Tuesday, January 12, 2016

The Predictinator Got It Wrong

What is The VFX Predictinator? Start here.

Last year, we were wrong.

For the first time since we created The VFX Predictinator, a single formula my wife and I created that correctly predicted the winner of the visual effects Academy Award from 1989 to 2013, our formula got it wrong. The algorithm predicted that “Guardians of the Galaxy” would take home the Oscar for visual effects, based on a number of quantifiable data points including its strong box office, overwhelming critical acclaim, and its inclusion of talking creatures as its primary visual effects. Sadly (at least for our formula as well as our egos), “Interstellar” took home last year’s visual effects Oscar.

As I wrote before last year’s Oscars, I was deeply concerned about the accuracy of the Predictinator’s prediction.  One could hear the anxiety in my voice, as I appeared on The VFX Show podcast to talk about our Oscar guesses.  Leading up to the actual ‘running of the numbers’, my gut said that “Interstellar” would be the front runner. My wife even said, just moments after she entered the numbers into the spreadsheet, “Well, I guess this time The Predictinator will be wrong.” She was right about that.

Last year's Predictinator results. Full article.

In the months since the Oscar telecast, we’ve recovered from the initial disappointment of getting it wrong, then looked back at the two films and tried to break down exactly why the Academy membership voted the way it did.

We built the formula to demystify the core values of the Academy voter. We realized a predominant and consistent value is what we define as the “prestige” factor. Academy members tend to vote for films directed by experienced filmmakers; inventive, beautiful movies that are regarded as “important”. Academy voters want to feel smart and forward-thinking, not just simply rewarding popular, popcorn films.

Critics saw “Interstellar” as flawed but beautiful.  It featured the much-loved Matthew McConaughey and Anne Hathaway (both Oscar winners) and a respected director, Christopher Nolan.  The visual effects were (arguably) not necessarily groundbreaking, but the gorgeous imagery visual effects created by the effects team was one of the reasons the film was highly regarded.  In the abstract, the visual effects in Nolan’s film also had a heavy ‘practical’ emphasis, pushed by the film’s marketing, which ingratiates itself to the Academy’s largely older voter base. The film was certainly considered the most prestigious movie nominated for the Visual Effects Oscar in 2014.

We represent this “prestige” factor in a few ways in The Predictinator: the Rotten Tomatoes score which measures critical acclaim, total number of Academy Award nominations, and Actor Prestige (if the lead actor has previously won an Oscar). By our numbers, “Interstellar” had a lot going for it, but didn’t earn it enough points to beat “Guardians"’s final score. Its Tomatometer value, while a respectable 72%, was the lowest in the category with crowd-pleasing superhits scoring higher. The film earned five Oscar nominations, which are hugely important to the Predictinator score. Even its star, McConaughey, had an Oscar under his belt (Actor Prestige).

The Academy has also historically favored movies with creatures--especially creatures that talk--in the visual effects category.  The Predictinator accounts for that, and in this area Nolan’s film suffered.  On the other hand, the Academy rarely looks kindly on any sort of sequel; a sequel is almost directly at odds with "prestige", since they are routinely looked upon as derivative, unoriginal, and cash-grabby.  In this case, “Interstellar” benefited. 

As you can see, “Interstellar” had the qualities the Academy wants in a winner, but our formula didn’t give it enough points to win.

We talked about everything on “Interstellar”s side; now let’s discuss why “Guardians” could not win the Oscar, even though The Predictinator gave it a higher score.

At its core, “Guardians” is a crowd pleasing, funny film. The Academy historically ignores comedies. Very rarely do comedies earn Oscars nominations, and even more rare is a comedy win (modern exceptions include Kevin Kline’s Best Supporting Actor win for “A Fish Called Wanda”, and Marisa Tomei's Best Supporting Actress win for "My Cousin Vinny"). The only comedy to win a visual effects Oscar is “Death Becomes Her” which, if you remember, was extremely difficult for the Predictinator to correctly predict. “Guardians” was tied for the highest Tomatometer rating of the year, but that score indicates overwhelming, general positive enthusiasm for the film, not “prestige”. To that point, “Guardians” only had one other Oscar nomination (for makeup), which hurt its Predictinator score.

“Guardians” is also part of a film series, a part of the Marvel Cinematic Universe. Although it’s not strictly a sequel, the film operates thematically and structurally within the confines of an overwhelmingly popular film franchise, and is branded as such. The Marvel name is as prominent as a “2” in the marketing of the film. As we’ve seen, Academy voters regularly avoid rewarding sequels with Oscars; we did not score the film as a sequel, but had we, it would have been dinged ½ point, giving “Interstellar” more of an advantage.

Finally, “Guardians” is regarded as another superhero movie (we will not quibble about the superhero powers--or lack thereof--of the protagonists of the film), which the Academy is also shy of rewarding. The Academy, largely made up of folks older than 60 years old (hey, look, the visual effects branch is the youngest branch, by far!), regards superhero films as childish fare, quite the opposite of important films that deserve to be lauded. Rarely do superhero films like “Batman”, “Superman” or “Iron Man” win Oscars (the lone exception since 1989 is “Spider-Man 2”). Academy voters are less likely reward a genre of film that regularly reboots and restarts itself (for example, Keaton, Kilmer, Clooney, Bale, Affleck... that’s an average of a new Batman every 5.5 years).

All this said, the Predictinator latched onto some key data points: “Guardians”’ extravagant box office haul ($333M), super high Tomatometer score (91%), and, most importantly, got 1.75 points because its primary visual effects consisted of organic creatures that talked. Groot and Rocket Raccoon were extremely impressive and main characters in the film, which usually is a big plus to Academy voters. But in this case “Guardians” was simply not prestigious enough to have this work get noticed.  

There was simply no way that the Academy would award such a film in the Oscars, but our Predictinator had no way to account for the factors mentioned above.

Understanding why “Interstellar” was always going to win, that “Guardians” was always going to go home empty handed, and with all of the potential changes occurring within the visual effects industry, my wife and I asked ourselves the fundamental questions:

Were we missing some key part of the formula all along? Should we alter the formula in some way? Has a changed occurred in the way the Academy votes for visual effects films? Should we just leave the formula as-is, and accept that it could be wrong in certain years? Or just give up and retire the whole stupid thing?

Stay tuned. Here is Part II, "Making The Predicinator Right"

No comments: