Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

This report reminds me of some of Jeffrey Lin (Riot Lyte)'s work on League of Legends. He found that 1% of players are toxic, frequently acting badly. But that only accounts for 5% of toxic behavior; 77% of bad behavior comes from people who are just having a bad day. That echoes this report's finding that 34 Wikipedia folks are responsible for 9% of abuse.

A related finding is Riot found that toxicity in LoL was contagious; people who played with an abusive player were more likely to be abusive in their next game. The Wikipedia phenomenon of abuse pile-ons seems similar.

Some links for Lin's work: http://www.nature.com/news/can-a-video-game-company-tame-tox... https://www.youtube.com/watch?v=nbYQ0AVVBGU



One of the interesting things about high functioning forums (like this one) is that there's a group of people who have as part of their job function the ability to enforce civility and reduce abuse. Like any community, HN also has "power users", in WP they're the run of the mill editors, but the difference is that here dang or somebody will show up to calm down incivility and point to the guidelines. In WP those people either don't exist or they're absent in providing that function.

I contribute here, I don't to WP and this kind of active community hygiene is exactly the reason why.


It has a lot to do with the community governance structure.

Wikipedia is something akin to a democracy. Every editor has the potential to become an Administrator or a Bureaucrat. This means that a lot of editors engage in metrics-gaming behavior and become involved in the Wikipedia community, which by its very nature means weighing on on various controversies, like whether or not to delete an article. There's nominally a focus on creating a consensus, but in practice, the opinion of the Right People wins the day. There's a process for everything, and everyone has a refuge in some guideline or another which leads to lots of rules-lawyering. It's difficult to get rid of problematic people who are good at rules-lawyering and garnering support.

On HN, dang and sctb are kings. If you cause problems, you are warned. If you continue to cause problems, you are banned. There are no higher appeals. Your interpretation of the guidelines is irrelevant; that of dang and sctb is irrefragable.

I think it is possible to construct well-run, friendly 'democratic'-type communities on the Internet, but very difficult. In practice, it's always seemed to me that dictatorial moderation creates the best communities - of course, there are communities with bad administrators, but they generally die out. I've seen some really excellent forums where moderators ban posters simply for not having sufficiently high-quality content, and the result was amazing.

Granted, for any reasonably active community, it means that someone is going to spend a lot of time doing the moderating. I think HN strikes a good balance with allowing users to downvote and flag, even though I feel that most people generally downvote for reasons I disagree with (disagreement.)

One other key factor I've found in community civility actually seems a little strange at first. In my experience, it is very important for no communication backchannels to exist - for example, no IRC channels. Those effectively breed the creation of cabals and all sorts of strategic interventions in discussions, or in the case of Wikipedia, talk pages and votes. That's been a major problem on Wikipedia. It is nominally community-run, but "community" really refers to the insiders who frequent the right IRC channels and are friends with the right people - connections you have no way of seeing simply by looking at Wikipedia talk pages. On top of that, people who are friends with the right people outside of Wikipedia always, simply by human nature, get much more leeway when engaging in abusive behavior than do people who aren't.


The much bigger difference is that Wikipedia conversations/disputes are long-running and relate to the state of permanent artifacts.

The technology of Hacker News makes it literally impossible to carry on a long-running conversation or collaborate on serious permanent artifacts. There’s no point in going back to last week’s discussion and vandalizing it with troll comments, because nobody is ever going to read it. There’s no need to invent complicated rules for productive collaboration, because there’s no work to collaborate on in the first place.


But the thing is, at the end of the day Wikipedia is still the best website on the web. It's amazing. It's continually accurate, inaccuracies get fixed very quickly because just about every article is someone's pet project, etc.

Yeah there are a few things - like some political articles - that you should be careful with, but everything has citations. You can see the historical edits, you can see the talk page, it's all very transparent.

Whether it's fun and nice to contribute to is irrelevant to me. To me it's a fantastic source of information. And so whatever they're doing, they must be doing right.


Many open-source projects and standards bodies operate the same way. Many's the time I've attempted to make a contribution and discovered that first one has to get on the inside of a particular clique, else be ignored.

It's rare for me to persist in such circumstances; life is too short to play other people's games. I usually abandon the contribution, or maintain a fork if it was of utility to me.


You've really nailed Wikipedia. That describes my experiences exactly, and it's a big part of why I left.


The civility and commitment to keeping discussion rational on HN is kept alive by the goodwill of its contributors rather than enforced by rules. When a power user shows up to nip incivility in the bud, they rarely have to employ more force than to remind the poster of the above ethos.

Rules and moderators alone cannot enforce civility. A common ethos such as the one which miraculously survives here on HN is also a prerequisite.

Rules can easily be subverted. They are a two-edged blade which can be forced to operate in an opposite manner to their original purpose. For example, the abusive users on Wikipedia often know all the rules, no matter how obscure, and they silence and confound their targets with such legalisms.

I have participated in forums where the first demand of hostile invaders was for a set of rules against abuse. Those forums where rules were provided quickly became toxic. Those forums where the moderators did not provide a rule-book to be subverted, but simply addressed abuse whenever it arose thrived.


Backing up what you said here, I've played a few online games with (at times) abusive communities (DotA, Overwatch), and in my experience:

- Games go for 30-60 minutes on average, and generally have a clear winner by the halfway point. Players are forced to sit through a defeat, building frustration.

- Games rely on team coordination on top of individual skill, so players who are scoring high (kills, etc...) may not actually be helping complete the objective.

- On death, there is enough downtime to vent frustration at your team, and blame them for the bad situation that caused your death.

- People often try to get the rest of their team to side against an individual, calling them out as the cause of other players dying / losing objectives.

It can be extremely toxic (to the point I've moved on), but I found pushing back with civility and calling people out often helped the situation.

Whenever I noticed it starting, I'd jump in and tell the frustrated player to take charge. I'd acknowledge the valid issues they call out, and try to reword them constructively and focus on the team rather than the individual. (eg: "Don't run in until we're all there to support you, we'll follow you now and see if we can outnumber them", or: "Try moving over there, their sniper is really good, and is protecting the area you were just in")


>- On death, there is enough downtime to vent frustration at your team

Now that you mention it, dead time in LoL (not sure what the player lingo for it is, "black and white" time?) really is a special timeslot for flaming. The angriest point for a player is also the time they can most easily communicate. If communication ability was uniform throughout the game, there wouldn't be flareups. When players are in action, they have only half the will to type/communicate. Suddenly that dampener is removed.


>but the difference is that here dang or somebody will show up to calm down incivility and point to the guidelines

Commonly, guidelines will be referenced, but the guidelines are very numerous and sometimes there are 'unspoken' rules backed up with 'a decision had to be made'. For example, blocking an article from being editing, and then marking for deletion, so you can't add content to it. Badmins will argue 'its not in the guidelines', so obviously they can do what they deem right.

We really need better wiki software so people can make their own.


This comes up all the time, but I think it just comes down to that HN is not a democracy, does not aspire to be one, and as far as I can tell, it never has had that aspiration.

I think the best metaphor for HN is that Paul Graham has invited all of us to his house for drinks and discussion. There are not a lot of limits on that discussion, but sure, if you are boisterous enough, someone will ask you to leave. There isn't due process and you don't have any rights, because we're just guests at a party. The rules are not applied by the admins as a judge might interpret statutes. Rather, they are there as helpful guidelines for the socially clueless as to how guests should behave.

I feel approaching HN in this way makes it very enjoyable. Personally, I have found the discussion at this gathering to be top notch.

Of course, it's not for everyone. If you would prefer discussion board as mini-nation state (another excellent model, in my opinion), I think reddit does a good job with that.


The really toxic people on HN are just sent down the memory hole though. I'm not saying that's a bad thing.


I believe this to be an example of the well known cognitive phenomenon that psychologists refer to as Monkey See, Monkey Do.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: