Movie Metrics: Is it Possible to Quantify the Experience of Film?

How can some number-crunching grad students really think this is possible? Analyzing the potential success of a screenplay with a computer?  Measuring brain reaction to find a winning story formula?

Some professors at Wharton have even proposed a metric for analyzing screenplays to predict the Return on Investment of a film(pdf here).   The method uses such criteria as vocabulary, genre, number of characters and sentence length.

“We compare the performance of our portfolio against two benchmarks. In the first benchmark, 30 movies are randomly chosen to form a portfolio. Each of the 81 movies in the holdout sample has equal probability of being selected in the portfolio. The ROI of this portfolio is recorded, and this procedure is repeated 1000 times. Then, the mean ROI based on random selection is calculated by averaging the ROI that results in 1000 portfolios. We find that the mean ROI in this case is -18.6%. In the second benchmark, we try to replicate more closely the way studios select movies by MPAA ratings.  Studios in general make roughly 60% R-rated movies and 40% non-R (G / PG / PG-13) rated movies
(MPAA 2004). Thus, we replicate this rule of thumb by fixing the number of R-rated movies to be made in the 30-movie portfolio case to be 18 and the number of non-R rated movies to be 12. Then we select 18 R-rated movies randomly from the pool of R-rated movies in the 81 holdout movies, and likewise for non-R rated movies. The mean ROI of a portfolio selected using this “MPAA-based” selection method is -24.4%. Our method outperforms both benchmarks by a significant margin. Indeed, as shown in Figure 6, our approach always produces a significant economic gain no matter how many movies are selected for the portfolio, suggesting that our model is able to capture determinants from the textual information in movie scripts and hence significantly improve the studio’s profitability.”

Their results are surprising, however I would guess that a thoughtful, experienced person could pick the most profitable scripts with more accuracy than this formula.

Analyzing the profitability of scripts based on comparable precedents is one thing – but trying to increase a movie’s impact by scanning the brain’s reaction to scenes is another.  Some well-intentioned brain specialists at MindSign have coined the term neurocinema.  My first reaction was that the whole concept is idiotic and deeply misunderstands the purpose and nature of Story.  In an NPR interview, Philip Carlsen of MindSign said:

Ideally, a director could send dailies out to us and we can scan the dailies. For those of you who don’t know, dailies are all the shots at the end of the day. They’ve got seven shots that look exactly the same, but we can give you the most activating shot of these seven, and you can use those. You can actually make your movie more activating based on subjects’ brains.

First of all, a story contains much, much more than a collection of images and scenarios.  Anyone who presumes to measure the impact of a story by measuring the impact of individual scenes is not a story-teller.  A story digs deep into universal human principles; it opens up the world of things unsaid and unseen – of intimation and connotation – and it affects us on different levels.  I don’t think this technique could even work on a popcorn flick, because the arc of a story is so much more than the sum of its parts.
Because this FMRI reveals individual brain responses to film clips, its most valuable purpose appears to be in determining what kinds of images excite the individual’s brain most.  I still have a problem with this, because an image or scenario is profoundly affected by what precedes it. For example, the image of a man running through a tunnel affects us differently if we sympathize with him or fear him. Is he chasing something or being chased? So this goes back to my previous point; a movie is more than a collection of moving images – it’s a story that is experienced as a whole.

I really hope that Dr. David Hubbard is related to L. Ron.