Sunday, October 15, 2006

The Movie Formula

This week's New Yorker had an interesting article by Malcolm Gladwell called The Formula. In it, he looks at the way data analysis is being used to predict things once thought pretty unpredictable: like movies and music. For instance, Mike McCready has come up with a way of finding music hits. He claims that after analyzing various factors of a song, he can then see it if the song falls into 1 of the 4 hit clusters he has discovered. He claims that hits from Beethoven to Gnarls Barkley will fall into 1 of the 4.

With movies the key seems to be on how much weight you put on the factors. An example of how this can work is illustrated by how some mathematicians outperformed the dog experts at Greyhound Races. It wasn't that the mathematicians discovered a new variable, it was that the dog experts weren't giving all the variables their proper weight. The experts knew about "race grade" (when a dog has been dropped or raised a class), but didn't realize just how important it was. This hinted at how a movie formula would work. It's not that anyone has to recreate the wheel, but perhaps the studios overrate some factors and underrate others.

All very interesting, but here was my only problem with the article. We never find out what the formula is. I guess this makes sense, the data analysis superstars want to make their money, but it was kind of disappointing.

It did make me wonder though. Is there a formula that predicts a total movie bomb? Here's my attempt. (Pauly Shore + Sinbad) * (number of sequels greater than 2) and then you square it all if it's NC-17. Which would be a really strange movie if you think about it.

Labels: ,

1 Comments:

Anonymous Anonymous said...

Chu,

I read the same article last night - I had to admit that it was a very interesting read. I enjoyed the whole Mr. Pink & Mr. Brown element it had going in it.

The problem I have with these models is that someone somewhere is always inventing the better mousetrap, yet when the rubber hits the road there's always some variable or other "kink" in the formula that they'll blame the faulty outcome on. I mean, if their movie predictor was really so hot, why wouldn't every producer in Hollywood come knocking on their door. Know why? Because they've probably heard this a thousand times.

Do you know how many fool-proof financial models have been created over the years???! How many of them actually work. The answer is none. If there were any true predictor of the future, there would be some guy (or gal) richer than Bill Gates, or at least on the Forbes richest 400 (and no one fitting this description was on the list this year).

Although it's a nice pipe dream and something to strive for, these things will always be just fools gold or another chance for a charlatan to bring forth their wares.

But the kicker was one thing - one of the main protagonists was getting all choked up over the movie Dear Frankie, which was a piece of a steaming turd of celluloid. He admitted crying to it, TWICE. Game over....

12:57 pm  

Post a Comment

<< Home