Ask HN: Algorithm arranging posts on HN?
I have noticed something on hn, that the posts on the main page gets arranged in a peculiar way. For example, a link posted 4 hours ago with less than 20 upvotes might be on the main page still, while another similarly aged link with twice as much upvotes might be on the second page.<p>I am wondering what algorithm is used in ordering them. Any idea?
4 comments
There are several small tweaks that I've noticed. There's something about the length of a post, possibly something about the number of comments, and possibly others. There are also penalties for various things. Items without a link are ranked more lowly than those with a link, and some topics du jour have a penalty applied.
All in all, it's actively curated by PG, with small changes to keep the system in balance and running smoothly. Enquiring too closely is discouraged.
It's also plausible that this item will get deleted. After all, this is a site for hackers, not a site about Hacker News.
All in all, it's actively curated by PG, with small changes to keep the system in balance and running smoothly. Enquiring too closely is discouraged.
It's also plausible that this item will get deleted. After all, this is a site for hackers, not a site about Hacker News.
Agreed!
I understand that there's always more chance that if its fully disclosed, it might be used to game the system.. but could stop my curiousity, hence the question :).
I understand that there's always more chance that if its fully disclosed, it might be used to game the system.. but could stop my curiousity, hence the question :).
http://news.ycombinator.com/item?id=1781013
quoting from PG's comment on that post:
(= gravity* 1.8 timebase* 120 front-threshold* 1 nourl-factor* .4 lightweight-factor* .17 gag-factor* .1)
or simplifying a bit:
Score = (P-1) / (T+2)^G
where, P = points of an item (and -1 is to negate submitters vote) T = time since submission (in hours) G = Gravity, defaults to 1.8 in news.arc
quoting from PG's comment on that post:
(= gravity* 1.8 timebase* 120 front-threshold* 1 nourl-factor* .4 lightweight-factor* .17 gag-factor* .1)
(def frontpage-rank (s (o scorefn realscore) (o gravity gravity*))
(* (/ (let base (- (scorefn s) 1)
(if (> base 0) (expt base .8) base))
(expt (/ (+ (item-age s) timebase*) 60) gravity))
(if (no (in s!type 'story 'poll)) .8
(blank s!url) nourl-factor*
(mem 'bury s!keys) .001
(* (contro-factor s)
(if (mem 'gag s!keys)
gag-factor*
(lightweight s)
lightweight-factor*
1)))))or simplifying a bit:
Score = (P-1) / (T+2)^G
where, P = points of an item (and -1 is to negate submitters vote) T = time since submission (in hours) G = Gravity, defaults to 1.8 in news.arc
I noticed that this morning too:
http://news.ycombinator.com/item?id=1966201
http://news.ycombinator.com/item?id=1966201
One word: Yes. (Duh!)
FAQ: http://ycombinator.com/newsfaq.html
FAQ: http://ycombinator.com/newsfaq.html
umm, the faq on this doesn't really explain the scenario...
the question was something like this... how come a 20 upvote link gets displayed higher than a 40 upvote link when both are carrying similar timestamps?
the question was something like this... how come a 20 upvote link gets displayed higher than a 40 upvote link when both are carrying similar timestamps?
Yes. Vote/Time Elapsed
You're stating the obvious without really replying to bhavin.