However, average karma per post gently nudges us towards avoiding circumstances where a comment might get low karma, such as on a post that is less popular.
Whereas, total karma encourages users to let a thousand flowers bloom. There's a disincentive to post anything that might get downvoted, and that's fine, but otherwise there's no harm in making a comment anywhere and not worrying about whether your seed has fallen on fertile ground.
You could "fix' things by weighting karma per post by eyeballs, or disregard average karma per post. I choose the latter course as I'm not a moderator and even if I were I'm unsure how the number relates to behaviour HN is trying to encourage.
The attitude you are advocating is an internet-age-old drumbeat since Slashdot introduced moderators, and we are better for it, but the fact is that karma does affect people's behavior is undeniable and statistically immutable; therefore your point doesn't really address raganwald's concern.
At this point in software history, all software is social software. Therefore, discussions about social behaviour and how it is affected by mechanisms and features is extremely relevant to the business of making software. So I'm fascinated by the subject of karma even if I don't try very hard to accumulate any.
I think a time-based system might go along with karma well. If you stay above a certain level for long enough, you get access to the next feature level.
You still have to work your way up the karma tree, but you're not doing it for the features.
HN is based on the quality of your ideas.
I'm sure a lot of good discussions never exist because of this. Even without avg. karma though, reputation is still at risk, so the problem won't be solved by removing/showing that statistic.
There's nothing wrong with caring about HN karma, so long as you care about karma less than you care about speaking your mind.
Since this may cause people to care about comment averages, here's how they're calculated:
; 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))
Comment averages are updated asynchronously.the game becomes that of other threaded comment systems where you get more up votes on heavily traffic-ed threads. it should push the long tail of comments further out, and have fewer threads with high traffic. someone trying to game the system might look for threads with say 10-15 points and then make a comment, hoping they get above the average. it's like investing in DOW companies during an economic upswing.
your karma actually should not be purely based on uptick, but should represent some amount of risk you took, and then measure that against how well the community responded to it.
so commenting on old or low-traffic threads could still be of value.
I am going to experiment with labeling things as experiments.
And it definitely discourages commenting in low-traffic threads when the average isn't normalized for viewers.
1. Median obviously is less suspect to a small number of very
popular comments which probably describes many users pretty
well. It's the "betting value" for a game where loss is
linear with error.
2. I'm not sure whether median would be higher or lower for
most. Poisson distribution suggests higher medians, but
many probably have a bunch of 1-karma comments to weight
the origin.
3. A HUGE psychological difference is that median would be
discreteness. Instead of seeming to be a smooth value
related (probably naively) to community engagement and
improvement it'll look more like a set of clubs. 0-medkarma
could be a rabble-rouser, 3-medkarma is consistently
meaningful. 10-medkarma means maybe you should run for office.Psychologically, it diverts attention away from a specific number, and toward a general concept of "goodness", segregated by club. Then, rather then concentrating on getting the next karma point, users try to enter certain club, and each club has a general identity to it (with the top being those who provide thoughtful, intelligent responses, and typically only speak when they have something to say).
Creating different clubs (creating classes of users) isn't arguably bad so long as there is easy mobility between classes based on merit.
The natural question is how and whether to indicate the "class" or "club" a user is in. Make it public? Indicate it in comments? PG tried a red dot next to high-karma users some time ago, then scrapped it; maybe there's some other way to reward users based on club achieved.
Both thoughts depend on figuring out more completely what club membership means; it's something I can't flesh out well despite having a feeling it might be beneficial.
What if the average rounds everything under some threshold (say -10) up to that threshold? So your total karma can get totally trashed by one heavily-downvoted comment, but there's a limit to how badly your average gets hit.
But really, I'd guess that 99% of comments that go below -10 are clearly intentionally trolling, so it's probably not an issue.
I think typing new function definitions into your live site seems very appropriate for a place called Hacker News.
Downvote threshold now 500. Glad I got to the old threshold (201) in time to see how it worked.
Think my possible upvote demographic for this comment is the set of all Hacker News users between 201 and 500 karma.
I'd love to have a running average of my last 10-20 comments rather than ALL of them.
My average at this moment is 7.64. By giving all this weight to avg. karma, the message from the site to the users is clear: "higher avg. karma is better, work to improve yours". In my case, this means I should make a comment only if I expect it to get 8 upvotes or more. This in turn means that I should not respond to replies to my comments, as they rarely get to 8+. I should also not reply to articles late in the game as it's unlikely to get to the threshold also. So what will I respond to? Early threads where I feel I can make a contribution. It is also unlikely that I will make 'eccentric' comments that are a risk in terms of votes, but may help break the echo chamber as this behaviour is not encouraged. The only other case I may make a comment is when I feel strongly enough about something that I am willing to take the hit, as in the case of this comment.
Now, many will say that I should ignore karma altogether and that I should just participate in the community as if it didn't exist. For one thing, this is orthogonal to the discussion here. pg has made this change to alter the community's behaviour, therefore he expects it to have an impact. Even if I chose to disregard it, nothing will change in the general case. Secondly, poster avg. karma affects ordering of comments and therefore the impact that your comments have. So it materially affects your experience in this community, and when I have something to say that I feel needs to be said, I benefit from it having the highest possible impact. This makes me care about avg. karma in all other situations.
On the other hand, submitting articles is not subject to such pressures, so I will happily continue doing that.
To me this is the biggest negative. A useful but non-controversial threadlet between members towards the end of a submission's life is IMO beneficial to the community overall but will be worse for average karma for participants.
I'll live with the lower average karma if it means still benefitting from the community in a way average-karma-targeting replies don't.
1. When your karma goes up/down, it's a good indication that you might have replies to your most recent comments.
2. Karma seems to be updated not so frequently.
I like the idea though of encouraging a good ratio vs total karma.
I would like to see total karma disappear altogether, because I don't think it measures anything helpful and may well be counterproductive. Looking down the HN karma leaderboard[1] you can see a number of users with very high total karma but very low average karma. I'm sure that some of these users make a valuable contribution, but some seem to post a vast number of very low-value comments that are rarely replied to and receive only occasional upvotes. I'm not sure that we want to encourage that style of participation.
I comment in a lot of new stories/links that pop up, especially people asking critiques because a) it's disheartening to post something about your work and have zero replies and b) I often get to see a cool product that's in a very raw stage and hear feedback from the creator themselves. In fact, I'd say less than 25% of my comments are in front page links.
But often these new stories rarely get traffic. And my comments will usually just get one or no upvotes at all from the creator, despite their thanks for the criticism.
Just saying that not all low-averaged scores are people not actively trying to make a contribution.
I would definitely love an equivalent of the reddit envelope, along with a view of replies.
Total Karma of a user can be seen on his profile page. Afaik there is no way to see another users average karma, unless he is on the leaderboard, which uses total karma as the ranking metric.
http://news.ycombinator.com/item?id=1741408
http://news.ycombinator.com/item?id=1485254
http://news.ycombinator.com/item?id=1349459
http://news.ycombinator.com/item?id=1349455
Those are all comments by me that were helpful to the parent, but which are "punished" by the public display of avg.
If you want it scaled to fit a continuum of "bigger is always implicitly better", a more reasonable number would be: average of (comment points/parent points)
Even if a number has no [absolute cardinal, > x] meaning, there is still a nominal information value to the datum.
This could help with another gripe I have with displaying average points (and comment voting in general), which is that oftentimes quality comments are downvoted simply for being contrary to the downvoter(s) views, and/or groupthink.
Personally, I like the average. Shows that HN is more 'adaptive', imho.
ETA. Whoa that was fast. I like the (abs., avg.), fwiw. That does seem best.
To join in the 'me-too' choir, rather than returning the arithmetic average, perhaps you could slap a confidence interval on it and return the lower bound.
One feature I'd love to see also, is the # of direct replies to my comments in the head bar. Also, while we're at it, a page listing them all too. That way, I'll know if somebody replies to a comment I made last week/month/year. For example, in 'My dalliance with smart drugs - and the lesson I learned' (http://news.ycombinator.com/item?id=1850750), a few people have ordered some of the drugs mentioned, but the story will be dead soon, so in a few weeks, when the drugs arrive, nobody will be able to share their experiences.
As of now, I have to re-read my threads in order to notice any updates.
it helps a lot with that. It has some bugs, but as far as I can see I can't get around them because of the markup used here ...
An average does more to encourage the desired behavior than a total does. Even as I am about to press enter here, I can't help but think of whether this post will make that number go up or down.
I'd also argue that there are only a few topics that are so polarizing that people up-vote or down-vote based on the position the author takes, and that the majority of discussions here aren't as opinionated, and that the number of up-votes a post receives is based on how interesting and thought-provoking it is.
However, it appears that how to display karma is one of the more polarizing ones :)
Order Who Total Ave
2 edw519 44607 22.29
39 grellas 11208 19
45 wheels 10752 12.15
4 patio11 37809 11.23
42 mixmax 11064 10.45
66 timr 8632 9.78
78 chaosmachine 7829 8.8
93 niyazpk 6885 8.28
70 alexandros 8353 7.98
53 jerf 9873 7.8
1 tptacek 45718 7.49
16 mechanical_fish 16598 7.4
17 raganwald 16421 7.27
97 maxklein 6751 6.68
34 ryanwaggoner 11976 6.63
76 _delirium 7895 6.56
49 lionhearted 10137 6.49
87 acangiano 7001 6.39
22 Anon84 14263 6.24
9 swombat 21933 6.09
75 abstractbill 7910 6.09One further thought: your "karma" in a site like HN is a very serious asset. Access to a post which reaches the front page can make or break a company. I would easily pay $100 for someone to post a choice blog post reference that would hit the front page, and I'm sure I'm not alone.
We also know about Digg's kabal of a few high-karma individuals who tilted the entire news site to the political right.
People don't game a site until it's important, and you don't find out the effects of your site's algorithms until the site becomes popular.
The owners of a site must be diligent, continue to understand how articles are floating to the front page, and be willing to make changes. We as users need to be supportive of the changes & experiments they make - even at the cost of our own "karma" (ie, real-world-dollars).
The reality is that the earlier you post in a story, the more karma you're likely to collect. To me, that just encourages "first!" type posts. A poorly thought out comment in a story that makes the front page will inevitably earn more karma than a well thought out comment that doesn't.
So is this about getting better stories to the front page, or about rewarding users for their participation?
As far as I can tell, it's one user, one vote (I mean, except in the common case of dupes) to my knowledge, higher-carma users don't get extra points to vote up stories.
"Ah, the karma point system is broken and they're showing the number pi instead, very funny."
(This is question, NOT a feature request, as I presume that Karma at HN serves the purpose of moderating the quality of content, rather than being a redditesque user-stickiness game)
I think those of us that care, but occasionally say something bad will be helped a lot by this. I'm always trying to ask myself "will this be upvoted" when posting things, as I'm trying to be as helpful as possible.
However, people that are actually trying to troll or can't tell that they're trolling won't really stop because of this change. I don't really have an answer, either: this is a Hard Problem.
Edit: I should also say that any of us can easily be in the "can't tell we're trolling" camp at any time; it's a per-comment thing, not an "I'm a good poster, what shall we do about those terrible bad ones!" deal.
I don't so much think average is important, possibly a derivative of it would be average amount of posts to receive down votes. We don't want to discourage people replying to threads that are falling off the front page, even if the only person who sees the post is the person that the post was in reply to, that still sounds like a net gain if it was useful or made a good point.
Karma doesn't matter as much to me because I come to the site for content, but I like to comment on a lot of new links (especially entrepreneurs wanting some design critiques) which doesn't get nearly as much participation as the front page and thus has a much lower chance of gaining a lot of karma per comment. A total karma score doesn't discriminate against that but an average score does.
Just keep on using the HN like you normally do, write insightful comments, provide utility for your fellow hackers.