9 ms·

When I worked with the HN post data, I noticed that some years ago HN users had correctly predicted the "Show HN" projects that later got funding. Those project

When I worked with the HN post data, I noticed that some years ago HN users had correctly predicted the "Show HN" projects that later got funding. Those projects had more upvotes. The more recent data has no such connection. It seems that the influx of users reduced the quality of judgement. So one way to improve HN submissions and comments is to weight points by the user's tenure on HN. I also suspect that early comments dominate late comments by the time factor alone. The sorting algo gives a brief advantage to new comments, but old comments are more visible. A post on the front page gets 30+ comments in the first hour, and latecomers can only post into the void. To address that, long branches could be collapsed by default, leaving only 2-4 visible messages per branch.

berntb 10y ago
Cool. And sad. But re old accounts: I've recently seen a few cases where old accounts with high karma dismissed Wikipedia -- without any references of their own. That is a "quality smell", if anything is. (One particularly shameless fellow just copied the primary link Wikipedia used -- and lied about the content. :-( ) EDIT: I wrote without any references of their own -- i.e. dismissed with hand waving. Wikipedia won't always be correct, but it has references and is whetted by quite a lot of people.
- 50CNT 10y ago
  My dad likes to dismiss Wikipedia, and he used to be an admin. They may just know how the sausage is made. His rule of thumb is "If it isn't a major article there's probably errors". His favorite example is that of an article about an ancient persian poet, which says that there's only 2 of his poems known, and only 1 remains. Next to that line, there's a picture of the other poem. Eh, who knows.
  - berntb 10y ago
    I didn't argue that Wikipedia always was correct (no source is that, I'd guess), just that it smelled really bad to dismiss it with just hand waving. It do have sources and have been whetted (for reasonably popular subjects.)
    - 50CNT 10y ago
      Wasn't trying to imply you meant that. It's more that I could see where someone like that could come from, dismissing the Wikipedia article out of habit, especially if they've been involved with the project. Not backing it up with additional sources is of course bad form, but finding and linking extra sources is additional work. Minor nitpick, vetting is someone checking something and approving it, whereas whetting is grinding away metal to sharpen a blade.
      - berntb 10y ago
        OK, you more or less agree with my argument. Enough about that, then. I am curious about something else. Do you have non anecdotal information about Wikipedia's general dependability? How is the dependability for large subjects? For subjects like a poet where less than a hundred people even recognize the name? (The last I saw on this was years ago, when Wikipedia edged out the Encyclopædia Britannica. It would be strange to be less dependable now?!)
      - berntb 10y ago
        It seems I was trolled, but in a funny way. I discuss how you can see bad behaviour from even old accounts at HN, dismissing serious references with just hand waving. This got a well written answer about how non serious my example reference was, arguing with a family anecdote. When I ask for better sources (I don't know the writer or his/her family from Adam/Eve) -- then I get no answer. :-)
  - nothrabannosir 10y ago
    You can see Wikipedia as a mixture of different "subpedias". I find articles on technical and mathematical topics to generally be extremely accurate, content rich, and surprisingly bias free (so not very, but still). Protocols, theorems, proofs, even very advanced topics (way over my head, so perhaps I can't tell...). You could do much worse than take everything Wikipedia says about mathematics at face value.* On the other hand, everything about currently living people, or worse, companies, you can essentially dismiss out of hand. * Please don't. I'm just making a point.
    - nickpsecurity 10y ago
      "I find articles on technical and mathematical topics to generally be extremely accurate, content rich, and surprisingly bias free (so not very, but still)." Same here. I usually just link the Wikipedia article because it's either totally accurate or accurate enough for what I'm trying to accomplish. I added a bunch of stuff to it as a result from my domain. I do spot marketing BS on occasion in there but the baseline makes it really stand out. Most people doing IT work are going to say, "Wait, that looks like a fan or advertiser wrote it" then dismiss that specific point.
  - jeff_petersen 10y ago
    > His rule of thumb is "If it isn't a major article there's probably errors" ...and if it is a major article, there are definitely errors, some likely introduced on purpose[1] (since major articles are the juiciest targets) [1] https://en.wikipedia.org/wiki/User:Ravpapa/Tilt https://en.wikipedia.org/wiki/User:Ravpapa/Tilt
- e12e 10y ago
  I always find it strange when people do that. I mean, it's a wiki - if there's an obvious, glaring error - just fix it? Then that particular article will be more correct? Wikipedia is absolutely useless as a dead book - the level of (professional) work that goes into vetting "real" encyclopedia articles is much higher - and the errors that slip through much more permanent. The great thing about Wikipedia is the potential steady improvement. When people complain about its accuracy it often just sounds like whining. Granted, there are a few articles that are "contested", and I understand why people don't have the time or energy to fix all of Wikipedia.
  - Tomte 10y ago
    No, your fix will often be reverted by those one or two authors who consider the article "theirs". If you fix it a second time, including a clear explanation in the commit message, you'll be accused of vandalism.
    - e12e 10y ago
      Do you know if anyone's done quantitative research on that? Should be pretty straightforward to sample some articles and see if that pattern is common? (I'm not trying to passive - aggressively suggest you're wrong, I'm honestly curious) [ed: I should add that I've started to form a habit of looking at the revision history and talk pages of wp articles - in addition to references. And I really do think the added transparency is a very real added value over dead tree encyclopedia. ]
      - berntb 10y ago
        I referenced this earlier, but got no answer. It is from 2005. I haven't seen anything after that. "Wikipedia is about as good a source of accurate information as Britannica, the venerable standard-bearer of facts about the world around us, according to a study published this week in the journal Nature." https://www.cnet.com/news/study-wikipedia-as-accurate-as-britannica/ https://www.cnet.com/news/study-wikipedia-as-accurate-as-bri... (Please Google a better source for that paper. :-) ) I am willing to believe that "sensitive" information about living (and relative unknown) people along with companies/political parties is being more gamed in Wikipedia today, than ten years ago. I'd just see some good references for that...
curiousgal 10y ago
Another factor might be the abuse of the bookmark feature. I tend to upvote stories that seem interesting when I am at a hurry just as a way to save them in the upvoted stories section for a later time. About the comments, pretty sure the sorting algorithm still takes into account a user's avg score. I usually make late comments but they show up well above other threads.
- berntb 10y ago
  I use the bookmark feature for articles that I want to keep, not read later. For instance, I seldom bookmark news. (I would, without data, guess your way is more popular.) Otoh, I often bookmark an article just to keep a reference to the comments.
paganel 10y ago
> So one way to improve HN submissions and comments is to weight points by the user's tenure on HN. I've been on this website for quite some time and I don't honestly care about "Show HN" links. I mean, I wish the submitters all the best in the world, it's just that I'm not visiting HN in order to see who is gonna get funding or not in a year or two, I visit HN for interesting links and some interesting comments.
tgb 10y ago
Interesting, but when I upvote Shown HN posts, it's because it contributed to my enjoyment of reading HN. That's largely orthogonal to whether it would be a good product, and often times is directly contradictory towards it (fun tech demos -> upvotes, products -> meh). I don't think that's a decline in the quality of HN, it's just a different focus. I use HN as entertainment rather than as a prediction market. But I can see that others would not like the shift.
- jschwartzi 10y ago
  Yes. And the analysis could be improved by coding each submission according to its intended status as a cool toy or a side project. I bet that as HN has grown we're seeing more toys and less startup projects on Show HN.
- tedmiston 10y ago
  You raise an interesting point; in comparison, Product Hunt is essentially the opposite.
tedmiston 10y ago
One consideration I'd like to point out is how the audience changed over time. I'm not saying anything like "HN is getting worse every year", but rather that it started at a startup accelerator, before startup accelerators were popular and at a time when startups weren't as popular as they are now. The crowd was more insiders at the beginning, and insiders may have more insight into what will / won't succeed from their experience. Another point, and I wish I had data but this is purely anecdotal, is that it seems like the side project --> startup transition is happening less. People are content making small side projects, doing hackathons, etc but not making that business their life (upsides and downsides to this).

berntb 10y ago

Cool. And sad. But re old accounts: I've recently seen a few cases where old accounts with high karma dismissed Wikipedia -- without any references of their own. That is a "quality smell", if anything is. (One particularly shameless fellow just copied the primary link Wikipedia used -- and lied about the content. :-( ) EDIT: I wrote without any references of their own -- i.e. dismissed with hand waving. Wikipedia won't always be correct, but it has references and is whetted by quite a lot of people.

50CNT 10y ago
My dad likes to dismiss Wikipedia, and he used to be an admin. They may just know how the sausage is made. His rule of thumb is "If it isn't a major article there's probably errors". His favorite example is that of an article about an ancient persian poet, which says that there's only 2 of his poems known, and only 1 remains. Next to that line, there's a picture of the other poem. Eh, who knows.
- berntb 10y ago
  I didn't argue that Wikipedia always was correct (no source is that, I'd guess), just that it smelled really bad to dismiss it with just hand waving. It do have sources and have been whetted (for reasonably popular subjects.)
  - 50CNT 10y ago
    Wasn't trying to imply you meant that. It's more that I could see where someone like that could come from, dismissing the Wikipedia article out of habit, especially if they've been involved with the project. Not backing it up with additional sources is of course bad form, but finding and linking extra sources is additional work. Minor nitpick, vetting is someone checking something and approving it, whereas whetting is grinding away metal to sharpen a blade.
    - berntb 10y ago
      OK, you more or less agree with my argument. Enough about that, then. I am curious about something else. Do you have non anecdotal information about Wikipedia's general dependability? How is the dependability for large subjects? For subjects like a poet where less than a hundred people even recognize the name? (The last I saw on this was years ago, when Wikipedia edged out the Encyclopædia Britannica. It would be strange to be less dependable now?!)
    - berntb 10y ago
      It seems I was trolled, but in a funny way. I discuss how you can see bad behaviour from even old accounts at HN, dismissing serious references with just hand waving. This got a well written answer about how non serious my example reference was, arguing with a family anecdote. When I ask for better sources (I don't know the writer or his/her family from Adam/Eve) -- then I get no answer. :-)
- nothrabannosir 10y ago
  You can see Wikipedia as a mixture of different "subpedias". I find articles on technical and mathematical topics to generally be extremely accurate, content rich, and surprisingly bias free (so not very, but still). Protocols, theorems, proofs, even very advanced topics (way over my head, so perhaps I can't tell...). You could do much worse than take everything Wikipedia says about mathematics at face value.* On the other hand, everything about currently living people, or worse, companies, you can essentially dismiss out of hand. * Please don't. I'm just making a point.
  - nickpsecurity 10y ago
    "I find articles on technical and mathematical topics to generally be extremely accurate, content rich, and surprisingly bias free (so not very, but still)." Same here. I usually just link the Wikipedia article because it's either totally accurate or accurate enough for what I'm trying to accomplish. I added a bunch of stuff to it as a result from my domain. I do spot marketing BS on occasion in there but the baseline makes it really stand out. Most people doing IT work are going to say, "Wait, that looks like a fan or advertiser wrote it" then dismiss that specific point.
- jeff_petersen 10y ago
  > His rule of thumb is "If it isn't a major article there's probably errors" ...and if it is a major article, there are definitely errors, some likely introduced on purpose[1] (since major articles are the juiciest targets) [1] https://en.wikipedia.org/wiki/User:Ravpapa/Tilt https://en.wikipedia.org/wiki/User:Ravpapa/Tilt
e12e 10y ago
I always find it strange when people do that. I mean, it's a wiki - if there's an obvious, glaring error - just fix it? Then that particular article will be more correct? Wikipedia is absolutely useless as a dead book - the level of (professional) work that goes into vetting "real" encyclopedia articles is much higher - and the errors that slip through much more permanent. The great thing about Wikipedia is the potential steady improvement. When people complain about its accuracy it often just sounds like whining. Granted, there are a few articles that are "contested", and I understand why people don't have the time or energy to fix all of Wikipedia.
- Tomte 10y ago
  No, your fix will often be reverted by those one or two authors who consider the article "theirs". If you fix it a second time, including a clear explanation in the commit message, you'll be accused of vandalism.
  - e12e 10y ago
    Do you know if anyone's done quantitative research on that? Should be pretty straightforward to sample some articles and see if that pattern is common? (I'm not trying to passive - aggressively suggest you're wrong, I'm honestly curious) [ed: I should add that I've started to form a habit of looking at the revision history and talk pages of wp articles - in addition to references. And I really do think the added transparency is a very real added value over dead tree encyclopedia. ]
    - berntb 10y ago
      I referenced this earlier, but got no answer. It is from 2005. I haven't seen anything after that. "Wikipedia is about as good a source of accurate information as Britannica, the venerable standard-bearer of facts about the world around us, according to a study published this week in the journal Nature." https://www.cnet.com/news/study-wikipedia-as-accurate-as-britannica/ https://www.cnet.com/news/study-wikipedia-as-accurate-as-bri... (Please Google a better source for that paper. :-) ) I am willing to believe that "sensitive" information about living (and relative unknown) people along with companies/political parties is being more gamed in Wikipedia today, than ten years ago. I'd just see some good references for that...

curiousgal 10y ago

Another factor might be the abuse of the bookmark feature. I tend to upvote stories that seem interesting when I am at a hurry just as a way to save them in the upvoted stories section for a later time. About the comments, pretty sure the sorting algorithm still takes into account a user's avg score. I usually make late comments but they show up well above other threads.

berntb 10y ago
I use the bookmark feature for articles that I want to keep, not read later. For instance, I seldom bookmark news. (I would, without data, guess your way is more popular.) Otoh, I often bookmark an article just to keep a reference to the comments.

paganel 10y ago

> So one way to improve HN submissions and comments is to weight points by the user's tenure on HN. I've been on this website for quite some time and I don't honestly care about "Show HN" links. I mean, I wish the submitters all the best in the world, it's just that I'm not visiting HN in order to see who is gonna get funding or not in a year or two, I visit HN for interesting links and some interesting comments.

tgb 10y ago

Interesting, but when I upvote Shown HN posts, it's because it contributed to my enjoyment of reading HN. That's largely orthogonal to whether it would be a good product, and often times is directly contradictory towards it (fun tech demos -> upvotes, products -> meh). I don't think that's a decline in the quality of HN, it's just a different focus. I use HN as entertainment rather than as a prediction market. But I can see that others would not like the shift.

jschwartzi 10y ago
Yes. And the analysis could be improved by coding each submission according to its intended status as a cool toy or a side project. I bet that as HN has grown we're seeing more toys and less startup projects on Show HN.
tedmiston 10y ago
You raise an interesting point; in comparison, Product Hunt is essentially the opposite.

tedmiston 10y ago

One consideration I'd like to point out is how the audience changed over time. I'm not saying anything like "HN is getting worse every year", but rather that it started at a startup accelerator, before startup accelerators were popular and at a time when startups weren't as popular as they are now. The crowd was more insiders at the beginning, and insiders may have more insight into what will / won't succeed from their experience. Another point, and I wish I had data but this is purely anecdotal, is that it seems like the side project --> startup transition is happening less. People are content making small side projects, doing hackathons, etc but not making that business their life (upsides and downsides to this).