30 Nov Film Fund-amentals: Beyond Magic

The fastest way to get the old guard in Hollywood steamed is to utter the phrase: “predictive model.” It works like magic as a room full of suits turns into a pack of Popeyes shooting steam out of their ears. Then for real fireworks, just add the word “database.” Suddenly, the whole place looks like Guy Fawkes Night.

The wild spectacle would be fun to watch if it weren’t so pathetic.

As often is the case with such profound hostility, the root cause is extreme ignorance. Few people in the movie industry have a clue about predictive models. That is one of the reasons why they always refer to the Arthur de Vany study, Hollywood Economics: How Extreme Uncertainty Shapes the Film Industry. Like a clove of garlic waved before a vampire, this book is used as the ultimate defense. Too bad it’s irrelevant to what I am discussing. De Vany does a fine job analyzing why you cannot predict success. But that isn’t the point in predictive modeling (I mean real predictive modeling, not the medicine show variety). The main point in predictive modeling is attempting to determine the basic parameters of what you will end up spending and, perhaps, loosing. Gee, if the dang thing is a hit, you will just have to learn to live with it. But it doesn’t matter to the modeling process.

What the process does is create a systematic guide for financial management. Yes, I just used the other dreadful phrase: “financial management.” Goodness, I’m starting to sound like Martin Luther touring the Vatican. Next thing you know, I’ll start yapping about “accountability.”

So let’s move on to more heresies (I’m starting to like this game). A traditional formula for producing a movie is to combine an A-list cast with a hot director and a so-called top script. This is a pretty standard way of raising lots of money in Hollywood and thereby allows the producer to take a swell vacation with his mistress (this snotty comment is courtesy of Sam Fuller). Maybe it will work, maybe it won’t. But these elements are only half-relevant to the predictive model system.

For the past several years, I’ve been involved with several other people in creating such a predictive model. It’s not designed to predict success. I cannot emphasis that enough. It is meant to create a financial outline that will give  filmmakers a guide to what they can expect to achieve at the basic, lowest level (which to be honest is the level most of them will be working at). In other words, your basic and probable financial returns from the investment placed into the movie. This system is not a Magic Eight Ball. It is a complex statistical system that will provide an outline of the movie’s most basic financial possibilities. While working with the system, we have even discovered some interesting things.

For starters, you can forget about movie stars. They sell magazine covers, not movies. In most cases, the real difference in box office between an A cast and a lesser cast is marginal. Most major stars can only justify their huge paychecks because the studio is willing to pay it, not because they earned it. Once movie stars are quantified into the model, they end up making very little difference. The rest is merely a fairy tale.

In the predictive model, the script is extremely important in terms of its quantifiable components. These elements include such things as genre (which is the main element partially determined by modern trends) as well as character type and the number of major characters in the script. These components actually break down into several distinct versions based on an analysis of the main characters’ complexity and the narrative direction. These elements can be quantified, though it is one of the most difficult parts of the model.

Directors are important in terms of these elements. Especially their experience in location management, character complexity, genre (though this element can float in strange ways with directors), as well as the actual number of past films completed and the size of the previous film’s budgets and running time (oddly enough, this is a critical element). Likewise, the producer is run through a similar process (along with the editor and the cinematographer).

Then we crunch the numbers. Actually, the process is better described as stomping, whacking, bouncing and generally shaking the whole set through multiple layers of algorithms that allow us to analyze various match-ups of the chosen components. It is from this process that we can then derive a bracket that presents the base financial prospects for the film. Please note the term “base,” as in “basic.” We are looking for the minimal basic amount the movie can make if it gets completed and receives some form of distribution.

So far, our test runs have achieved 95 percent accuracy. Since I also adhere to Sam Fuller’s comment about film being a battlefield, I expect the ratio to drop in the heat of real application. Most likely it will average out somewhere around 75 to 85 per cent, which is still pretty good.

Notice that not once have I said anything about “success.” It’s like balancing your checkbook. You do so to see what you’re dealing with in your bank account. You don’t balance it in hopes of finding a million bucks you didn’t know about. If you did find a million dollars, well gee, ain’t that something. But you don’t run your account based on that notion. The same is true with a predictive model for film financing.

But this is the way Hollywood tries to run its own business. It’s all a warped combination of magical beliefs, false assumptions, misbegotten expenses, and – if it works – lots of dumb luck. No wonder they hate the predictive model. After all, we’re rude enough to attempt something that faintly resembles science at a witchcraft convention. Well, excuuuuse me!

So I understand the hostility. I even predicted the hostility. I don’t really care about that part. I once had a job conducting health care surveys among old folks down in Sarasota, Florida. They are ten times meaner than anyone in Hollywood (heck, I was the first person at the company who could do the job and not have an emotional breakdown. Oh sure, there were a few complainants filed against me by these folks, but I wasn’t the one having the breakdown).

However, I’m tired of this “you can’t predict success” crap. So let me make this quite clear: That isn’t the point of the process. Success is an elusive thing, but the rest is manageable.