By Colin Brown
Those that think Twilight, The Green Hornet, The Beach, The Abyss andBoston Shuffler are a bunch of movie titles are only partially correct. They are all nicknames for just a few of the algorithms used on Wall Street to give firms those precious milliseconds of trading advantage. Some two thousand physicists and mathematicians work in the financial sector cooking up these computational black boxes – and a handful are now applying such predictive modeling and risk evaluation skills to the film industry in order to determine why some movies click and many others don’t. With worldwide spending on filmed entertainment climbing towards an annual $100 billion and beyond,
the pay-off is self-evident.
Cinema historians will point out that movies have always been subject to generic modification in order to maximize their appeal across multiple markets and cultures. Studio distribution executives will tell you that older-female-targeted films tend to “over-perform” in Germany and Australia, while Latin America has an outsized soft spot for family fare. And Hollywood tailors its films accordingly.
But such factory-belt fine-tuning is nothing compared to the scientific wizardry being applied today. Consider the following:
- Netflix, awarded $1 million to the team that created the collaborative filtering algorithm known as Pragmatic Chaos that predicts user ratings for films based on previous ratings. According to TED speaker Kevin Slavin, in his memorable talk on how algorithms are shaping our world, this one piece of code accounts for 60% of the movies that end up getting rented.
- As mentioned in a previous Filmonomics blog, Epagogix uses artificial neural networks to analyze screenplays with a view to determining their probability of commercial success. Meanwhile, professors at NYU Sternand Pennsylvania’s Wharton School claimed to have a devised
a reliable screening method for choosing movie scripts based on “textual information” – including the use of specific words.
While different in their methodologies, these algorithmic approaches share a fundamental trust in data analysis as the best filtering system. For those who think that the creative vetting process is best left to humans, there is a contrasting set of computational tools that might be broadly termed social recommendation engines. These rely on the curatorial instincts of one’s peer-groups. These influential tribes could be your friends on Facebook banging the drum for the film they just saw at the multiplex or film festival. Or they could be the membership network of 300 movie professionals whose votes determine those as-yet-unproduced screenplays that deserve to be included on The Black List.
Conceived by Franklin Leonard, VP of Creative Affairs at Will Smith’sOverbrook Entertainment, The Black List is perhaps the nearest we get to an instant snapshot of Hollywood’s collective taste buds. And it’s persuasive too: more than 125 past Black List scripts have ended up getting produced and being released theatrically, generating $11 billion between them. Successes include four of the last eight Oscar-winning screenplays.
Building on its hit-making potential, The Black List announced last year the launch of an online members’ community that will make algorithmic screenplay recommendations based on individual tastes. Users can now explore real-time updated lists of Hollywood’s most liked scripts. By offering
a blend of human insight and artificial intelligence, The Black List is affirming the need for both sets of tools in order to decode the DNA of filmmaking success. As we are learning, each inspires the other. It’s a no-brainer.
“We all fear the ‘future script’ spit out by a robot. But we can take comfort in the fact that the human brain is its own vast data storage program. A serious artist does his or her own rigorous study of successful works to lodge the common elements and patterns into their unconscious and inform their artistic output,” notes Jennine Lanouette, a story consultant who lectures at both Lucasfilm and Pixar (and will soon be contributing periodic blog posts on screenwriting analysis for Slated).
“As I see it, there is room for both – the organic artistic process and the set of objective measurements. The best application I can imagine of these computer prediction methods is in determining the size of the budget for a given project, and discovering niche market or sales territory potentials. The danger, in our winner-take-all economy, is everyone wanting to compete for the same gold ring – the billion dollar box office take. This is when art loses out. A computer might be able to maximize return on investment, but it will never create art. You need a human soul to do that.”
Colin Brown
Editorial Director of Slated
