The recent implementation of a -5 karma penalty for replying to comments that are at -3 or below has clearly met with some disagreement and controversy. See http://lesswrong.com/r/discussion/lw/eb9/meta_karma_for_last_30_days/7aon . However, at the same time, it seems that Eliezer's observation that trolling and related problems have over time gotten worse here may be correct. It may be that this an inevitable consequence of growth, but it may be that it can be handled or reduced with some solution or set of solutions. I'm starting this discussion thread for people to propose possible solutions. To minimize anchoring bias and related problems, I'm not going to include my ideas in this header but in a comment below. People should think about the problem before reading proposed solutions (again to minimize anchoring issues).
This rule is asinine.
If I see a post at -3 that I desire to reply too, I am incentivized to upvote it so that I may enter my comment.
Furthermore, it stifles debate.
Look at this post of Eliezer's at -19 In the new system, the worthwhile replies to that post are not encouraged.
In the new system, instead of people expressing their disagreement, they will not want to reply. The negatives of this system grossly override any upsides.
I have not noticed a worsening trolling problem. Does anyone have any evidence of such a claim?
Can someone please provide hard data on trolling, to assess its kind and scale? I can only remember a single example of repeated apparent trolling - comments made by private_messaging and presumed sockpuppets. I'm not very active though, and miss many discussions while they're still unvoted-on.
Instead of trying to stop noise, you can filter it. Instead of designing to prevent errors, you can design to be robust to them.
I'll repeat something I said in the other thread:
To the extent that all the griping over signal to noise is about a desire to control what you see, and not control what others see or say, there are decades old solutions to discussion filtering. The fancy shmancy Web has been a marked deevolution of capabilities in this regard. It's pitiful. No web discussion forum I know of has filtering capabilities even in the ball park of Usenet, which was available in the 80s. Pitiful.
I also suggest that any solution which is not fundamentally about user customization is a failure under my assumption above, because one man's noise is another man's signal.
You've made me understand the root of one of my own dissatisfactions with the current system. If I look through my post history and roughly group my posts into bins based on how I would summarize them, this is what I see:
Silly posts in the HP:MOR threads: ~ +20 karma
Posts of mine having little content except to express agreement with other high-karma posts: ~ +10 karma
Important information or technical corrections in serious discussions: ~ +1 karma
Posts which I try to say something technical which I retrospectively realize were poorly worded but could have been clarified if someone pointed out an issue instead of just downvoting: ~ -5 karma
Perhaps I exaggerate slightly but my point is that if I were to formulate a posting strategy aimed at obtaining karma, then I would avoid saying anything technical or discussing anything serious and stick to applause lights and fluff.
On top of this, I tend to watch how the karma of my most recent comments behaves, and so I notice that, for example, a comment might have +5 upvotes and -3 downvotes, with no replies. This is just baffling to me. Was there something wrong with the post that three people noticed? Were the three separate things wrong with it? Was it just a response to the tone? What about the upvotes, is it being upvoted because of the witticism at the end, or because of the technical content in the middle? My point is something like Slashdot has a system where things are voted "funny" or "insightful" would be infinitely more useful.
Perhaps I exaggerate slightly but my point is that if I were to formulate a posting strategy aimed at obtaining karma, then I would avoid saying anything technical or discussing anything serious and stick to applause lights and fluff.
That's about right. Also, stick to high traffic threads. Hit the HPMOR threads hard!
As I pointed out that people want different things out of the list, you finish by pointing out that the karma votes themselves are clearly used differently by different people. They're also used to a different extent by different people.
One nice thing that Slashdot does is limit your karma votes. That keeps individual Karma Kops from have a disproportionate effect on total score. But I don't think the Slashdot system of multiple scores is that helpful.
From my experience in the grand old days of Usenet, the most useful filters were on people, and the important ease of use features were a single screen view of all threads, expand and contract, sort by date or thread, and sort by date for a subset of threads.
Wiki
We would benefit from more and better wiki articles since they seem the best way to compress information that is often scattered across several articles and dozens of comments. This should help us maintain our level of discussion by making it easier to bring users up to speed on topics.
I used to think the most straigthforward fix would be:
Eliminate the trivial inconvenience of creating a separate count for the wiki. Lets just make it so you use your LW log in. Also perhaps limit edits to people with more than 100 karma, since I hear they had some problems with spamming.
Let people up and down vote edits. Let karma whoring work for us!
But when talking about this on IRC with gwern and he thought it probably wouldn't do much good and isn't worth the effort to implement. What do fellow rationalist think might be a good way to encourage more quantity and quality in the wiki?
(as I noted in the buried thread)
The mental model being applied appears to be sculpting the community in the manner of sculpting marble with a hammer and chisel. Whereas how it'll work will be rather more like sculpting human flesh with a hammer and chisel. Giving rather a lot of side effects and not quite achieving the desired aims. Sculpting online communities really doesn't work very well. But, geeks keep assuming the social world is simple, even when they've been in it for years.
The recent implementation of a -5 karma penalty for replying to comments that are at -3 or below has clearly met with some disagreement and controversy.
How about we wait a couple weeks to try the new feature; instead of jumping up in outrage and proposing even more complicated schemes?
I'd be in favor of an official "no complaining about feature X for the first two weeks" rule, after which a post could be created for discussion. Like that the discussion could be about what actually happened, and not about what people imagine might happen.
It's not as if two weeks of using an experimental feature was some unbearable burden.
It's not a huge burden, but we're already seeing some negative effects worth discussing. For example I have now twice paid the 5 karma penalty replying to downvoted comments which were not trolling at all; they were downvoted because people disagreed with what they were proposing.
If we decide to wait two weeks, we need to decide on specific criteria that we will judge in two weeks' time to decide whether to modify or remove the new feature. If the new feature stays anyway because a few people decide unilaterally, then we might as well discuss it now.
Proposed solution: remove the karma penalty and do exactly the same thing we were doing before. That is, if someone is pretty sure that they will not benefit from reading the replies to a particular thread, they don't read them. No disincentives from posting such a reply needed. What is the problem with that system?
Edit: As of this edit, if one more person decides they don't like my comment, then no one can tell me why they don't like my comment without loosing 5 karma. One of many reasons the new system is terrible.
Consider the prior art, here. The first place I saw the "reply to any negatively-scored comment inherits the parent's score" concept was at SensibleErection. That policy has been in place there longer than LW has been a forum (possibly longer than reddit), so it seems to work for them.
Prior art for special "oldschool/karma-proven" sections: hacker news. Paul Graham is intensely interested in keeping a high-quality forum going, and is very willing to experiment. Here's the normal frontpage, here's the oldschool view, and here's the recent members view. Hacker news also has several threshholds for voting privileges.
One more step HN took is hiding comment scores, while continuing to sort comments from highest to lowest. It's dramatic, almost draconian, but it definitely had an effect on the karma tournament system
I think we forgot to hold off on proposing solutions.
I'm not proposing a solution. I'm thinking about the problem for five minutes.
edit: Well, it didn't even take five minutes!
We need a reliable predictor of troll-nature. I mean, I'm not even sure that P( troll comment | at -3 ) is above, say, 0.25 - much less anywhere high enough to be comfortable with a -5 penalty.
Of course, I'd be comfortable with asserting that P( noise comment | at -3 ) is pretty high, like 0.6 or something. Still not high enough to justify a penalty, in my opinion, but high enough that I can see how another's opinion might be that it justifies a penalty. If that is the case, well, the discussion is being severely negatively impacted by conflating noise and trolling.
I might go and figure out how to get some data off of LessWrong commenting system, to try and determine a good indicator for troll-nature. (I don't plan to try and figure out noise-nature. That's the problem that the Internet has faced for the last 15 years, I'm not that hubristic.) That in turn would would put some numbers into this discussion. I don't know that arguing over how many genuine comments can be inadvertently caught in a filter is any better than arguing over whether there should be a filter at all, but to my mind it's more constructive.
The proposals I have (not all of which are mutually exclusive) are :
Make comments within highly downvoted subthreads not appear on recent comments. Since the main problem with trolling is drowning out of recent comments, this will solve many of the issues. Moreover, it will discourage continued replies.
Have a separate section of the website where threads can be moved to or have a link to continue to. This section would have its own recent changes section. Moderators could move threads there or make it so that replies went to that section, and would be used for subthreads that are fairly downvoted. This has the advantage of quarantining the worst threads. This is a variation of an old system used at The Panda's Thumb which works well for that website.
Use the -5 penalty system but adjust either the trigger level or the penalty size. It isn't obvious that -3 and -5 are the best values for such a system if it is a good idea. The fact is that -3 isn't that negative as comment scores go, so something like -3 can be obtained without saying that much about a comment's quality. -5 and -5 or or -5 and -1 may be better values. The second would offer softer discouragement for more severe comments.
Use the penalty system but have karma penalties be restored if the comment it replies to is subsequently voted up more. This may be slightly more technically advanced, but this will help encourage people to speak out when they think a comment is unfairly downvoted.
Downvoted for putting more than one suggestion in a single comment.
Punish me for this anti-social act if you must, but as one of the dudes who tries to act after reading these suggestions (and tries hard to discount his own opinion and be guided by the community) this practice makes it much harder for me to judge community support for ideas. Does your comment having a score of 10 suggest 2.5 points per suggestion? ~10 points per suggestion? 15 points each for 3 of your suggestions and --35 for one of them (and which one is the -35?)?
Can we please adopt a community norm of atomicity in suggestions?