Ask HN: How is the average karma calculated? | Hacker News Reader

Ask HN: How is the average karma calculated?

5 pointsmrb8 comments

How does HN calculate the average karma for a given user?

I noticed my average karma changes relatively quickly. Over 2-3 months it varied between approximately 2.0 and 4.0. Given that my account is 4 years old, it seems that HN takes a sliding window of the last X posts, or perhaps posts made in the last X days, to calculate the average karma.

Also it seems HN does not take into account points from submitted articles. 2 of my last 15 submissions over the last 3 months received hundreds of points, but this did not significantly increase my average karma.

I also noticed the average karma seems to be recalculated infrequently (no more than once a week).

Out of curiosity I downloaded my posts over the last 2 years, and tried running a sliding average karma calculation, but nothing matches what HN is reporting.

HN staff: care to enlighten us?

View on HN

It's based on comment scores only. My sense is that it's an irrelevant distraction and we should get rid of it.

For example, we caught a bunch of users who were gaming it by deleting any comments that brought their average down. We fixed that by treating deleted comments as comments of score 0 for average-computing purposes. It's an example of how, once you publish a metric, people start to care about it and do things based on it, regardless of how meaningful it is.

Comment average used to be used by a few algorithms (like comment ranking) but we turned that off as an experiment a while ago and nothing seemed to get worse. If anything, I think it may have helped a little.

Even if it's not being used algorithmically, it may still be useful as a form of gamification. I mean, people deleting their poorer comments is probably a good thing - particularly when it doesn't effect the placement of their comments on the page.

Developing a willingness to delete my less productive comments was a big step toward improving my skill at self editing. Since average is not being used algorithmically, it's hard to see an advantage in penalizing self-editing.

Learning self-editing is great, of course, and you could still do that without a published comment average.

I think optimizing comments for upvotes is a bad idea. You should optimize for saying substantive things.

Sure I agree.

My sense is that Karma may correlate with what is substantial and awareness of its trends may correlate with awareness of one's own patterns. That awareness can produce, as you observed, heightened self editing.

Now perhaps we disagree that more self editing is or is not a good thing or whether it outweighs the negative effects of deleted comments. Personally, I think more comments could stand self deletion and have observed few instances where a deleted comment has had much impact on an important discussion.

I totally agree. But what you say does not jive with what the system encourages.

Get rid of it!

The coup de grace would be to figure out how to do away with karma altogether. If you will it, it is no dream.

When I look at my profile, there's the following text. I think it's actually from an HN extension I use and not HN itself but I don't remember for sure:

Average karma is calculated by averaging the scores of your last 50 comments except the comment with the highest score and your 5 most recent comments. It is generally recalculated every few days.

Average karma is based off the average karma of your past n submissions and comments. It is not an average of all submissions.

It also doesn't mean anything.

Reply on news.ycombinator.com