Ask HN: How are stories ranked on HN?
Hello, I'm a few weeks new here and learning the ropes. I'm curious to learn more about how stories are actually ranked on HN?
I understand per the FAQ that "The basic algorithm divides points by a power of the time since a story was submitted." Is there any further specifics about the basic algorithm that anyone knows and willing to share?
Moreso, the FAQ makes it clear that "moderator intervention" is a factor. Is there anything more known about this how this manual human aspect affects what we as a community will see ranked higher or lower?
Appreciate the opportunity to learn more about how stories are actually ranked on HN. Hopefully no-one takes any issue with asking or discussing transparency here. Thank you.
Edit to add: I see the moderators removed "actually" from the title, that's interesting.
12 comments
[ 3.9 ms ] story [ 44.7 ms ] threadYC has at least two vested interests in keeping it under wraps. 1) If everybody knew the algorithm, it'd get SEO-style abused even worse than it does already. 2) YC aren't in the business of running a tech discussion web forum, they're VCs aiming for 100x exits for their investments, and this site is one of the tools in their arsenal to improve that chances of that happening, so to expect "transparency" on an asset like this website is kinda insane.
Ultimately though, unless you're trying to game the rankings (in which case I couldn't care less how difficult you're finding it), you should just treat that FAQ answer as "close enough, barring details of fraud prevention", and get on with reading/voting/participating.
(If you're just being intellectually curious, there have been old copies of the arc source code around, you'll need to learn a specific dialect of Lisp to understand it, and the publicly available code explicitly leaves out most off the fraud prevention or any other YC-specific modifications this site runs...)
I've seen little evidence since joining in 2014 against the hypothesis that everything here is simply about maintaining quality and topical interest.
(Sorry, I can't even begin connect the dots about how some cool hack (say Pac Man that fits into a boot sector) being #4 on the front page generates ROI for some VC, but I had a good chuckle.)
I also don't think that blatant spammers would be easily able to manipulate the ranking even if they knew exactly how it is done. I have faith that HN would be on to their antics in a heartbeat.
Pac Man in a boot sector and similar stories keep you (and me) coming back here.
Both of these objectives are not targeted at investors, but are without doubt helpful in assisting YC backed companies find and hire "A players" at a higher rate that they'd be able to without this site, and that is almost certainly a significant factor in some/all of their 100x (or 20x or 3x) successes.
And if you don't think blackhat SEO-type spammers wouldn't destroy this site in a heartbeat if they had an accurate understanding of the algorithms and tools used to moderate it, perhaps you should cast your mind back to the days before SEOs and Google "ruined the internet". :sigh:
Did the question and context come off as though I felt owed something from someone?
I am intellectually curious. On that note...if the YC fork is proprietary, then how would learning a specific dialect of Lisp and running an old copy of the publicly available arc source code help answer the original two elements in question on the YC fork - namely algorithms and moderator influence?
Yeah, it did little. (Although perhaps less so now rereading the original post...)
"Transparency" is - at least in my mind - a word often used by entitled people demanding information about things they have no need to right to know about. Sorry if I misread your intentions there...
Of course you want the facts, and not some fictional story.
Most adverbs add nothing to a corpus. Before hastily putting down any adverb, if you stop and pause momentarily to think whether it is really necessary, your writing will totally improve.