There was an intriguing piece today in the WSJ applying the principles of The Wisdom of Crowds to predicting the outcome of the 2008 presidential election. The logic of it was persuasive when it invoked Wikipedia, which I find to be far more useful and reliable than detractors claim (when people say it’s inaccurate, I want to know, Compared to what source of such breadth and depth?)
It was less persuasive in the preceding sentence, when it said,
This collective intelligence also accounts for why Google results,
determined by an algorithm reflecting the popularity of Web results
matching a search, are so relevant….
Today, wearing my vice president hat, I heard a presentation on new vistas of user-specific smart online advertising that the presenter described more than once in “Big Brother” terms — not as a bad thing, but in terms of Big Brother’s storied effectiveness and, I suppose, intrusiveness into private thinking patterns.
But you know what? So far, I’ve been hugely unimpressed by the effectiveness of software that is supposed to get to know me well enough that it can predict what I want. Take Netflix, for instance. I have freely given Netflix more than its share of info on my preferences. I have, for instance — and I’m embarrassed to admit this — rated 1,872 movies on the one-through-five-star system. Yes, that’s one thousand, eight hundred and seventy-two. Any time Netflix has said I need to “rate more movies” — and it seems to have an insatiable appetite in this regard — I have taken a few moments (in the evenings, of course) to oblige.
I have done this in a vain attempt to give Netflix enough info to at least make a wild guess as to what sort of movies I like. It still doesn’t seem any deeper or more intuitive than what a clerk at an ’80s-style video store might have guessed after less than a dozen rentals. Or so it seems to me.
For instance, Netflix is convinced I’ve got a fierce hankering to watch “Classics” — you know, movies with Clark Gable or Myrna Loy or whatever. Apparently, this is based on the fact that I’ve given high ratings to, for instance, “It Happened One Night” and “The Thin Man.” But of course I give those high ratings! Any literate movie fan would! That doesn’t mean I want to see them again, or that I want to see lesser films with the same actors in them! I don’t have a black-and-white jones here, people. I just acknowledge quality, and I think my judgments along those lines are fairly conventional, really. What I need you to do is extrapolate what I might like among films I haven’t seen or heard about…
Whatever. Anyway, this sort of software hasn’t figured me out, even when I’ve wanted it to. It’s more like somebody’s stupid second cousin than Big Brother.
Related Posts
My TiVo thinks I’m gay.
Does “rating” movies for Netflix mean “rating” in the classical sense? In other words, you rate movies that you have actually seen?
If so, what the heck?!?
You have rated 1,872 movies…meaning you have actualy watched that many? If so, Wow. I didn’t think I had a life, but compared to you, I’m the fricking president. I don’t think I’ve watched a tenth of that number…but then of course I don’t watch many movies anyway.
The other thing is had I watched that many, I certainly don’t think I’d have taken the time to rate many of them…certainly not all 1900. David
You’d be surprised. Once somebody starts running the titles by you (along with an image of the poster), you’ll realize you’ve seen a lot more than you think you have — either on TV or at the theater, or on video.
You do reach a certain saturation point… the last few times I’ve agreed to rate movies, I’ve gone through a couple of screens without finding anything I like, and get bored and quit. I may never crack 2,000…