Hacker News Reader

Comment by jwecker | Hacker News Reader

Here's what gets me- in the first couple of days there were some pretty good posts that got _tons_ of marks. There are some users in the top 20 rankings that haven't had any activity since the first couple of days- like http://news.ycombinator.com/user?id=beau . I have to be honest, while the customer service article was nice, I don't think it would get more than 20 points these days- I don't think any article would.

So I thought there must have been a lot of people using it at the very beginning- excited and contributing and the karma flowed like milk and honey and then the novelty wore off and a bunch of people dropped out. I see the opposite in the graph. I guess the usage of the site is what changed considerably.

View on HN

The reason those early articles got so many points is that the decay function was broken initially. Instead of accelerating downward off the frontpage with age, the velocity of stories' fall actually decreased. Because of this bug some of those stories stayed on the frontpage racking up votes for 4 days.

Looking at old comments[1], though, indicates that the overall volume of karma flow was much higher in the early days. Of course, it could be that the bug-driven hyperkarma on posts led users to be more generous with karma on comments.

[1] In particular, the feature request thread.

It might be interesting to try a decay function on the leader board.

Now, it's probably not cool to take away karma points, but it would reward activity. Plus, it'll get around the issue of single posters who post a single ridiculously popular post. They would have their hayday in karma but then slide back in the rankings in favor of the regulars.

The karma of active users would continue to rise whilst the dead users stay steady and eventually get overtaken.

Perhaps a 'most karma in the past week' leader board would work?

I like the idea; this mirrors some discussion in the "Thoughts on using news.YC" thread, viz. http://news.ycombinator.com/comments?id=2987

There are a lot of people who follow the links yet dont vote even if they like them. An example is the facebook YC news group. It has 42 members yet the story announcing it only has 13 points.

Could this be an example of the 1% rule? http://technology.guardian.co.uk/weekly/story/0,,1823959,00.html

I tend to think that it is more the result of a bit of cognitive dissonance, particularly with respect to what the points values mean. They can be viewed either as a metric of popularity, or a metric of quality. Most people {{handwaving}} conceive of points as some mixture of the two, hence there is a reluctance to give a point to a story with, say, 20 points, even if you liked it, because you don't think it is a "21-point quality" story.

Reply on news.ycombinator.com