Is there a Hacker News flagging mafia?

[UPDATE: I normally put updates at the end, but this one is important to know up-front: the data point that this post is based on might be wrong. Adjust your skept-o-meter accordingly.]

Today in my Quora feed I found the best explanation ever of why is it hard to estimate a schedule for software development projects. I decided to submit it to Hacker News just on the very remote chance that no one had beaten me to it. Unsurprisingly, it had already been submitted a day earlier. But what was surprising was that the submission was dead. It had four points and no up-arrow. HN is not explicit about why posts are killed, but apparently this one had been flagged to death as "inappropriate content".

Now, if there was ever a piece of content on the web that was appropriate for Hacker News, this was it. It is pure gold, destined to become a classic among software developers, a fifth level warding-off charm against pointy-haired bosses. Someone should have written this ten -- no twenty -- years ago, but better late than never.

And it is dead on Hacker News, the premiere gathering spot for entrepreneurial software engineers on the web.

I fear that there is a Hacker News flagging mafia. This has been a problem on Reddit for a long time now. There are people out there who blindly downvote every new submission. Why? Because Reddit has so many eyeballs that getting an article on its front page has tremendous value, so the competition for upvotes is fierce. But there is a positive feedback effect that makes early votes count much more than later ones: articles that get a lot of early upvotes are much more likely to attract people's attention -- and their upvotes. Articles that get a few early downvotes are much more likely to just sink into obscurity. Because of Reddit's (and HN's) time-based ranking algorithm (newer articles rank higher than older ones, all else being equal) simply keeping an item off the radar for a couple of hours is usually enough to kill it.

Reddit introduced a countermeasure to this some time ago where scores are not shown for an item until it is one hour old. The hope was that by hiding the score for that amount of time, enough votes would accumulate so that the score, once revealed, would be more or less free of early-vote bias. The problem is that the number of votes that even a popular article gets in the first hour is fairly small, and it doesn't take a very large group of people acting in concert to heavily bias the results. Hence the rise of downvote mafias, groups of people who downvote everything except articles that they or their friends (or their customers) submit.

HN doesn't have downvotes, but it has flags, which makes it even more attractive to game the system. Running a flagging mafia is even easier than running a downvote mafia, because if you can get an article to accumulate enough flags then you have killed it once and for all. It is not even possible to resubmit the same link to give it another chance (on Reddit you can).

This is an inherent problem in all collaborative filtering sites. Once a site has a large enough audience, it becomes economically attractive to try to game the system.

Sounds like a good problem for a startup :-)

[UPDATE: If you liked this, please upvote it on HN before it gets flagged to death.]