The Self-Moderating Forum, or, Who Moderates The Moderators

OK, so let's imagine an, uh... hypothetical... set of forum software that, uh... hypothetically... let the community participate in this moderation process. Stop. Do not reply now. Read what I am about to write first, because HYPOTHETICALLY here is how it would work, in this fantasy imaginary software that clearly does not and cannot exist, because such a thing would be impossible:

1. "enough" trusted users flag a post as offensive. Could be 2. Could be 20. Depends on trust levels and forum threshold settings. Could be one, if that person is a formal mod with infinite trust.

2. The post in question is immediately hidden.

3. A private message is sent to the author of the post, describing what happened, in the friendliest imaginable language, and letting them know that a considered edit of any kind is enough to un-hide the post. It also lets them appeal this flag to a mod, if they want to.

Now some what ifs.

• The post author edits, the post is un-hidden and nothing else happens: success! No harm no foul, and notably no formal moderators had to be around for this to work!. The users who flagged gain a little trust, and the post author gains a tiny amount of trust. (There are other ways to gain trust outside this process altogether, even by just reading the forum a lot; I'll probably describe that in a later post.)
  
  
• The post author does not edit, does not appeal: the post is never un-hidden and that user loses a fairly sizable amount of trust. (if their trust level gets low enough, they lose posting privileges, at the lowest levels of trust, they will gently be turned away at the door of the forum altogether.)
  
  
• The post author appeals to a mod. The mod agrees that the flags were not correct. The flaggers lose a little trust. The post author gains a little trust.
  
  
• The post author appeals to a mod. The mod does not agree. The post author loses a LOT of trust. Like.. a lot a lot. The flaggers gain a little trust.
  
  
• The same post is hidden a second, third, fourth time through another round of flags. The trust levels amplify in each round, so by the third time a post is hidden, a simple "do nothing" by the post author might be enough to have the system gently turn them away from the forum.

The goal here is for the community to be able to protect itself from the worst users, even without a moderator present. But it works even better with a moderator, as the moderator can accelerate the process by handling the appeals, which amplifies the speed at which trust is gained or lost.

Or you could disable it altogether, and go back to a world where every raised flag must be manually handled by an official mod, if that's how you roll. But this'll be on by default, so moderators who start a forum and walk away forever are not an inevitable death sentence to the community. Cough. You know who you are.

What the hell are you doing??

Sanctioned spambotting?

Well, I'm asking your opinion on this process: would it work? Is it crazy? Is there something else I'm not thinking of, but should be?

Sorry, was that not clear?

And I guess the other assumption is: how much do you trust your fellow forum users? Clearly you have to trust the mods to some degree, but should ALL trust be concentrated in the mods? That did not work so well for Q23.

Wumpus is crowdsourcing a business venture yet again.

Speaking of self-moderating...

I just thought it was interesting that when the concept of moderation came up, there was (apparently) no way to conceptualize forum moderation outside of formally appointed moderators. Is it too radical to think of a forum that has strong community moderation aspects, that work in tandem with the moderators -- when and if they happen to be around?

See, that's the thing: forums always scale with the size of the community. Moderators generally don't.

Imagine a forum that worked even if the moderators are woefully overworked or understaffed, even if the moderators are not present at all, even if the community no longer trusted its moderator(s). I think that's possible under the system I described.

It's not too radical. It's just stupid & poorly thought out.

Moderation inherently requires the judgement of individuals. Individual humans. And your idea of crowdsourcing moderation, like almost all attempts at crowdsourcing, is absurd.

The answer to the problem you're trying to solve is simple: the moderator hires more moderators. That's how moderators scale. It's a solution that's been around forever, and there's a reason why it's been around forever: it works. It only fails to work when you get a jackass control freak like Mr. Chick who refuses to acknowledge his limitations. And frankly, there's no helping someone like Mr. Chick.

For all the moderators that aren't Mr. Chick, your proposed software solution is completely unnecessary.

How so? A given post that gets flagged, then PM, then edited, is reacting to the judgment of at least three "individual humans".

And if the flag is appealed, then the judgment of a fourth person is involved, a moderator -- who is also an "individual human".

Yeah... you've thrown in all these stupid uneccesary steps.

You know what else works? The moderator for that forum reads the offending post and decides to act (or not act) based on his or her judgement.

You can divide up the moderation responsibilities in all sorts of different ways; that's the domain of management theory. But your proposed software adds nothing valuable to the system. Nothing.

I beg to differ. It adds the following:

• a system of moderation that scales alongside the community as it grows
  
  
• the formal moderators do not have to be present, either through absenteeism or overwork or neglect, for community moderation to be effective
  
  
• the formal moderators do not even have to be competent for community moderation to be effective
  
  
• if the formal moderators are around to handle flag appeals, their effort is amplified through the community flags, making them more effective.

And if you don't like it, you can effectively turn it off (though it will be on by default) -- thus reverting to the old world of moderator(s) handling a flag queue.

Anyway, moderators need at least one flag to even know to look at a problem in the first place, right? So flagging clearly needs to be encouraged as a net positive for the forum. Leveraging the flags you already get seems sensible to me.

I mean, this is slashdot et al, just with the extra 'editing the post unhides it'. Given that people will be flagging posts without providing a reason to the offender, is this a significantly more helpful design than just going with Slashdot's "well you fucked up this time and nobody, try to post better in the future" approach?

I guess I could have elaborated, but there are a few subtypes of flags you can cast, no more than 5, probably 4. These are also configurable, if you want a HERP DERP flag with a description of "herpy derpy" you could do that. But out of the box we pick some sane defaults for common badnesses, a McDonalds value menu of bad forum behavior, if you like.

The one in the above example is "content a reasonable person would consider offensive" but it doesn't really matter.

To your point, you would definitely get a bit of flag-specific information versus "a flag was cast!"

My God, it's full of stars!

To be clear, 100% open source in this case.. so crowdsourcing "with benefits".

That will be explained in the fourth thread.

It'll all be clear in Volume 3, I swear!

But seriously -- does anyone have any reasonable critiques of this system I've proposed? I want honest feedback, I already have "your system is stupid for unspecified reasons!" and I'd like specifics if possible.

In the absense of a mod as you've described, what would keep a group of posters from ganging up on a member who's not well liked and hiding even his non-controversial posts to the point he ends up essentially banned? Would that be a bug or a feature?

That's immediately what sprang to mind when I read it. If anything, it sounds like it would encourage factions.

Also, wumpus, the way you're soliciting feedback (Three threads? Really?) is kind of obnoxious. If your proposed system were in place here, you'd probably be hide-rated into oblivion by now.

Three threads is world-ending supernova? OK, I'll go ahead and remove the other two topics based on your feedback.

And, Done.

Thanks. I apologize for being obnoxious, that was not my intent, I honestly thought it was better to have two topics there since the original text changed so much.

Well, the assumption is that you have a moderator around even in desultory fashion at least some of the time, as a checks-and-balances thing.

Speaking as a programmer, I'd say that repeated flags from the same people would not count as much. For a person to be truly toxic, they need to be flagged by a wide variety of trusted users, not just the same 5 over and over.

As an extreme example. At the point where someone has had their post content hidden by flags from 50% of the trusted users on a forum at some point, I think we can say with reasonable statistical accuracy that person is kind of a jerk.

How soon until the brett-ocalypse

The main issue is that the content will suffer from the tyranny of the majority, but I guess that's 'by design'.

WTF HOW DO YOU COME UP WITH IDEAS SO BAD?

Seriously though, there are two massive problems with that: 1) as mentioned FACTION WARFARE OPEN PVP ZONE 2) it's not even tyranny of the majority, it's tyranny of the vocal minority. Your idea is sort of like giving the keys to the nuclear launch codes to the Tea Party.

Not to mention that by design it seems to give mods more to do instead of less, since they'll probably end up spending most of their time dealing with appeals.

I like the self-moderating system on Stack Overflow. Presumably wumpus has heard of that. ;)

I think for a site like Stack Overflow, where it's Question/Answer, and the questions are usually pretty straightforward, the system works very well.

I think it can work, but the problem you're going to have is calibrating the amount of trust it takes to hide a post. To much, and you're still relying on moderators. To little and you're in the dystopia that the other people assume you'll fall into.

Also, it's within the realm of possibility that you could game your system given a few user accounts working together, in an attempt to raise the trust of all accounts.

And to be clear with this:

I assume that's only the negative trust impact would amplify in each round? You shouldn't reward people who can't get it right after retry #1 with a amplified positive trust.

If there's no moderator, the system is broken, because the part that makes it work is the "appeals process" where a moderator gets to determine whether or not a previous crowd-mod decision was correct or not.

If there is a moderator, the system basically comes down to "people flag posts, moderator reviews" with only uncontroversial auto-spam handled without moderator intervention.

So: Will it work? Yes, given a good community and a good mod. Will it work without those things? No.

Also, systems that require input more complex than "vote up/down" from users, (eg slashdot, which has the same voting system as you propose) are really not that much more informative for offenders.

The data we have from Stack Exchange usage says otherwise; users are quite willing to flag when there are severe problems, and those are by far the clearest and least ambiguous flags to cast. This system is even better, since the automatic resolution of those flags requires zero moderator intervention -- the user edits the post, it becomes less offensive or inflammatory, and everyone carries on as before. With slightly higher trust levels.

Yes, systems of opinion are very very different than systems of science and fact. That's why this is an interesting and complementary problem space. Also, sometimes you really do want to just hang around and shoot the shit and post LOLcats.

True. But I have some experience with designing these kinds of games, since I designed Stack Overflow.

Yeah, only negative parts would amplify, as you are dealing with repeat "offenders".

True of voting and I agree completely.. nobody needs "brofist" and "hug" and "like" and "dislike" and "omg" and "lol" to choose when voting. But flagging is different, if a policeman writes you up for jaywalking that is a very different thing than a policeman writing you up for, say, grand theft auto.. or murder. Flagging works along a continuum and offering people a "value menu" of common pitfalls and problems to flag with is pretty straightforward.

But what is the difference between "flagging" and "voting down"?

If I get this stuff correctly, very quickly every forum under this system would develop a "character" and this would then be used to remove dissent. Unless the forum is VERY balanced along dividing lines, you'll get the majority voting to supress "one or two troublemakers", and that would be the start of a very pretty but destructive landslide. In BF, for example, we would have started by agreeing on the perfectly reasonable step of banning Octonoo, Dawn Falcon, and Vetarnias, but quickly we'd progress to getting rid of stupid but not offensive dissenters like Brett, and this would continue on, with greater homogeneity leading to a heightened awareness of dissent, leading to action against broader categories of people. The same would be true of the terrible terrible human beings that dislike Dark Souls. and so on and so on.


so thanks, but no thanks. Just do what everybody else does, and get the more or less reasonable lifers to help out (HI Elyscape)

Dark Souls aint no Crusader Kings 2.

I vote to stake Astromarine

As opposed to leaving him around and turning every discussion about him? There's something to be said for getting a little more homogenous.

DO YOU SEE?