Predicting Hollywood's Blockbusters
from the good-luck-with-that dept
A professor from Oklahoma State University says he’s developed a computer program that can predict if a movie will be a box-office hit before it’s even been made. The program judges films on seven criteria, including its rating, genre, cast, special effects, whether it’s a sequel, on how many screens it will open and its competition on release, then puts it into one of seven categories ranging from flop to blockbuster, depending on how much money it’s predicted to make. We’re skeptical of the system, and not just because it’s accurate only about 35% of the time, or because it doesn’t take a movie’s actual quality into consideration. If the professor wants to predict the future based on past events, so can we. Two years ago we wrote about a similar computer for pop music, which, judging by the amount of crap still on the radio, hasn’t worked out. It’s hard to see the movie computer doing any better, even though one studio is reportedly in talks to develop it further.
Comments on “Predicting Hollywood's Blockbusters”
the full quote
The actual quote, from the article is:
So 75% of the time the model will predict the boxoffice results within $60 mil, and 37% of the time the results will be within $20 mil. Sounds pretty useful to me for ballparking purposes.
Re: the full quote
Nope,sounds like a bunch of crap.
Re: the full quote
nope, sounds like BS. Believe me, I work in the industry. You have to take into consideration, number of theaters its opening, number of screens, how much press its getting, the hype, the story, who directed, what style they directed and so on. his program is BS
You know where this could be useful? If it delivers the message that derivative crap will fail at the box office, and one studio head believes the message and instructs his employees to work on better, more original material. Even if the program is a piece of crap, it could still prove beneficial if it forces Hollywood out of one or more of its bad habits.
Unfortunately, for the computer, the whims of movie goers change constantly. What may be a hit this month may bomb next month in a different location, or even at the same one.
Our superficial culture of the day cannot be programmed to predict any sort of organized reponses, other than society is not very predictable.
One way to test it would be to try a whole lot of awful movies and see what it makes of them.
On the other hand, why not a movie about a movie designed by the movie-success predictor? “According to this program, this movie should be a smash hit!” Then the program can be the protagonist (or the antagonist) of every sequel! How many people were lured by the program into making bad movies, deliberately by the military-industrial-entertainment complex?