3 ms·

In this case, the algorithm is open, so there's no guesswork: http://www.righto.com/2009/06/how-does-newsyc-ranking-work.html http://www.righto.com/2009/06/how

by Natsu 10y ago

In this case, the algorithm is open, so there's no guesswork: http://www.righto.com/2009/06/how-does-newsyc-ranking-work.html http://www.righto.com/2009/06/how-does-newsyc-ranking-work.h... HN has normally been transparent whenever they modify something manually.

cLeEOGPw 10y ago
So it's exactly that. > An item with too many comments (at least 20 and more comments than upvotes) gets the contro-factor penalty. What actually happens is an organized group of shills comment on an article and it gets burred. Also what it does is pushes all the interesting articles that facilitate discussion to the bottom. That's my another complaint about HN - there are barely any interesting stories on the first page even. And it doesn't look like there's a correlation between how interesting is the story and how high it is shown. The idea of time factor is cool in theory, but it fails in practice when most people visit once a day or only once every few days, and instead of getting the most interesting submissions they get a bunch of newly submitted links with barely any traction.
douche 10y ago
I've been noticing a pattern lately where anything that touches on a few different hot-button issues gets blasted away with flags the second it creeps onto the front page. I don't think its the mods having anything to do with it, but it feels a little broken.
- dang 10y ago
  It's usually that users flag these stories, presumably because of HN's (really the public internet's) inability to discuss them civilly and substantively. It's a tough problem. Those stories often relate to important issues which hold intellectual interest, if not necessarily gratifying intellectual curiosity (the mandate of this site). In that sense they're not all off-topic. But it damages this site and community for it to host ideological flamewars, much as it damages a neighborhood to host rubbish fires. My reading of the user flags is that they're partly trying to protect the neighborhood, and there's value in that.
  - Natsu 10y ago
    It's a difficult balance to be sure, and I have to say that I respect your work in keeping HN a pleasant place to be. I mean, there's stuff like this which is pretty important: http://imgur.com/gallery/ebUd5 http://imgur.com/gallery/ebUd5 But... if you look at that, you can see 'flamewar' written all over it, even though it's getting very little coverage. So I wouldn't submit it to HN as a story, even though probably quite a few people here care about it.
  - cLeEOGPw 10y ago
    So what it achieves is silences any discussion and pretends that no controversial topics exist. I don't think there's much value in this HN's "fear of controversy". Fact is, almost ALL things that are "gratifying intellectual curiosity" have some controversy in it, so the system doesn't deliver what it's made to do. After this too aggressive filter on topics, what's left is some very niche topics without controversy, since they are too niche for the majority to have any opinion or prior knowledge on it, or so uninteresting nobody even bothers to comment. Just open HN at any given moment and tell me honestly - how many articles do you actually read and think they are genuinely interesting? I can tell you this - a couple years back, when HN was just starting to get popular, the algorithm was different, and I found almost 90% of top stories so interesting it was worth to read them, and the remaining 10% I usually read anyway because it was something I just didn't knew about yet, but could be safely said they were interesting to people who knows what it's about based on the way stories were ranked. Now I open NH and AT BEST I read 1-2 articles, often 0, because it is full of either some obscure stuff, often without even upvotes, but new, or it is something everyone agree on, which is often an article about common sense. Sad.
    - dang 10y ago
      I'm open to hearing about ways to make HN more interesting, but it's hard to have this discussion on the terms that you're presenting. For one thing, people disagree strongly about what's interesting. For another, this perception tends to be skewed by nostalgia: most of us think that HN is worse than it was when we first got into it, but that's because we mostly think that about everything. (Things have always been getting worse, as some wit once said.) There are other biases too: if HN doesn't interest you (i.e. any of us) for a few days in a row, you're likely to think that it's going to seed, but in fact regular fluctuation is bound to yield such stretches from time to time. If you'd care to find an example or two of a front page from when you think HN was better, that expresses what was better about it, I'd be happy to take a look. You can do that either using the Internet Archive, or by using HN's front page archive like this: https://news.ycombinator.com/front?day=2016-07-23 https://news.ycombinator.com/front?day=2016-07-23, for any day since 2014-11-11. For fun and a small bias check, if you do look at frontpage history, keep track of the number of days you looked at before finding suitable examples. e.g. if you come up with two good examples, it would be interesting to know if they were the first two you looked at, vs. 2/10 or whatever. Btw, the above offer goes to any HN reader who wants to spend time finding examples of how HN was better in the past. Just make sure to give us specific links. Otherwise it's too hard to say anything meaningful.

by Natsu 10y ago

cLeEOGPw 10y ago

So it's exactly that. > An item with too many comments (at least 20 and more comments than upvotes) gets the contro-factor penalty. What actually happens is an organized group of shills comment on an article and it gets burred. Also what it does is pushes all the interesting articles that facilitate discussion to the bottom. That's my another complaint about HN - there are barely any interesting stories on the first page even. And it doesn't look like there's a correlation between how interesting is the story and how high it is shown. The idea of time factor is cool in theory, but it fails in practice when most people visit once a day or only once every few days, and instead of getting the most interesting submissions they get a bunch of newly submitted links with barely any traction.

douche 10y ago

I've been noticing a pattern lately where anything that touches on a few different hot-button issues gets blasted away with flags the second it creeps onto the front page. I don't think its the mods having anything to do with it, but it feels a little broken.

dang 10y ago
It's usually that users flag these stories, presumably because of HN's (really the public internet's) inability to discuss them civilly and substantively. It's a tough problem. Those stories often relate to important issues which hold intellectual interest, if not necessarily gratifying intellectual curiosity (the mandate of this site). In that sense they're not all off-topic. But it damages this site and community for it to host ideological flamewars, much as it damages a neighborhood to host rubbish fires. My reading of the user flags is that they're partly trying to protect the neighborhood, and there's value in that.
- Natsu 10y ago
  It's a difficult balance to be sure, and I have to say that I respect your work in keeping HN a pleasant place to be. I mean, there's stuff like this which is pretty important: http://imgur.com/gallery/ebUd5 http://imgur.com/gallery/ebUd5 But... if you look at that, you can see 'flamewar' written all over it, even though it's getting very little coverage. So I wouldn't submit it to HN as a story, even though probably quite a few people here care about it.
- cLeEOGPw 10y ago
  So what it achieves is silences any discussion and pretends that no controversial topics exist. I don't think there's much value in this HN's "fear of controversy". Fact is, almost ALL things that are "gratifying intellectual curiosity" have some controversy in it, so the system doesn't deliver what it's made to do. After this too aggressive filter on topics, what's left is some very niche topics without controversy, since they are too niche for the majority to have any opinion or prior knowledge on it, or so uninteresting nobody even bothers to comment. Just open HN at any given moment and tell me honestly - how many articles do you actually read and think they are genuinely interesting? I can tell you this - a couple years back, when HN was just starting to get popular, the algorithm was different, and I found almost 90% of top stories so interesting it was worth to read them, and the remaining 10% I usually read anyway because it was something I just didn't knew about yet, but could be safely said they were interesting to people who knows what it's about based on the way stories were ranked. Now I open NH and AT BEST I read 1-2 articles, often 0, because it is full of either some obscure stuff, often without even upvotes, but new, or it is something everyone agree on, which is often an article about common sense. Sad.
  - dang 10y ago
    I'm open to hearing about ways to make HN more interesting, but it's hard to have this discussion on the terms that you're presenting. For one thing, people disagree strongly about what's interesting. For another, this perception tends to be skewed by nostalgia: most of us think that HN is worse than it was when we first got into it, but that's because we mostly think that about everything. (Things have always been getting worse, as some wit once said.) There are other biases too: if HN doesn't interest you (i.e. any of us) for a few days in a row, you're likely to think that it's going to seed, but in fact regular fluctuation is bound to yield such stretches from time to time. If you'd care to find an example or two of a front page from when you think HN was better, that expresses what was better about it, I'd be happy to take a look. You can do that either using the Internet Archive, or by using HN's front page archive like this: https://news.ycombinator.com/front?day=2016-07-23 https://news.ycombinator.com/front?day=2016-07-23, for any day since 2014-11-11. For fun and a small bias check, if you do look at frontpage history, keep track of the number of days you looked at before finding suitable examples. e.g. if you come up with two good examples, it would be interesting to know if they were the first two you looked at, vs. 2/10 or whatever. Btw, the above offer goes to any HN reader who wants to spend time finding examples of how HN was better in the past. Just make sure to give us specific links. Otherwise it's too hard to say anything meaningful.