How I ended up with so much Hacker News karma
blog.jgc.org
blog.jgc.org
I don't think I've ever posted a comment and felt that the score (or attention) didn't fairly reflect the value of the comment[1]. Something about the uniformity is very reassuring, unlike reddit where you need to either be in the first 5% of comments or provide something seriously exceptional that appeals to every reader to even have a chance of being listed near the top of a submission.
From my experience the "best" scoring comments are easy to pinpoint why they've scored well and if someone was so inclined they could very easily spend their day "farming" karma here, but if someone were to do so they'd be benefiting the site because the only real way to obtain karma here is to provide value. Even if that value is trivial (like mentioning a relevant article, or giving a brief anecdote that validates a point) it's still value.
[1] Sometimes controversial topics result in "unfair" scores, but they're also very easy to expect before hand and generally if you post something controversial you'll get good enough discussion from other users that it's never a question of "I wish I knew why people gave me that score!" which is the most frustrating thing.
So I upvoted you because you wrote first. You have to first mover's advantage. And I believe it is fair.
Follow-up responses will tend to be received in the same vein as the initial post, and if the initial post gained traction the one's follow-up is likely to garner upvotes as well.
For example I do not think that http://news.ycombinator.com/item?id=4843141 is anywhere close to being the best comment that I've made in the last month, but I have to go back to http://news.ycombinator.com/item?id=4684916 to find one that is higher rated.
It's a powerful disincentive for putting much thought into comments.
http://news.ycombinator.com/leaders
here on Hacker News from time to time is that pg's last major change in the site code (to make comment karma scores invisible to persons other than the person posting the comment) really seems to have made a positive difference in the site. That change was quite controversial when it was made.
Site founder pg began the experiment by opening a thread, "Ask HN: How to stave off decline of HN?"
http://news.ycombinator.com/item?id=2403696
After looking at the discussion there, he announced an experiment
http://news.ycombinator.com/item?id=2434333
601 days ago of not publicly showing comment karma scores, but rather just letting each user see her or his own comment karma scores when logged in. That change was protested in several LONG threads in the first few months after the change, but what I noticed is that I started seeing the comments of the users from whom I learned the most, the most often, more consistently sorting to the top of threads, and on the leader board I noticed that the AVERAGE comment scores (still disclosed there) of the 100 highest-karma users began better matching who was making substantive, helpful comments as contrasted with who was making comments that were, in pg's words, "a) mean and/or (b) dumb that (c) get massively upvoted."
A reality check on what kind of comments get upvoted is to look from time to time at the bestcomments list,
http://news.ycombinator.com/bestcomments
which is still not exactly really a "best of" list of HN comments, but which has converged to better quality since that software change. The user jgrahamc appears to have seen his average comment karma score rise under the new software settings, deservedly so. Other users whose comments are especially thoughtful, whose user profile pages I visit to look up their most recent comments, also appear to have gained in average karma over the last year and a half that that software change has been in effect.
My best posts (on QM or molecular dynamics or color theory) don't get that many points but I still write them because I feel like they are informative and somebody could get a lot of use out of them. I occasionally also write semi-controversial posts that I know won't get many upvotes, but I still feel the arguments I make can be useful to others or hopefully at least spur thought.
On the other hand, submissions remain a mystery to me. My highest voted submission (466 points on old Russian photographs) nearly slid off the new page before being caught right at the end with a surge of upvotes. However, my submission on Titan (now the world's fastest supercomputer) got no interest whatsoever, which really surprised me given the demographics of this place.
Or people have noticed that you tend to write comments that are either well-designed, or that others widely agree with -- and because you've developed that brand, some people now upvote you because they recognize your name instead of anything to do with your posts.
Basically, people get used to seeing your name as one of the top comments, and upvote you more often, even if they believe your comment to be equal to others.
There is also the timing aspect however; habitual commenter are more likely to be in early to a thread.
If you see upvotes (U) as a product of the number of people viewing HN at a specific point in time (V) and the percentage of people who will upvote the post (Q) multiplied by the number of points in time (T) during which the post is viewable, we find U=VQT.
We can increase U by posting when V is highest, by posting comments when the most people are visiting HN. We can increase U by increasing the % of people who like the post, which usually means writing a high-quality comment. We can increase U by increasing T, the amount of time the post is available, simply by posting it earlier.
Thus comment quality is still the long term determinate. Particularly, now that snark and meanness are less acceptable, more thoughtful comments are required.
What I have found is that writing comments that get upvotes provides valuable feedback about the mechanics of writing comments that get upvotes - just as writing downvoted comments provides feedback (more quickly). Writing one good comment increases the likelihood of writing another one - writing bad comments, also increases the likelihood of good comments via downvotes.
HN provides editorial feedback in a form other than LOL's - comments and votes. Editorial feedback is a proven way to improve writing, and improving writing is the surefire way to get higher comment scores.
I suspect the styling on the site, which presents non-comment bits in a lighter grey color, helps to de-emphasize the authors' identities in favor of the comments themselves.
But after you hit like 1000 karma...you just don't care anymore. At least, I've pretty much stopped looking at my karma score.
Edit: This could be related (someone please correct me if I'm wrong), comments are allowed some time at the top before they "mature."
(let mature (table)
(def delayed (i)
(and (no (mature i!id))
(acomment i)
(or (< (item-age i) (min max-delay* (uvar i!by delay)))
(do (set (mature i!id))
nil)))))Also, PG has made so many unpublished modifications - for obvious reasons - to karma, spam filters, voter-ring detection, I wouldn't put too much stock into sussing out the inner-workings of HN through the original Arc download.
The main point is that it's not instant (i.e. your last comments are not taken into account) and it uses only recent comments, rather than over the life of your account.
Why did you care in the first place?
Silly I know, but I also recognize that it is what made me try to write good comments when I first joined HN. Which is exactly what karma should do - make you try to be a better poster.
However, I would think that patio11's, edw519's and others' reputation attract karma more easily than the "regular" commenters. I mean that people are more easily convinced by whatever they say because they trust their reputation.
>> he sorts automatically to the top of the page. Which then gets him more karma, etc etc.
Patio is one of my favorite posters, I'm not denigrating him at all. He certainly deserves all the karma he gets.
Now /leaders isn't highlighted anymore, I spend less time contributing and more time lurking. This is neither good nor bad as I am statistically insignificant in the great tide of stuff on here. But we're all motivated in different ways in different contexts and I've learned over the years a main motivator for me is credit/recognition/"glory" (and not the points per se). Now I know that's a 'quirk' of mine, I take advantage of it as a way of getting my work done and stopping procrastination, rather than trying to do well on HN or wherever ;-) Except for this comment, surely..
Glib and offhand attempts at humour are not much appreciated here, even though they might contain some wit, which reddit would gobble up.
The HN audience seems to upvote comments that are analytical, insightful and informative. This takes some adjustment for reddit heroes who can reap great karma with a well-placed meme.
Because there's nothing worse than seeing 7,000 comments on a Reddit article and 99.7% of them being complete garbage. That's discouraging to a community. Why even bother posting when someone's lame joke is 347 votes ahead of yours. So I'm a complete asshole on HN. When I see jokes or one liners I downvote away. I feel guilty and sympathize for the poster but... I like the noise free environment too much. So I wipe away the tears and downvote away.
An equally good article/discussion is what led me here. Of the aggregation sites I know of on the web, this place is currently the only one that retains (for the most part) it's signal to noise ratio in the comments.
Reddit anymore, even though being subscribed to very specific subeddits, is just frustrating to use. It seems to me that the average age of the user has shifted downward as the site has grown. Couple that with a lack of moderation, and comments are all now one line, or a 'funny gif' response, which, by the way, are strictly outside of reddiquette...
Slick attempts at humor don't certainly get downvoted, they may not be the top comment. I still think that the reddit model when brought down to Subreddits still works. I am part of some very tiny subreddits and there is still a free flow of high quality information.
http://news.ycombinator.com/item?id=4631280
http://news.ycombinator.com/item?id=4431975
http://news.ycombinator.com/item?id=4754809
These comments were punished by HN:
http://news.ycombinator.com/item?id=4789162
http://news.ycombinator.com/item?id=3515735
http://news.ycombinator.com/item?id=3515784
I would like to find an online community that welcomes and values the latter kind of participation.
1 (FizzBuzz): You're responding to a strawman ("unrealistic audition programming for a live audience") and dismissing, rather than responding to the very point the article makes at great length ("Perhaps the hiring pool really is that bad"). I wouldn't have voted it down, but it's honestly not much of a value-add.
2 (Craigslist middleman): Again, I wouldn't have downvoted, but you're being assertive for one side in the debate without backing your assertion by reasoning or arguments. Also, it might just be me, but I catch a whiff of aggression in your comment.
3 (same thread): You're aggressive and you're asserting that some nefarious secrecy is going on, when no-one else has claimed that.
Granted, in both of 2 and 3, there's a likelihood of you being voted down because it's a story about an entrepreneurial guy doing well, and you disagreeing with the legitimacy of what he's doing. This is a very entrepreneurial forum, so there's a bias against that.
Downvotes which one feels are unwarranted are best considered editorial feedback. Take them as evidence that your writing did not match the expectations of the audience which received it. This may be due to misassumptions about the nature of the audience, or due to a miscommunication of the point one is trying to make. The solution to downvotes one feels are unwarranted is improving one's writing.
The point is that I am questioning what those expectations of the audience are. If you have read those threads, preferably in context, would you mind giving me your own personal judgement about whether or not my voice in those discussions is unwelcome or should be removed from the room? Particularly when set aside those comments of mine that were deemed to be welcome?
How could your posts have been better written?
Was their content thoughtful enough to be worth
submitting in the first place (or not deleting
after having been posted)?
Before posting the grandparent, I gave the links a cursory review. It was pretty clear that there was nothing outstandingly insightful about the comments. There was a bit of meanness, however. Their tone was a bit argumentative without being informative, as well.
All one can do is support their writing with effort and bounce back from mistakes and rejection as an author.
[Edit] I will add that comments do get systematically downvoted in a thread when someone disagrees. This morning, all my comments in this thread were downvoted by one, in a very short period of time. It is not unsurprising given what I compared and the entity I criticized. If I was certain I had never done the same thing, I might complain about it. http://news.ycombinator.com/item?id=4864977 But I think I might have had a bad day, myself.
Granted, in both of 2 and 3, there's a likelihood of you being voted down because it's a story about an entrepreneurial guy doing well, and you disagreeing with the legitimacy of what he's doing. This is a very entrepreneurial forum, so there's a bias against that.
It's difficult for me to imagine anyone reading those threads and missing that conclusion. It seems to me that it is the primary driver.
The fact that you write "I agree with the broad points" doesn't change the fact that your comment disagrees with the broad point (that FizzBuzz is a good and useful test). Also, the bit you quote seems unrelated to your concerns. How do you disagree with the interpretation that 40% passed the test? Any interview situation is "for an audience" and stressful. Not testing actual coding (rather that talky-talky skill) is deeply unfair to those many talented coders who are not great sales people.
For two and three, the comments are still not good ones, they're just further being ill-considered because of the bias. There are plenty of good comments that go against the bias of this site, but still get upvotes, because they're good.
The broad point of the article was that FizzBuzz (still) reveals an amazing number of unqualified applicants. "Why, look what happens when we do FizzBuzz... You should too."
I did not disagree that FizzBuzz does that very thing. My objection was this: "Only 40% of programmers could correctly write out a FizzBuzz program on paper during our timed test in the middle of a job interview" is not equivalent to "Only 40% of applicants even have the skill to code FizzBuzz." The difference between the two is vital in a way many people here don't see, or at least won't acknowledge. That was the subject of my comment.
Maybe entrepreneurial types are fundamentally more at ease with that kind of thing than coders who aren't wired that way. I don't know how to enable you to see into that blind spot. I could try, in a meaningful back-and-forth with those who disagree with me. But to simply dismiss my initial objection outright, as something that doesn't even belong in the conversation? To employ HN moderation that sidelines it so it won't be seen? That leaves nowhere to begin.
I value being able to argue with intelligent, tech-minded people who disagree with me. It matters so much to me that there be an online venue like HN to meet unfamiliar points of view, to test my own biases, to figure out how I actually see things by engaging in meaningful debate. But we can't have such a low tolerance for disagreement and for challenges to biases that a whiff of a problem with someone's tone makes that person's words seem "mean." We can't be so quick to silence people at square one just because we don't (yet) see their point.
Please believe me that this is sincere and not pedantic. I just want to be clear: When you compare the first set of three comments above with the second, you really see the first as more valuable to HN than the second? Let's agree that no submission or comment is going to be all bad or all good and that we won't quibble about details. On balance, you would say the second set is unwelcome when compared with the first?
OK. You made that point very poorly. I only understood that now. Also, I disagree, but that's offtopic here.
You were not downvoted because you made an unpopular point, or because people disagrees with you. You weren't "silenced". Understand that. There is an extremely high tolerance for disagreement here, but it has to be well phrased and well argued.
You were downvoted because you made your point very poorly, so it read like an argumentation-void grasping at a strawman. I'm completely willing to accept that wasn't how you meant it to come out, but it's your responsibility, and only yours, that your comment comes across as intended.
I never responded to the first three comments - they were upvoted because you hit a popular vein in that particular discussion. That's pretty common. By the way, here's one for challenging the bias on this site: http://news.ycombinator.com/item?id=4827567 - it's got 53 points. It's not a particular well written or insightful comment, it's a fairly trivial anecdote that happens to contain the magic word "YC". It's probably one of my all-time highest scoring comments, to think that means it's the best depresses me. A lot.
Whenever I get one, regardless of how many upvotes I've gotten, I delete the post immediately.
Sometimes, I'll also delete any other recent posts done in the same reading session, if they're of a similar quality.
This had the effect, over time, of increasing my karma. It also caused me to write things that are generally uncontroversial and, well, bland.
Second, there's the point of HN karma, especially of the magnitude that can never reach the leaderboard, is completely worthless. I'm sure being on the leaderboard (ie. minor HN royalty) can open some doors, if you want it to, but yours and mine never will.
It's true that jgc writes a lot, but I'd say his average article is pretty average. His best articles are great, but it seems somehow he manages to score a run for a blog post that's really worth a ball or a single. It's hard not to believe that a voting ring is involved somehow.
Take that back. It's not true: I have no voting ring for my posts.
And here's some data to back that up. According to the HN Search API I have submitted 373 items that are posts on jgc.org (out of a total of 670). Of those the following karma has been gained:
Points Number of posts
1 81
<5 102
<10 32
<50 74
<100 47
<200 6
<500 31
All those posts account for 12,624 points of karma (from my total of 18,119 points obtained from submissions, not comments). So 56% of my submissions were for my blog and resulted in 70% of my submission karma.
I suspect that there is significant overlap between his regular readers and HN participants. People read his blog and submit it.
Furthermore, articles about writing and karma acquisition are often topics of interest to HN's readership, as are articles about HN.
There's no conspiracy needed to explain it. Not to mention that his karma score from comments alone would put the author on the leader board.
Just for fun, here are the posts jgc makes by domain:
And here are the posts of the blog.jgc.org domain by poster:
If I had any articles that are great, I'd be very happy indeed. I might even submit everything I wrote, just to let the community decide which are great.
And in fact as I can see there's no much use in HN karma - smart comments are more useful. At least, they can do you good when you apply to YC (that's what PG says at least :)).
He then proceeds to do exactly that, and give us all a good tactic for increasing our karma: Post Consistently Interesting Stuff over a Long Time. Sure, we all read that and think, "_duh_, of course that's how to get it", but it's nice to see that there's data backing it.
And I understand your reasons and they're quite good. I don't think that it's a bad article to vote up - moreover I like the article too. There's some "pure math" magic in it that I also like (although I'm a humanitarian as I've mentioned before).
Take the following scenario: you've got some interesting topic that most HN users will click on, and the #1 comment is like this:
--------- parent comment line 1
--------- parent comment line 2
--------- parent comment line 3
...
--------- parent comment line x
>>>>--------- reply comment line 1
>>>>--------- reply comment line 2
If you're in that jittery state of mind when you just want some infotainment without reading dozens of lines of dense and subtle text, isn't it more likely that you'll just read the reply to that top comment first, since it's short and snappy? It may even make you feel that you've made progress towards understanding what the parent comment was about, without actually spending the required amount of time on it. Topics on HN obviously have a strong bias towards solid intellectual content (and I don't necessarily mean academic topics), so it's probably unlikely that much insight can be gained from very short comments - unless the poster is particularly observant and skilled in condensing information, of course.
That's just one of the reasons why I'm sceptical about karma. I think it does serve a purpose, but if the majority of users go into a topic with no real intention to read or comprehend anything thoroughly, karma becomes a less meaningful measure of contribution.
I think people put too much weight on things like this. I've been on message boards for over 12 years. People know me because of my name and what I contribute, my karma (which we don't have) or likes or friends don't matter.
(def update-avg (user)
(= (uvar user avg) (comment-score user))
(save-prof user))
; Ignore the most recent 5 comments since they may still be gaining votes.
; Also ignore the highest-scoring comment, since possibly a fluff outlier.
(def comment-score (user)
(aif (check (nthcdr 5 (comments user 50)) [len> _ 10])
(avg (cdr (sort > (map !score (rem !deleted it)))))
nil))Personally I don't much like this: average karma is a fairly good proxy for what I want to be optimizing in my HN participation, but the way HN reports it more or less pessimizes it as a feedback tool for me. (I'm a fairly infrequent and "bursty" commenter; HN's reporting of average karma isn't up-to-date enough to tell me what HN thinks of what I've been doing recently, nor is it long-term enough to tell me what HN thinks of me overall.)
The right answer, of course, is not to care much about karma. This is made easier by the fact that the correlation between HN karma and my own judgement of comment quality is weak, and I'm arrogant enough to think the latter is more accurate.
From my point of view however, getting good points on a comment is a lot more satisfying and worthwhile than your general Karma count (once you're over the various thresholds).
If you get pithy, insightful, argued, or just convenient comment in on a top story before it becomes a top story, you'll see an order of magnitude more points rolling your way than if you contribute research or elucidation in a day-old thread. A similar situation happens with submissions - you need reasonable content, but it can die with 1 point if the timing or title don't hit the sweet spot.
Content functions only as a high-pass filter.
In all seriousness, I really wouldn't worry about it at all.
If you're not just being yourself, asking questions with candor when they come up, and chipping in with answers when you can, you're going to miss the point of HN.