Wednesday, January 13, 2016

Making The Predictinator Right

What is The VFX Predictinator? Start here.

In the last article, we discussed how The Predictinator got it wrong. Before deciding if we should continue predicting the Oscar winning visual effects film with our current formula, alter it, or (gasp!) abandon it, we examined two questions that might help us make a decision:

• Has there been a slow change in how visual effects films are perceived by the Academy, causing members to vote differently?
• Are we missing something that has been in front of our eyes the whole time?

First, let’s take a look at the potential changes that have occurred over the last decade.


WHAT HAS CHANGED?
In the early days of digital visual effects, only a few filmmakers were confident and comfortable in helming multi-million dollar visual effects blockbusters. Oscar winning visual effects films were dominated by directors like Robert Zemeckis, Peter Jackson, Steven Spielberg, James Cameron and the Wachowski siblings, largely making wondrous, successful crowd-pleasing blockbuster visual effects films.

But as time went on, complicated visual effects became more accessible to a larger pool of filmmakers at greatly reduced costs. In the modern era, visual effects became a tool of art-house, ‘important’ and ‘prestigious’ filmmakers like Martin Scorsese, Ang Lee, and Alfonso Cuaron, filmmakers who regularly make Oscar-caliber films.

Incidentally, all three of those directors have directed Oscar-winning visual effects films within the last decade. And six out of the last seven visual effects Oscar winners were also nominated for Best Picture.

With breakthroughs created for films like “The Abyss”, “Terminator 2” and “Jurassic Park”, audiences (and Academy voters) were amazed by the fidelity of digital creatures, something audiences had literally never seen before.  In fact, the reason The Predictinator’s data set begins in 1989 is because that was the year “The Abyss” was released, which ushered in the new era of digital effects.

The next two decades saw films that had a heavy presence of digital creatures win Oscars. Year after year, films featuring Gollum, Davy Jones, King Kong, a talking pig, and talking polar bears were winning Academy Awards. This is why our formula gives so much weight to two criteria: organic creatures and organic creatures that talk. However, the full Academy seems to be valuing these breakthrough digital creatures less each year, particularly noticeable in the disregard of the two recent “Planet of the Apes” films featuring astoundingly realistic all-digital apes. While audiences still make blockbusters out of films with digital creatures, the Academy seems less likely to award a visual effects Oscar to a film solely for the presence of its quality digital characters.

Audiences now expect Gollum-quality digital characters in their films, and the Academy may no longer reward a film solely for hitting this quality standard. The bar for the “wow” factor has been raised. In fact, four out of the last five Oscar winning visual effects films had no significant digital characters.


WAS THERE SOMETHING ESSENTIAL MISSING?
Our formula couldn’t differentiate between “Interstellar” (a prestige-leaning ‘classy’ Oscar film) and the well-reviewed, popular crowd-pleasing comic book film “Guardians”. Were there characteristics of “Interstellar” that could be quantified to boost its point value? Were there essential characteristics of a film like “Guardians” that could be quantified to reduce its point value? And would these be universal truths, correctly impacting past and future contests?

Like we mentioned in the previous article, “Interstellar” had a respected, auteur director (could we quantify that, somehow?). The film was touted as a heavy ‘practical effects’ film (is there a way to quantify miniatures and physical sets?). Are we hurting ourselves by using the Tomatometer as a gauge for critical acclaim, since crowd-pleasing popcorn films can garner higher Tomatometer scores than prestige pictures like “Interstellar”?

As for “Guardians”, it had a strong comedic bent (how on earth does one assign a point value to humor? I mean, even the esteemed Golden Globes doesn’t have any idea how to classify a comedy. Oh, look, “The Martian” just won the Golden Globe for Best Comedy!). “Guardians” is part of the Marvel Cinematic Universe, but isn’t a sequel (so do we just ding it for being a Marvel film? That seems too specific).

WHAT DO WE DO?
We were stuck.

We were not comfortable with abandoning our little formula, just because it broke once. We would not let it die; we’re too proud of it. And we were not ready to accept that it would occasionally be wrong. That’s no fun at all.

We could have fractured the Predictinator into two distinct time periods, breaking it up into two separate formulas, to account for our perceived changes to the industry we outlined above. That option seemed inelegant; our goal was a single, polished formula to rule them all. We could have tinkered with the criteria and how each piece of data is weighted against one another, but, even with the slightest tweak, we quickly realized the entire formula would fall down like a house of cards. This thing is pretty well-calibrated, and can’t withstand much tinkering without destroying itself.

The only option remaining, we thought, was to add a new magic piece of criteria that would solve our problem, without breaking the rest of the formula. And we wanted to add something that had a universal truth, not a statistical hack. For example, we weren’t going to add the criteria “Does the title of the film include the word ‘Guardians’? Minus one point!”

So we thought about it. For nine months. (Hey, we were busy!)


THE “I’VE GOT IT” MOMENT
One day, my wife was in the kitchen, I was in the living room. She just blurted out, “I’ve got it.”

And it was staring us in the face every time we looked at the full spreadsheet of films from 1989-2014. She had figured out a new piece of easily quantifiable criteria that would simultaneously de-emphasize digital characters as well as bolster the power of a ‘prestige’ film.

“Is the film based on a comic book? Minus one point.”

Duh. It seems so obvious now. Historically, and for the foreseeable future, Academy members simply do not give Oscars to films based on comic books. Peering into their headspace, I suspect they consider these films to be pure popcorn and not worthy of the honor of receiving an Academy Award. We posit that subtracting a Predictinator point for films “based on a comic book” will not only slightly de-emphasize digital characters, but will strengthen point values of prestigious films that frequently win visual effects Academy Awards.

My wife executed a simple masterstroke that not only fixed the “Interstellar” year, but also bolstered the historical data. Since 1989, only one film based on a comic book has won the Oscar. “Spider-Man 2” is the lone victor, with 12 comic book films going home empty handed. Even our most difficult year, 1992, was made easier with this tweak. Instead of “Death Becomes Her” squeaking past “Batman Returns” by a narrow margin, it now crushes the caped crusader with the new criteria. (By the way, the single comic book winner, “Spider-Man 2”, still commands its victory over its next contender, “Harry Potter 3”.)

Here is what the data looks like. Again, the only thing that we changed to the formula is a single line item, “Is the film based on a comic book?” If so, one point gets deducted. No other criteria values were changed. “Interstellar” now wins the Academy Award for visual effects. Presenting The VFX Predictinator 2.0.

(((RT Score/ Sum of all noms' RT Score) X 5)^2) + (BO (millions)/ BO Total of all noms) + (Academy Noms (only if 4 or more) X .25) + (((Month of Release / Total Month of Release) X 2.5)^2)* + (Sequel = -.5) + (Prior Sequel won Oscar = -1) + (Primary FX organic creatures = 1) + (Primary organic creatures include facial acting = .75) + (Lead Actor an Academy Award Winner = 1) + (Film based on comic book = x(-1)) = Final VFX Predictinator Score

*value has an upper limit of 1




Looking toward the future, the frequency of comic book films is increasing. Marvel films are a cinematic juggernaut- the twelve Marvel movies have grossed over $9 billion worldwide, and every movie studio is building their own universes, flooding the future movie market. The vast majority of these films are aimed at young audiences, consistently hitting explosive levels of risk-averse, family-friendly PG-13 mayhem. These films will continue to earn visual effects Academy Award nominations, because the best visual effects facilities in the world are creating outrageous, spectacles and otherworldly characters for them. The films will continue to perform well at the visual effects bake-offs, and will continue to earn Oscar nominations.

Prestige pictures, however, they are not, and I think the Academy will continue to resist rewarding comic book films.

We’re going to stick with this, The Predictinator 2.0, for a while. Well, at least for this year.

The nominations for the 88th Academy Awards will be announced Thursday, January 14, 2016. We will run the numbers through The VFX Predictinator 2.0 sometime after.



6 comments:

  1. My fellow scientists, I look forward to your results!

    ReplyDelete
  2. An elegant choice for a more civilized age.

    ReplyDelete
  3. I think two more factors should also be taken into consideration including whether heavy use of physical practical effects and miniatures were used and if movie had large portions shot in larger format (Imax). Dark Knight was heavily considered for vfx nom for these two factors. I think this might also be factors pushing Star Wars for an oscar win over Mad Max because of these two factors.

    ReplyDelete
  4. Darkmoon3d - how would you quantify 'physical practical effectxs and miniatures' use in a film? Also, "Dark Knight" didn't win the VFX Oscar.

    ReplyDelete
  5. Not sure how but best way would be if the film production mentions using it heavily. For example, Interstellar, Mad Max and Star Wars have said this. If effect sequence have option of being created CG or physically and physically is used instead then this would qualify. Concerning Dark Knight I meant that it received a vfx nomination not a win. VES gave Nolan a Visionary Award (http://www.hollywoodreporter.com/news/christopher-nolan-honored-visual-effects-46643) evidently due to his work on Dark Knight and Inception with use of physical effects.

    Didn't realize but does seem hard to quantify whether a film used physical effects as a majority over use of CG.

    ReplyDelete