Trusting Users Too Much

An anonymous reader writes to alert us to an article at Forever Geek on sites that trust users too much and the users who game them. From the article: "Trusting users is a good thing. But implicitly trusting users is no good. If Digg has moderators who approve a story before it goes live on the front page, shouldn't they have moderators checking spam reports? Social sites give so much power and emphasis on users yet a handful still have the power to wreck these sites. Until these issues are properly addressed, social sites will continue to be gamed."

It happens.

[nametaken]

Digg exists so that people can easily tell each other what they want to hear. Sometimes it's cool, sometimes it's bogus. You can't have your cake and eat it too.

Ummm....wikipedia?

[catbutt]

Maybe they could add a reputation system!

Digg Reliability

[klenwell]

I ignore the home page and just check out the day's most popular page once or twice a day:

http://digg.com/view/technology/popular/today [digg.com]

It's not the front page of the NY Times and it's no doubt influenced by the much lamented front-page gaming, but I still usually find one or two interesting things that I hadn't heard about yet.

Evidence is all over the place but hard to prove.

[twitter]

Snotman wants evidence of social gaming. Evidence of past behavior abounds. First, let's reiterate the potential problem. From the Fine article:

Register them on Digg. Have them randomly digg 5 stories a day. Then scrape the top 100 users on Digg, and add them randomly across the 100 fake users. Simmer for a week or 3, and then *bam* - start reporting any story dugg by the top 100 users as inaccurate.

The ease of creating a botnet of Windoze machines eliminates all evidence. Instances of actually catching a company hired troll like Barkto [essential.org] are rare. Even obvious astroturf, like the M$ PR created Apple Switcher [slashdot.org] are hard to detect. If that's not proof enough for you that some dishonest companies are abusing the net for there advantage, I'm not sure what is. Oh yeah, you can look up the court proven public disinformation campaign against DRDOS by M$.

As for evidence of gaming sites like Slashdot, visit these these losers one day. [anti-slash.org] Use a text browser and a condom or you might walk off with more than you want.

Yes, it's pathetic but people do that kind of thing.

Re:It happens.

[finiteSet]

I tried Digg for a while but ultimately gave up - and I think it is because they trust the user too much. Though the average Slashdot discussion is well short of, say, an academic journal, even the worst Slashdot discussion I've read was better than the best Digg one. I'm not trying to troll: I'll explain. Allowing everybody to moderate every post of every discussion, in my experience, results in a discussion that reflects the views of the majority by silencing the views of the minority. You may disagree, but I find that Slashdot moderators put more thought into how they shape the discussion - I know that I will mark a post insightful if it shows insight, regardless of whether I agree. Furthermore, leaving the majority of Slashdot posts remain unmoderated allows more room for both sides of the debate to be heard. I know, I know, proper tweaking of settings on both Digg or Slashdot can reduce some of these problems, but in the balance between trusting the user with too much control over the conversation and too little, I think Slashdot is a lot closer to optimal.

Re:It happens.

[telbij]

You may disagree, but I find that Slashdot moderators put more thought into how they shape the discussion - I know that I will mark a post insightful if it shows insight, regardless of whether I agree.

I agree with this. I'm always astonished at people who talk about how shitty /. discussions are and then talk about how they've switched to Digg. The /. system is far from perfect--moderators can't always be experts--but by and large there is much more opportunity for modding up of cogent arguments that go against popular opinion. Of course you still have the issues with earlier posts getting more moderated and so forth, but I don't see how anyone can even for a second say that Digg discussion is anywhere near the level of /. Digg's strength is in volume of stories.

That said, I think the best communities are smaller and more focused. Any large and general community is gonna be pretty ignorant about any given issue. If I'm really interested in a topic I don't want moderators to ham-fistedly attempt to pull out the knowledgeable posts. I'd rather go to a focused community where everyone has something relevant to say.

Re:It happens.

[fredrated]

I think it ramains an interesting experiment.

Any given topic is going to be of interest to a subset of the Digg community. Within that subset, if the users were eevenly split on a topic, and equally likely to digg, undigg or ignore a given comment, a comments diggs would be a random walk centered at 1 (the submitters digg).

Five or more undiggs and a comment doesn't show, at the default cut off. At that point a comment is less likely to be viewed by the community, only those that expand the comment.

This would seem to give power to the early posters, if a group acted in cencert to undigg posts they didn't like. If the desire is for the exposed posts to reflect the communities position,

On the one hand you probably want to view the 'voice of the community', assuming you respect the communities position generally, but on the other hand, 5 undiggs can be reached perty quickly, and for the post to be cut off from general ispection that soon can result in a small group with inordinant fitering power, or just a generally scewed picture of the communities position by random walk.

Anyway the gist of this rant is that if a post that has been dugg down doesn't repeatedly get dugg down over page views, it should eventually raise back into view, so more people will decide to move it up or down.

online communities will continue to increase in complexity, and I think the digg model can continue to be extended to produce interesting results.