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

Hacker News metrics (first rough approach)

mjg59.dreamwidth.org

Post reply on HN
Print view View original

61–70 of 72 posts

Re: Hacker News metrics (first rough approach)

#61

Post by tptacek »

pron wrote: post #60
tptacek wrote: post #54

Earlier quoted context omitted.

Disabling comments on sensitive stories has the effect of privileging the viewpoints of people who happen to win the front-page contest, and endorsing whatever framing happens to accompany whatever happens to be the most popular telling of that story. That seems like an extremely unfortunate dynamic to adopt. Please, HN, do not allow people to post stories on the site that can't be contextualized or even rebutted in…

I actually think this would encourage more thoughtful rebuttals in the form of long blog posts. One great thing about this community is that usually there are enough people to promote stories on opposite sides of the debate to the front page.

The assumption here is that blog post followups to popular front-page stories will themselves naturally find the top of the front page. But of course they won't. For one thing: simple causality guarantees that they won't be available until hours or (depending on the quality of the post) days after the original story first hits the front page, when it's at the zenith of its gravitational pull for HN votes.

Re: Hacker News metrics (first rough approach)

#62

Post by nknighthb »

dang wrote: post #44
nknighthb wrote: post #42

Earlier quoted context omitted.

You said: "The Hacker News ranking algorithm does not penalize stories that address social issues." And then explained the observed effects as being from flags and vote ring detection. People flag stories that address social issues, thus, the ranking algorithm is penalizing stories that address social issues. That (the collective) you did not specifically intend this makes a nice case study in unintended consequences…

By your logic, the software must also be biased against the things users downvote, since after all the code records the vote. Of course the algorithm penalizes (most) stories that users flag. But it has no knowledge of what those stories are about. If some topics get flagged more than others, that is a function not of the algorithm but of the inputs users give it. This seems so obvious that I fear this dispute is pur…

... Wow.

Does it help if we change the word "algorithm" to "system"?

Stories entering the HN system on certain subjects are more likely to be penalized by that system. Humans are an integral part of HN. HN would neither function nor exist without them. You cannot simply ignore them and pronounce the system unbiased.

Does that make more sense to you?

This is really surprising to me. I usually disagree with you on, well, everything, but you have not previously seemed like someone who suffers from this particular technologist disease of utterly missing the big picture because someone used a word that you define narrowly in a broader way.

Re: Hacker News metrics (first rough approach)

#63

Post by SamReidHughes »

nknighthb wrote: post #62
dang wrote: post #44

Earlier quoted context omitted.

By your logic, the software must also be biased against the things users downvote, since after all the code records the vote. Of course the algorithm penalizes (most) stories that users flag. But it has no knowledge of what those stories are about. If some topics get flagged more than others, that is a function not of the algorithm but of the inputs users give it. This seems so obvious that I fear this dispute is pur…

... Wow. Does it help if we change the word "algorithm" to "system"? Stories entering the HN system on certain subjects are more likely to be penalized by that system. Humans are an integral part of HN. HN would neither function nor exist without them. You cannot simply ignore them and pronounce the system unbiased. Does that make more sense to you? This is really surprising to me. I usually disagree with you on, wel…

Because clearly this dispute is caused by his inability to understand you, not by your inability to understand him.

Re: Hacker News metrics (first rough approach)

#64

Post by nknighthb »

SamReidHughes wrote: post #63
nknighthb wrote: post #62

Earlier quoted context omitted.

... Wow. Does it help if we change the word "algorithm" to "system"? Stories entering the HN system on certain subjects are more likely to be penalized by that system. Humans are an integral part of HN. HN would neither function nor exist without them. You cannot simply ignore them and pronounce the system unbiased. Does that make more sense to you? This is really surprising to me. I usually disagree with you on, wel…

Because clearly this dispute is caused by his inability to understand you, not by your inability to understand him.

I seriously have no idea what you're talking about. If you'd like to point out something specific you think I've misunderstood, I'm listening. If you just want to engage in content-free sarcasm, I'm uninterested.

Re: Hacker News metrics (first rough approach)

#65

Post by SamReidHughes »

nknighthb wrote: post #64
SamReidHughes wrote: post #63

Earlier quoted context omitted.

Because clearly this dispute is caused by his inability to understand you, not by your inability to understand him.

I seriously have no idea what you're talking about. If you'd like to point out something specific you think I've misunderstood, I'm listening. If you just want to engage in content-free sarcasm, I'm uninterested.

... Wow.

You're here calling @"dang" disingenuous when he has posted nothing but straightforward explanations of what inputs the HN ranking algorithm uses. Then you make a reply laced with the pretense that he doesn't get what you are saying, or are trying to say.

Really he's (I'm just inferring here, because it's soooo obvious) 100% clued in to your conversational gambits.

Re: Hacker News metrics (first rough approach)

#66

Post by dang »

AndrewDucker wrote: post #26
dang wrote: post #15

Earlier quoted context omitted.

> Do certain types of content set off the ring-voting detection more frequently? Sure—people posting their startups, for example, set off the ring detector more frequently than the typical news article. Friends voting for friends' blog posts, too. But is that a bias of the algorithm? That strikes me as a stretch.

Interesting. How do you guard against coincidental rings? (i.e. a subgroup of people who all love, say, Lisp, and therefore always vote up Lisp stories. They're not an organised group - but they do tend to vote in concert.)

I used to worry about that too, but it turns out not to be a big problem in practice. (I realize that's not a satisfying explanation, but this is one area where we can't give out details without enabling people to game the site. Having put a ton of effort into HN's anti-voting-ring measures, I dread the thought of having to climb all the way down the mountain and push that boulder up again.)

Re: Hacker News metrics (first rough approach)

#67

Post by dang »

JoshTriplett wrote: post #13
dang wrote: post #9

Earlier quoted context omitted.

According to HN's software, this story has been heavily ring-voted. That's what has affected its rank. Call that "penalized" if you want, but it has nothing to do with the content, so that tweet is misleading—and ironic, too, given what provoked the "penalizing". (Also, in case anyone's wondering, moderators haven't touched the post.)

While I realize you can't go into too much detail about ring-voting detection to prevent people from finding ways around it, can you elaborate a bit on what ring-voting means to HN, beyond the obvious? The obvious definition would be a clique of accounts that all upvote each other's stories, but that seems rather unlikely in this case; is there a secondary criteria? Are you looking to prohibit what Wikipedia calls "m…

I'm happy to elaborate on nearly any aspect of Hacker News, but this is one area (maybe the only area?) where I resist the impulse to explain. We're in something of an arms race against a lot of people who try to get their content on the front page through non-organic means. Accidentally helping them do so would be a genuine shot in the foot.

But let's say what good voting means to HN. It means voting for something that you personally find intellectually interesting, regardless of who posted it and regardless of what you or your friends are hoping to promote.

Re: Hacker News metrics (first rough approach)

#68

Post by aptwebapps »

nknighthb wrote: post #62
dang wrote: post #44

Earlier quoted context omitted.

By your logic, the software must also be biased against the things users downvote, since after all the code records the vote. Of course the algorithm penalizes (most) stories that users flag. But it has no knowledge of what those stories are about. If some topics get flagged more than others, that is a function not of the algorithm but of the inputs users give it. This seems so obvious that I fear this dispute is pur…

... Wow. Does it help if we change the word "algorithm" to "system"? Stories entering the HN system on certain subjects are more likely to be penalized by that system. Humans are an integral part of HN. HN would neither function nor exist without them. You cannot simply ignore them and pronounce the system unbiased. Does that make more sense to you? This is really surprising to me. I usually disagree with you on, wel…

Those two words are not interchangeable in this context. The algorithm is the code. The system, in your example, is the whole thing.

There's nothing wrong in making it clear that the code itself does not target social issues.

Re: Hacker News metrics (first rough approach)

#69

Post by nknighthb »

aptwebapps wrote: post #68
nknighthb wrote: post #62

Earlier quoted context omitted.

... Wow. Does it help if we change the word "algorithm" to "system"? Stories entering the HN system on certain subjects are more likely to be penalized by that system. Humans are an integral part of HN. HN would neither function nor exist without them. You cannot simply ignore them and pronounce the system unbiased. Does that make more sense to you? This is really surprising to me. I usually disagree with you on, wel…

Those two words are not interchangeable in this context. The algorithm is the code. The system, in your example, is the whole thing. There's nothing wrong in making it clear that the code itself does not target social issues.

> There's nothing wrong in making it clear that the code itself does not target social issues.

I agree, and I suppose if you choose to read my comment completely free of the context of the rest of the thread, my attitude might not make sense. But when you realize he had also said to 'somehow' that he "thought that was the main point of the article?", it's obvious he was missing the forest for the trees.

Re: Hacker News metrics (first rough approach)

#70

Post by nknighthb »

SamReidHughes wrote: post #65
nknighthb wrote: post #64

Earlier quoted context omitted.

I seriously have no idea what you're talking about. If you'd like to point out something specific you think I've misunderstood, I'm listening. If you just want to engage in content-free sarcasm, I'm uninterested.

... Wow. You're here calling @"dang" disingenuous when he has posted nothing but straightforward explanations of what inputs the HN ranking algorithm uses. Then you make a reply laced with the pretense that he doesn't get what you are saying, or are trying to say. Really he's (I'm just inferring here, because it's soooo obvious) 100% clued in to your conversational gambits.

Yeah, not interested in engaging with someone like you.
Post reply on HN