I tracked and analyzed 13,159 Hacker News posts. Here's what I learned

ankle.io

33 points·mellosouls··7 comments

I think this style of title is deliberately designed to be clickbait. I will not click on this.

I found the article to be bona fide data analysis, containing exactly what the title claimed.

Perhaps, but "I did something interesting. This is what I found." style, which is the two-sentence style revealing nothing, seems to be a psychological bait.

The problem is that title fail to summarize the content of the article, and tries to play on the readers curiousity to get them to click on it. Any self serious source of information, like Encyclopedia Britannica wouldn't title an article that way.

My main takeaway is that the HN system mostly works. You get some minor improvements by choosing the time well.

If I wanted to game HN, I see 2 suggestions:

* You can create upvotes from fake accounts. I'd assume the moderators to notice this very quickly.

* You can find a group of people with 10k karma and ask them how to optimize your content to be more HN friendly a.k.a. interesting. Then everybody wins, so I suppose I can live with it.

This was an interesting analysis, however I think for histograms with not a lot of bins it's better to use kernel density estimation[0] to better visualize the spread and not have to control the bin size.

[0] https://en.wikipedia.org/wiki/Kernel_density_estimation

Interestingly OP didn't post at the best time nor day to reach the front page, although not at the worst either.

Reply on news.ycombinator.com