Ask PG: have you seen a change in the cumulative comment karma? | Hacker News Reader

Ask PG: have you seen a change in the cumulative comment karma?

32 pointsjonmc1217 comments

I was curious, when you look at the statistics of total comment karma before and after the karma points were removed from comments, is there a big change? Are users upvoting / downvoting at the same frequency?

View on HN

Comments with high scores seem to have slightly higher scores than they would have, but comments with low scores seem to have about the same. Probably because the -4 limit on displayed comment scores was already concealing the actual score.

A helpful yet not quite profound comment of mine from last night is now my highest voted comment in 1.5 years on the site. Based on this single data point I'll attest to well-received comments scoring higher without visible scores.

People don't know if it's at +5 or +500 so they might be more likely to click the upvote on a comment that's already got plenty of karma.

Different people have different views of the purpose of karma, so having it not be visible will affect their voting in different ways. What you propose is one possible outcome, but another is that people stop upvoting at all. What's the point if there's no visible effect?

Please bring back the comment scores. It helps a lot in parsing the comments and assigning a proportional weight to each when reading them.

I had to think about this a bit, and I disagree so far. I'm finding that I'm not pre-judging comments as much. It's nice to be able to read someone's comment without knowing first that 70 or 80 or 3 other people thought it was worthwhile.

However, it also eliminate the ability for me to reference a comment with a high score with the understanding it's probably more correct. After all, once I read a comment, I want to know if what others think. Comment scores help that. Granted, it's not the only solution. Having some other visual indicator that suggests a comment is of high quality would be good without needing to resort to simply displaying points.

Replies to comments solve the "I want to know what others think" problem. I agree with the previous commenter: having the comment ratings hidden was unusual at first but I definitely do acknowledge that I used to mentally pre-judge a comment based purely on others' responses to it, despite having dealt multiple times with comments that I disagreed with which had higher upvote counts than the ones I felt were better answers.

I disagree. First, disconnect the actual score with something that indicates a high quality comment. You can display something as high quality without showing a score.

Next, a high quality comment need not have comments. Some of the best comments do not need followup. We should also discourage comments that merely parrot the parent and offer no content of their own. This is probably the biggest argument I have against not having any soft of indicator. If a comment is good, it deserves an up vote to indicate quality.

Next, replies still do not solve the scanning problem. For me to deduce the quality comments using replies requires I read through all the comments. Sometimes this is not an issue, but posts with a larger number of comments, this becomes problematic.

Finally, the above also means that replies of comments, and replies of replies, suddenly makes it really difficult what others deem quality content.

At the end of the day, up voting comments is meant to distinguish good, high quality comments from the normal run of the mill comments. If it's not being used for that, then up votes are merely being used as a way to obtain karma. In fact, I'd go so far as to say that by removing karma from showing which posts are of higher quality, it serves no real purpose other than to obtain more karma. I'm more interested in reading quality comments than obtaining karma, something the current system doesn't assist with. It only makes karma something to "brag" about.

or something else that helps rate a comments importance beyond the near-meaningless ordering. Should work within a thread and at the top level equally well.

This result suggests that a limit of +4 displayed points would be equivalent to the current situation, while leaving more visual guidance about how valuable a comment is perceived to be. If you chose to do that, maybe you should color-code it instead of displaying a number, though.

FWIW: I like the comments with hidden scores better than before.

So what? Does it really matter if comments have higher scores or not? The only thing they are good for now is getting karma to show off on a profile page? Why not just remove karma all together then?

The value of the points on comments before was to help distinguish higher quality comments. Comments that added value generally got a higher score, and we could browser for that.

That can no longer happen.

By all means do not display the points on comments, but do not remove the ability to scan for higher quality comments. You can use some other visual method or indicator to suggest that certain comments are higher quality.

Otherwise, the karma does nothing except provide an encouragement to post comments and links so you can get karma...? I just don't see the value.

My guesses:

Voting was used to set a comment to the "correct" level. Now no one can see where they're at, so good comments probably get more than they used to (depth).

Overall people don't think about points as much, so overall commenting is down (breadth).

Voting was used to set a comment to the "correct" level. Now no one can see where they're at, so good comments probably get more than they used to (depth).

That sums up my behavior. While I might up or down vote a comment simply based on its intrinsic value I tended to also consider "correcting" vote tallies. For example, if it looks like a comment is getting high up votes simply because of "me-too-ism" rather than because it is offering a striking observation I might down-vote (since I want to discourage generally trite but popular opinion comments).

The lack of feedback, the inability to see the results of my voting, and the inability to see how others are appraising comments detracts from the site and make me less inclined to vote at all.

I doubt that good commands get more upvoted than before. Visible, high points serve as a signal that the comment is worth a read. Also, I believe that other things being equal people tend to upvote high-point comments more. In other words, if you did an A/B test that would (resp.) add/subtract 20 points to the displayed points of an already well upvoted comment, people seeing the high-point version would upvote it more.

The more important question is: are populist comments getting fewer upvotes? The reason they might is that voting has become for many voters more the individual task of judging the worth of the comment and less the group task of deciding what is popular.

Looking at the distribution of comment karma before and after the change might give some clue as to that. Analysing two popular threads might be better, if we get the methodology right.

I'm finding that I miss the comment points, but I didn't think I would.

PG should do another experiment where he hides the submitter and commenter names while showing the comment points for N hours (N tbd) after the comment goes live.

The only thing I miss is an indicator that I actually did hit the upvote button and didn't accidentally downvote someone.

I use HN from an iPad about half the time, and I don't really trust my ability to accurately hit the correct arrow.

Reply on news.ycombinator.com