Can HN make the leaders page more useful by showing average karma per comment? | Hacker News Reader

Can HN make the leaders page more useful by showing average karma per comment?

5 pointsmojuba11 comments

I think total karma says nothing but total time and effort spent on HN, which rather speaks negatively about the person. On the other hand, someone who joined later may turn out to be a more valuable contributor.

Two suggestions on making this calculation (average karma) easier and more or less meaningful: (1) I think it would be OK to just store monthly averages to avoid needing to go through all historical data. An average of monthly averages for, say, last 6 months can do the job pretty nicely and in any case is more interesting than total karma. (2) Showing up in the leaders chart can be triggered only after certain amount of comments made, say 50.

Should submission karma be included? I don't know. Honestly, in terms of merits your submission karma says very little about you, unless you submit your own articles, of course. But then, just be clever in the comments.

I'd expect the effect of this change to be even more interesting discussions and probably less uninteresting and unstimulating one-liner remarks here and there.

Calculating/displaying average karma discourages the type of behavior we want to cultivate on HN.

Users that care about their scores will only contribute to highly visible conversations in ways that are more likely than not to be up-voted. This will not only have the effect of squelching debate and the voicing of contrary opinions, but will also lead to an HN where only popular, big ticket posts are discussed while the rest are left to fallow.

SearchYC displays such data. For mojuba it is http://searchyc.com/user/mojuba

It displays avg points for both submission and comments

Also the leader board... http://top.searchyc.com/points_per_comment

Cool... there has to be a threshold though, some 50 or 100 comments minimum.

Or do something like k/(n+10)

Any chance to see the median instead of the average?

It's interesting to see spolsky at #3 with 14.88 on 87 comments and pg at #20 with 7.88 on 4721 comments. How do the underlying distributions compare?

"make the leaders page more useful" suggests that the leaders page has some utility.

What is that utility and why would increasing it have value?

> I'd expect the effect of this change to be even more interesting discussions

I'd be surprised if any change to the leaders page had any effect on discussions.

One easy experiment is to simply remove the leaders page and see whether there was any effect on the discussions.

This gets back to the earlier discussion of orange usernames for users with the highest average recent karma per comment,

http://news.ycombinator.com/item?id=467181

an idea that has since been rejected. But if the server load wouldn't be too bad, I wouldn't mind a display of the same figure on each user's profile page. I largely agree with your rationale for reporting this figure.

I also think submission karma should be included or separately reported, as that is how some users mostly contribute to the community, by making good submissions.

In terms of merits your submission karma says very little about you, unless you submit your own articles, of course

Similarly, popular comments say nothing about you other than that people agree with you. Short comments get more upvotes. Does that make short comments necessarily more valuable than longer ones?

Why have the leaderboard at all? What value does it provide, when karma says little other than activity and longevity on the part of the user?

Actually, this is a better implementation of 'orange' users. Have a page where users with the highest average comment rating are listed.

Why stop at average karma, we can also print a little histograms (sparklines) of comment karma, I bet the target audience of HN would love it :)

Another way for front-page ordering would be (total karma * exp(-beta * age)), you can tweak beta for time decay.

Reply on news.ycombinator.com