Top Stories On HN vs Reddit | Hacker News Reader

Top Stories On HN vs Reddit

thintz.com

20 pointsthecombjelly13 comments

"I took two numbers that depend on the size of a certain population, let's call it X. Those numbers were X1 and X2, and were, predictably, of the same order of magnitude, since they're just different measurements of the same thing (the upvote activity of the entrepreneurial community on social news sites).

Then, I took the first number and multiplied it by 10. To my surprise, after this operation, the first number was about 10 times bigger than the second number!"

Not to mention the essayist seems to assume that all Reddit traffic is allocated to the programming section.

A reddit/HN-like voting system is inherently unstable, as some users just skim over top stories and upvote them, creating a positive feedback loop. If you have few submissions, this effect is weak. As the number of submissions, users and votes all grow linearly with respect to each other, the number of top stories and the number of stories an average user can read stay the same. This means the stream of new stories gets less and less readable, while top stories stay readable, which makes larger and larger percentage of users just skim over the top stories, and quality suffers. It would be interesting to create an accurate mathematical model of this behavior, maybe this would let us significantly improve stability.

A better way to compare might be to look at the top n stories over the past m days on the two sites, and see how much overlap there is. You can see the top recent stories on HN at http://news.ycombinator.com/best.

Or you could just take snapshots of the two frontpages every 12 hours or so for a couple days.

I applaud your effort in looking into this. However, I believe that when people say HN is turning into Reddit they don't mean that the same stories are being submitted, they just mean the same quality of stories.

http://www.paulgraham.com/hackernews.html

PG says that the comment quality is the most worrying part of the site ("Bad comments seem to be a harder problem than bad submissions. ")

Perhaps a better way of measuring the difference, would be to parse the comment threads and compare length (PG: length of a comment is directly proportional to the intelligence/civility of the poster) or frequency of memes.

I suspect this would show HN to consistently score higher, but not as dramatically as your analysis of posts.

I think the comment quality is quite high here given that it's such a dynamic site. I'm hard pressed to think of another site with such story traffic that matches the comment quality.

It seems the solution to that isn't really technological, it's the community peer pressure the site creates. It's cultural, in terms of the culture of the site.

One thing that could be done would be to go back to a -8 points limit for down votes which may reinforce better what are really bad comments. At the moment a comment with an unpopular stance that is decently written can go -4 along with plain stupid comments that would normally get down voted a lot more.

Alternatively; a two value system measuring popularity of the sentiment expressed and separately the quality of the post. So for example, a poorly written post with good ideas would be: legibility: -4, popularity: +4. This could be expressed as an rounded average or different users could filter by their own preference of legibility/popularity.

Of course this would all be a fairly major change to a system that isn't really broken so it's just speculative thinking on my part.

Reddit-like submissions don't bother me as much as Reddit-like comments. I used to love reading the comments on HN. In the past I was also much more likely to take part in the discussion. These days though I just skim them, and I hardly comment. In the past, there weren't nearly as many "me too"-type comments (etc), and those that were posted got instantly down-voted. The turning point came when people started upvoting complaints that noone was interested in superficial anecdotal opinions that added little value. That was the signal to me that HN's "mindset" was beginning to shift to a Reddit-like state. Not that there's anything wrong with Reddit, but the Web doesn't need two of them.

The dip in quality doesn't surprise me though. A year or two ago I said the same thing about Reddit in comparison to Digg.

What your data shows is that Reddit and HN do share some of the same popular stories which would lend credence to peoples complaints that it is turning into reddit.

Also your upvote statistic doesn't seem particularly meaningful.

reddit's frontpage sucks

reddit/r/programming half-sucks

reddit/r/coding is like HN, but nobody goes there

so, what's the conclusion? really?

Reply on news.ycombinator.com