posted it

Right, but that doesn't necessarily mean you should upvote it. Comments like the one above, while they add to the discussion, are not on par with comments that also make good, clear arguments, cite sources, and link to relevant resources.

ETA: Yes, I do realize the tension between this and the notion that I wouldn't post something if I didn't think it was upvote-worthy. If everybody agrees with me on this stuff I'm going to have to go back and un-upvote a lot of my own comments.

How did I miss this post entirely?

I'd been wondering that too... Anyway, no matter: I think there's interesting stuff in both. (Though I hope LW won't turn into a community whose main topic of discussion is the mechanics of running the community.)

Since I would not be able to upvote my comment, upvoting someone else's comment would suggest that I think their comment is better than my own

Huh? If you have no ability to upvote yourself, why would upvoting someone else's comment indicate it's better than yours?

Suppose that I don't have the ability to upvote my comment, but I have the ability to upvote someone else's. So even though I can't say that my comment is better-than-average, I can say someone else's is. Thus, if I think my comment is the best, I shouldn't vote anyone else's comments higher.

Okay, I was reaching a bit with that one. It could, after all, be argued that whatever concern I have in that situation, is identical to the situation where everyone can upvote their own comments and always does so.

However, in the situation where everyone doesn't do so (which would perhaps be more likely if karma was not tied to the auto-upvote), I have the option of re-evaluating my comment in light of other people's, and removing the upvote (or even downvoting!) if my opinion of my own comment changes. In such a situation, upvoting someone else's comment just means, "This comment is as good as mine".

If you are a well known commentator, I think it would help to know what share of the comments that you read you have been voting up and down. Of course, less knowledgeable readers should vote less often.

You guys are fussing over voting too much.

This is the perfect example of the kind of comment that should automatically get hammered to -10.

Think about the asymetry of effort. If the reply is simply "No we are not." then the site has already degenerated and is close to death. If the reply makes the simple point about the Eternal September and sites being killed by trolls, crap flooders, and general mediocrity, that is also a problem.

The person who is willing to contribute a substantial comment has been hi-jacked. Instead of making a interesting response to some-one elses substantial comment, he has spent his time on typing in stuff everybody knows. He will soon feel that his time on the site is poorly spent and will leave the site to die.

Whoops! If it is hidden, so are its children. I guess it has to be left visible, a ship on the beach being a lighthouse to the sea.

I'm not sure what you mean. Voting seems like it's one of the key features of this site, and it's still not entirely clear what it's for, how to use it, or whether it's doing its job. It seems to me that we've been fussing over voting just about the right amount.