Tell HN: My experience with karma average staring me in the eye . . . | Hacker News Reader

Tell HN: My experience with karma average staring me in the eye . . .

13 pointstossedout99915 comments

The inclusion of the karma box in the top right has me caring about my per-post average for the first time.

I am more gun shy about posting comments that might get a below the average response.

I find myself hunting to comment on "fresher" stories which almost always = more karma.

I know I know solution is "stop caring" but I feel somehow more watched or more accountable.

Looking forward to PG sharing results of the experiment.

And hearing other folks comment about positive / negative.

View on HN

Would it help to suggest that, on average, half of your posts should get a below average response? Most of them, actually, if you take into account how much the average can be skewed by a handful of high-scoring posts.

If you're being disinclined from posting one-word answers and empty "me too" responses: cool. No one would miss those. But if you're stopping yourself from posting interesting or helpful responses to old or unpopular threads? That's when you should stop caring about "karma" the poorly-quantified score and worry more about "karma" the concept. Being directly helpful to a dozen individuals in otherwise ignored threads would usually give a lower karma average than simply posting a clickable version of a link, but you'd get a lot more real respect from it.

Also keep in mind that anyone that sees your average karma will see your total right next to it, and if you game either one too much at the expense of the other it could give a worse overall impression than posting normally. I'd rather see a 1000-total 3-average poster than a 100-total 8-average or a 5000-total 1.1-average.

What you say about averages is very very true.

Linus Pauling once said "If you want to have good ideas you must have many ideas. Most of them will be wrong, and what you have to learn is which ones to throw away."

which can be extended to:

"If you want to have good HN comments you must have many HN comments. Half of them will always be below average, and what you have to learn is which ones never to post."

I'd say the single most significant thing pg could do for the karma system is to weight votes by the total number of votes in the story.

Right now it is way easier to get karma in animate, controversial topics rather then in insightful, complex discussions.

I think I understand why he hesitates: weighting removes the simplicity of one vote - one point. But if karma is supposed to overcome blind popularity, I think it's necessary.

Easy karma fraud.

Not if weighting is below 1.0, i.e. threads with lots of comments get less karma, but sparse threads get close to 1.0.

Something like

  karma per vote = #comments / #votes

Okay, actually that sounds reasonable.

Hah wait that's not actually what you said, which still doesn't sound reasonable.

Also a better formula might be something closer to

  karma per vote = thread karma / total comment karma in thread

Oh, and of course the average would have to be weighted by thread karma.

  average = total user karma / sum of thread karma over all threads commented in

The "gun shyness" aspect is unfortunate.

At first glance, it makes it sound like "grumble grumble good, now HN comments will be less like Reddit grumble grumble". But, in reality, people will be discouraged from participating in less popular discussion threads that either aren't "hot" or fill an interests niche, of which there's been many.

Or at worst, it will encourage my friends to mindlessly upvote my comment in order to help me keep my karma average going.

Aside from that, I appreciate pg's experiments, or rather, the data results of his experiments!

By comparison, I've barely noticed it. In fact, I'm actually commenting more than I used to, instead of just writing something and then closing the tab because I don't feel that it will add to the discussion. I just say things when I feel like I have something to add, while filtering out useless comments, which is exactly the behavior which I think should be encouraged.

My average karma is very low, I've noticed that I rarely hesitate to post something, regardless of the thread size. I think this is because I benefit sometimes from other late-thread low-karma posts and want to reciprocate.

Much has already been said about the current state of HN and karma, but one thing I'd like to add is this: comments that provoke discussion (i.e. comments that are responded to) should perhaps be valued higher than ones that do not. Excluding those that are downvoted and responded to because of that, obviously.

I am more gun shy about posting comments that might get a below the average response.

This is what makes this a dangerous road to go down.

My Average Karma hasn't changed? Does the algo only calculate new averages after a certain threshold?

I believe that the average is computed of all HN posts you have made. That means that it becomes harder and harder to change the average response, particularly if you have posted a lot. Your performance is locked in by your history and the law of large numbers.

It might make sense to have a time window on karma and have the average reflect the last 30 days or the last 100 posts or something of that ilk.

I had the exact same experience when I realized you got to down vote when you had "enough karma". That is, I found that just a total karma had the same effect. I've only just noticed the average karma, so call me slow.

Reply on news.ycombinator.com