A CNET blog post by Harrison Hoffman reported on the efforts by fans to elevate the new Batman film The Dark Night to the number one position on IMDB‘s list of the top 250 films of all time. Hoffman uses this and the Digg HD/DVD key incident (discussed by Clay Shirky in Here Comes Everybody) as examples of the failure of the wisdom of crowds.

I can sort-of buy the idea that the IMDB gaming incident represents a failure, though I would argue that it’s only a minor one. Having The Godfather drop to number two or even three on the list seems like evidence for a temporal mobbing weakness in the ranking implementation. Why wouldn’t time be considered? If The Godfather consistently received votes to rank in the top spot for ten years, then how could it be pushed off in weeks or even a few months by a newcomer? If I were designing a ranking system for films, I would certainly try to factor in time. You wouldn’t want a ten year top film to be invincible, of course, but I think there’s a balance to be found.

(As an aside, I was intrigued to find out that IMDB uses a Bayesian estimator to calculate its movie scores. At least they’re not simply tallying the total votes or scores for each film.)

I’ll speculate a bit — another aspect of the IMDB gaming incident is that the group involved was likely connected a little more closely than the rest of the voting population. In other words, it’s not really a case of the general crowd failing to make the right decision; instead, it’s a case of a small sub-group manipulating the results of a faulty voting system.

The Digg HD/DVD incident is rather more controversial, I think. I don’t view it as a failure of the wisdom of crowds. I can understand how it can be viewed as such, though. More thoughts later if I get to it…

And maybe “wisdom of mobs” is a better term for these degenerate cases.