Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Stories that Hacker News removes from the front page

sangaline.com

Post reply on HN
Print view View original

101–110 of 327 posts

Re: Stories that Hacker News removes from the front page

#101

Post by foob »

johnlbevan2 wrote: post #74

Other theories: - This is a bug / happens randomly; you just noticed it because you were looking (i.e. as you analyse this data); all the posts it's happened to before and since went unnoticed. That's supported by the evidence of your analysis; most of the results don't look any different to other posts. - It's not the link, but the related activity. Presumably if you're running analysis on HN data, there are a lot o…

Good points! I use the official Hacker News API [1] and my requests come from an ip address that is completely disconnected from my personal account. Even if the API usage were a red flag, there would be no way to automatically connect it to me.

You're definitely right about there being miscellanious rules in there. Something that I mentioned in passing in the article is that many stories exhibit a significant drop in position once they're 15 hours old. If you look closely at the typical story trajectories you can also see various other jumps of about 10-30 positions which I would guess are triggered by these various rules.

The stories listed in the article exhibit very different behavior where they jump hundreds of positions instantaneously. It's absolutely possible that this is triggered by some automatic mechanism but if that's the case then there's an enormous amount of signifance being assigned to the corresponding rules. If there's some random component to the ranking then I highly doubt that it would be responsible for jumps of this magnitude.

I try to emphasize in the article that I do think it's possible that there's a hidden flagging threshold that's responsible and that the data can't tell us with certainty whether or not that's the case. I just personally find it unlikely that that's what happened for all of these stories. If you ran a site like Hacker News then would you put an admin link next to each post that pushes it off of the front page? I know that I would.

[1] - https://github.com/HackerNews/API

Re: Stories that Hacker News removes from the front page

#103

Post by Jerry2 »

Interesting. Another aspect of HN that I've noticed are stories that are not censored but actually promoted. Some time ago, I've noticed that some stories were few hours old and had 4 votes yet they were in top 15 position on HN. I never understood how that's possible without some kind of manipulation.

I've collected some of these anomalies. Peruse them and analyze them in this album:

https://imgur.com/a/6OvnE

Maybe OP can find a pattern in these.

Re: Stories that Hacker News removes from the front page

#104

Post by foob »

azrazalea wrote: post #83
ploggingdev wrote: post #71

Earlier quoted context omitted.

OP uses this story as an example of having "moderator fingerprints on them" i.e moderators censoring posts, which is not true. What happened was the opposite : the moderator uncensored a post flagged by the community.

Well, the author tried to account for that by saying that normal flagged posts don't go down near as fast. If what is being said in other threads is true, high karma accounts clicking flag could push a post off the front page without showing the "flagged" tag, then that could be an explanation instead. Which would then mean that maybe these high karma users have too much power.

That's an interesting point. I hadn't even considered the possibility that a single user would have the power to drop a story by hundreds of positions. If that's the case then those users do have a lot of power.

Re: Stories that Hacker News removes from the front page

#105

Post by djsumdog »

laurent123456 wrote: post #12

As much as I like HN, I'm not a big fan of the secrecy around moderator interventions - what gets censored, what posts get re-titled, etc. I can understand they might want to keep the ranking algorithm and anti-spam techniques secret, but stuff that are manually censored by a moderator should be indicated as such, maybe by some automatic message like "This post was removed due to [reason]". Some websites manage to fi…

I remember a few weeks back I was looking for a comment reply to my comment. The original post got flagged and the comment itself didn't appear in my comments block!

I was really disappointed as the comment I was looking for wouldn't even show up in search.

Re: Stories that Hacker News removes from the front page

#106

Post by ethbro »

brudgers wrote: post #96
Mithaldu wrote: post #85

Earlier quoted context omitted.

I have show dead on since a long time. The things i mentioned have no UI to note them. Show dead only makes, well, "dead" posts render their text.

There's also the |contact| link to ask the moderators directly about anything that looks nefarious.

I think the perspective some people are taking here is that if it's invisible in the UI, then how would they ever know to ask? Short of running a continuous scraper and then compiling statistics as article did.

I'd be all for tagging a generic "Downvoted by moderation" onto any article this happened to. The poster will probably still be annoyed, but I imagine most would agree with moderation most of the time. And if they like, they can contact.

What they're less likely to agree with is this happening invisibly. And if it's invisible, then how do we know whether we (as a community) agree with it or not?

PS: Side note, strongly against "toggle a non-default option" as a solution. Filter bubbles are incompatible with democracy and true, diverse community.

Re: Stories that Hacker News removes from the front page

#108

Post by leephillips »

ballenf wrote: post #93
LeifCarrotson wrote: post #37

Earlier quoted context omitted.

> Security through obscurity isn't a thing, and the better response is just to make a better algorithm, not try and suppress knowledge about it. Obscuring an algorithm or making it more tedious to reverse may not make it perfectly secure, but that's not the goal. It's not like actual information security, where loss of the encryption keys means your product is broken or your database is on the Internet. You're just t…

Once it's public (even unofficially), people will begin to rely on the current implementation and then complain if it is ever tweaked or changed. They build towards the value it grants and then complain that the value was arbitrarily stolen from them.

That's for sure. I rarely visit SEO websites or forums, but the few times I did I saw lots of complaints from people who seemed to think that Google owed them something after tweaking their page ranking algorithms.

Re: Stories that Hacker News removes from the front page

#109

Post by belorn »

moomin wrote: post #13

It's very interesting to note the number one pulled story: https://news.ycombinator.com/item?id=13741276 . If you take a look at the comments, it's theorized there that the story got pulled not because of moderator action, but because people abused the flagging mechanism. Given the content, and given the principal person under discussion, this seems pretty likely to me.

Why would it be abuse if people flag it as being off-topic for HN? The article is about some community code of conduct text which the article author comments on by adding their own interpretations to it. It look political and it uses inflammatory language.

The tug of war between upvotes and flags, as commented by dang, seems a strong sign that the article is very much political in nature. The question is if it also gratifies one's intellectual curiosity, but personally, in this case it did not do that for me.

Re: Stories that Hacker News removes from the front page

#110

Post by foob »

Sean1708 wrote: post #76

I don't think I understand your figures correctly, because none of them (except the last) seem to show any significant drop. There's a drop of about 10 positions in one of them but I wouldn't call that particularly significant, and it even climbs back up after twenty minutes or so.

Sorry if I didn't make this clear enough in the article but all of the stories in "Suppressed Story Trajectories" either drop off by hundreds of positions or disappear completely after their plotted line ends. They all match the criteria laid out in the previous paragraph. I used a larger y-scale on the last figure to show the full trajectory because that was a special case where the story didn't disappear completely and the trajectory after the jump was relevant.
Post reply on HN