3 ms·
Well, this is one "newbie", but not all "newbies" are the same. I also think it depends on the time span during which a post gathers upovotes and comments. You
by valevk 12y ago
Well, this is one "newbie", but not all "newbies" are the same. I also think it depends on the time span during which a post gathers upovotes and comments. You can't derive the position of a post based on points, comments and actuality of a post.
I think some things changes since pg posted the algorithm [0], but I can't find anything. Also make sure to check Google for "Hackernews ranking algorithm". [1] [2] [3]
[0] https://news.ycombinator.com/item?id=231209 https://news.ycombinator.com/item?id=231209
[1] https://news.ycombinator.com/item?id=1781013 https://news.ycombinator.com/item?id=1781013
[2] http://www.righto.com/2013/11/how-hacker-news-ranking-really-works.html http://www.righto.com/2013/11/how-hacker-news-ranking-really...
[3] http://www.quora.com/Whats-Hacker-News-ranking-algorithm http://www.quora.com/Whats-Hacker-News-ranking-algorithm
- NicoJuicy 12y agoI do think that the post should be higher then the ones before it... Looking at the raw metadata, it's more relevant and more recent. The whole point of the popular section is seeing more popular and recent news that you could have missed out. The chance you missed out an article from 1 day ago is higher then one of 2 days ago. The chance you like it, is bigger if it has 90 votes then 50 votes. So there is no reason at all for putting it lower. PS. I recreated HN for a client (it included tags and custom input fields), so i do know what HN is in the core.. I do not know the algorithm specifics though, but neither does Google :)