April 25, 2009

Movie statistics?

Over time, sports statistics hobbyists have become professional analysts employed by sports teams. Bill James started out analyzing baseball statistics in the 1970s as a boilerroom attendant and now he has a very nice job with the Boston Red Sox. The less mature field of basketball statistics has been losing its leading hobbyists over the last few years to NBA front offices.

With basketball statistics, it's been kind of like the early 1940s when the physics journals that had been full in the late 1930s of of exciting atomic discoveries suddenly had a shortage of new papers from Americans and Brits about splitting the atom. The Germans eventually realized that there was a very big reason for this sudden silence.

One of the main trends in baseball and basketball statistics has been away from analyzing existing box-score statistics to making up new statistics by analyzing video. For example, baseball statisticians have now been reviewing games to see which section of the field balls were hit too.

But let's think about other ways that quantitatively-inclined males can waste time.

For example, are there any movie statistics besides box office totals? The only use of movie statistics I can recall involved the lawsuit against Oliver Stone that was aided by legal novelist John Grisham over Stone's movie "Natural Born Killers," a vile film about a couple who conduct thrill kill murders of random people. A friend of Grisham's was randomly murdered by a couple who had just watched "Natural Born Killers" a whole bunch of times in a row while taking drugs.

"NBK" was made, right after "JFK", near the height of Stone's considerable powers of cinematic razzle-dazzle. To establish the amount of effort put into making "NBK" psychologically powerful, the plaintiff's lawyers counted up the extraordinary number of cuts in the film and compared it to an ordinary film.

So, number of cuts per minute would be an interesting statistic. Which directors or editors use the most, which use the fewest. Is there a sweet spot? Does it change over time for a director? Stone's "W." from last year has many fewer than "NBK."

That's just one thought, but you might have more.

My published articles are archived at iSteve.com -- Steve Sailer

44 comments:

Unknown said...

Stone footnote: A friend, a dedicated leftist and union man, worked on NBK. He said Stone was a monster who worked his crew to exhaustion and beyond. An example would be early am (like 2 am) resets after spending hours getting things ready for a shot He said he would never ever work for that SOB again.

Anonymous said...

Some of this work has been done. For instance Woody Allen is noted for his long takes.

Michael said...

They have that statistic already. It's called Average Shot Length (ASL) and is defined as the length of the movie divided by the number of takes in the movie. The Trivia section on a movie's IMDb page often lists it. A typical modern Hollywood film will have an ASL of 3-4 seconds. Cinemetrics is a website attempting to build a database of ASL's.

greenrivervalleyman said...

Adam Sandler Erdos number: the number of stars Roger Ebert awards a film is equal to the minimum collaborative distance between Adam Sandler and any one of its leads (i.e. if your movie is starring Rob Schneider, you are hosed).

Also, the number of Oscars received by any literary adaptation is proportional to the film's budget divided by the print-run of the original novel.

TCO said...

There are a lot of derivative ones from the box office. number of screens, size of screens, number of showings, duration, rev/day curve, etc. repeat views, audience demographics, etc. time of launch

seems like something could be made of time. rev/minute and the like.

therare probably stats related to star lead, whether you have one and the individual trends. although the star may be picking winners too...as well as making them.

genre is obviously useful for categorization.

on the cost side, in addition to total cost, also the breakdown in categories: acting, other labor, permitting, marketing, other?

I thought there was some expert ptogram in SUPERCRUNCHERS that could predict success of movies and even books with a regression off of a few categories.

Bill James said...

Sabermetrics for movies ... that would be a nice complement to your movie reviewing job. go for it!

agnostic said...

There are ratings from customers and critics, collected at IMDb and Rotten Tomatoes.

I've used the IMDb ratings at my personal blog to show that under-18 girls tend to have better tastes than older women. They're more gung-ho about Alfred Hitchcock, to pick just one example.

There's another use that I won't discuss, since I'm trying to write that into a journal article.

But maybe something like what I did over at GNXP, looking at the lay public's ratings (at IMDb) vs. the critics' ratings (at Rotten Tomatoes).

Another thing I've suspected is that comedies age very poorly, compared to, say, dramas or thrillers. Someone could look at IMDb ratings of a bunch of comedies from the 1970s and '80s, and then for dramas and thrillers form the same years, and see if the comedy-drama gap is bigger than today (by making the same measurements for movies made in this decade).

Or maybe take a bunch of comedies from 1970 to 2008 and plot their average rating over time, do the same for dramas, and see if the comedy one decays faster as you move into the past.

There seem to be a lot of subtle topical things in comedies -- little mannerisms or words whose context we just don't pick up on now -- whereas dramas and thrillers play to universals more, and so translate better across time.

The exception for comedies are slapstick ones like the Marx Brothers or gag-oriented ones like Woody Allen. These are more oriented toward universals than topicals.

JCS said...

I think imdb.com already tracks the number of times a movie is referenced by another movie or in some other medium; whether or not they perform this task accurately and objectively is another question. For example, say a TV show uses the phrase: "Here's looking at you kid." Casablanca would be referenced.

It would be interesting to know from such a metric things like: What types of films are most likely to be referenced? What types of films are most likely to reference other films? (I'd guess comedies) Do certain writers, producers, and directors tend to reference the same set of films (perhaps their own) or do they access the full spectrum? What are the historic trends for a given film? (For example, just how much did references to the Matrix series drop after the second and third films?)

I believe there's probably a bunch of grad students studying film who know the answers to these questions, but don't have the statistical inclination to write this stuff down in an easy to tabulate format like the Baseball Prospectus.

greenrivervalleyman said...

Actually, getting back to the inspiration for this thread, I think the next frontier/unexploited resource/gross market inefficiency in sports is sustained performance. A Hall of Fame-style career in professional sports (which is a strong causative factor in the success of any franchise lucky enough to secure the long-term services of such an athlete) requires more than just raw talent (a necessary but insufficient condition). It also requires the mental discipline to apply a strong work ethic to what is essentially a kid's game, and to do so year-in and year-out for at minimum a decade and long past the point where the marginal utility provided by the athlete's salary is exceeded by the opportunity cost of their free time.

Baseball is the worst culprit in this regard. To have a Hall of Fame-style career you must put in at minimum 15 years of duty and spend 9 months out of every year at work (regular season + spring training). The only part of the year you get off is during the holidays, when people are with their families and the sorts of fun of interest to young men are temporarily out of the picture. It's no wonder, then, that Jose Canseco once whined during his pennant-winning A's days (without the slightest hint of irony): "How come it's always us that have to go to the playoffs." J.D. Drew, another player with prodigious talent and little work ethic, could just not be motivated- when Cardinals' manager Tony Larussa once tried pointing out to Drew all the money he could be making if he just tried a little harder, all Larussa got was a shrug and the retort that Drew (of modest background) already had enough money to financially set him for life. And, of course, one wonders what kind of monster numbers Babe Ruth and Mickey Mantle could have added to their career totals if they had not lost several years to partying during their prime.

Yet obviously the worst problem is with young, grossly talented and even more grossly immature black athletes- the proverbial baby boys- living up to their potential. How does one turn a Randy Moss or a Keyshawn Johnson or a Terrell Ownes into a Jerry Rice, especially when money is no longer a lure? The team which can corner the answer to this riddle will dominate for at least a generation.

Deckin said...

The noted film scholar David Bordwell has made a career out of analyzing the length of scenes and of shots within a scene. He's about as quantitative a film scholar as there's ever been. You should look him up.

Usually Lurking said...

How about the number of words said in a movie? I am thinking about The French Connection which, IIRC, had fewer words per minute than just about any mainstream movie in the last 40 years.

Jorn said...

This is the quick slippery slope to the semantics of storytelling: how many edits, distribution of short vs long, closeups vs longshots, males vs females, heroes vs villains, exposition vs drama/suspense, violent events, violent emotions, nonviolent emotions... and you find yourself looking for a typology/ontology of emotions, and of story-plot-elements. I think it's fun, but rewards are very sparse.

josh said...

Steve you have written about "Game" a couple of times,and lots of those guys read your blog.If u can care about movie cuts,let alone sports statistics,to this degree,I can just imagine what wouldve happened had you dedicated yourself to the Venusian arts. You'd be like the love child of Style crossed with Tyler Durden!! But what a blog!

Anonymous said...

"But let's think about other ways that quantitatively-inclined males can waste time."

Can't talk, busy analyzing relationship between rank of coup leader and the viability/success/humanity of the ensuing junta.

I'm seeing negative correlation so far and have tentatively dubbed the phenomenon "The Rawlings Paradox" after Flight Lieutenant Jerry Rawlings of Ghana. As an African Big Man aficionado, you'll agree with me that you could do much worse than Jerry Rawlings.

Oh, and in case you were curious Aaron Neville has the highest notes per syllable ratio of any popular musician, it's insane, a full two standard deviations beyond runners up Boyz II Men.

"A friend of Grisham's was randomly murdered by a couple who had just watched "Natural Born Killers"

Fun fact: one of the couple is a distant cousin of mine.

jody said...

a good topic for boxofficemojo.com

as for world war 2, i always think the american chest thumping over developing the fission bomb is kind of silly. as if it was such a feat to develop it out there in the desert, across an ocean, and 3000 miles away from any war front. almost like some guy walking into the middle of bar fight where 1 guy is going toe to toe with 3 other guys, landing a few punches in a 4 on 1 fight, and then bragging about it for the next 50 years.

what the germans were doing was hard. maintaining that much industry while fighting on all fronts. their own fission bomb efforts were physically sabotaged by fins or swedes if i recall.

nothing the americans did in world war 2 was particularly difficult relatively speaking. of course this sends hardcore america first redneck types into keyboard pounding mode.

Anonymous said...

Who uses the most? Among current mainstream directors, I'd guess Stone, Meirelles, Luhrmann, Scorsese, and Spike Lee.

Who uses the fewest? Among recent American directors I'd say Gus van Sant in his later arthouse stuff (Elephant, Gerry). Among earlier foreign directors, there's Mizoguchi, of course, and Ophuls.

GMR said...

Well, there are all kinds of statistics that you could do besides box office (and I think the present value of the box office is more relevant for comparisons). # of tickets sold could also be used I guess.

Body count (could be split into hand to hand vs. gunshot I guess)

Boobies (# of, seconds on screen, etc. This was satirized in Knocked Up).

Goofs (already exists in IMDB, so tabulation by category)

Product Placements

Amount of footage added to directors cuts

# of times that movie is referenced by other movies

testing99 said...

There are plenty of stats Steve, many don't ever see the light of public day.

Contrary to your assertion that Hollywood is mostly middle class, most movies are made by and for the SWPL yuppie hipster class, lose money domestically, and only break even or make a small profit via foreign sales (a situation now changing according to the WSJ with the collapse of presales and foreign distribution under piracy and local flims).

Stats include gross receipts and profit margin per film for actors/actresses. Nicole Kidman, for example, does not come out well. Adam Sandler does, but of course movie industry insiders care more about internal politics-social dynamics than making money.

There's other stuff, such as foreign revenues per actor, Brad Pitt being a good example of an actor with no domestic appeal but considerable foreign appeal.

There's "top of mind awareness" for actors/actresses, the vaunted Q rating. The presales via the online box office services that Deadlinee Hollywood Daily often has. Indicating awareness of a movie before it opens, presales of tickets.

Captain Jack Aubrey said...

Boobies (# of, seconds on screen, etc. This was satirized in Knocked Up).Me, single, 14-years-old: fast forward past the plot points to get to the boobies. Pause.

Me, married, 32-years-old: Fast forward past the boobies so I can get on with the plot. Even when my wife's not around. I have a healthy sex drive. Last month the entire first season of "The Tudors" was available for free on Comcast On Demand in edited format. I didn't feel the least bit slighted.

As for length of shot, it seems that the real trend is for long continuous shots, or alternating between rapid cuts and continuous shots. This has been true at least since ER started doing it in the early 90s.

Anonymous said...

Steve, the idea you're talking about is called "featurization" in the machine learning community and is very powerful. Make a vector of quantitative features from anything you want -- movies, music, etc. -- and start correlating it with interesting response variables (like box office, or more interestingly with societal patterns & behaviors among the people who like such movies).


Along those lines, I'm surprised that no one here has mentioned one of the most obvious -- namely culture war stats. Take a look at the trailer for this movie:

www.areyouobsessed.com

That reality inversion was written, casted, and produced for explicitly political reasons.

I think statistics on the ethnicity of film makers and those of protagonists/antagonists would be very revealing.

Moreover, many movies can be rated as yea/nay votes on various issues that legislators were also working on (e.g. "Milk" is a vote in favor of any imaginable gay right, "Rendition" is a vote against torture, and so on).

That would allow a correlation of movie features with votes.

Another very interesting one would be to scrape Facebook or IMDB and correlate political attitudes with the list of movies tagged as favorites. You'd get pretty big datasets.

Eddy Elfenbein said...

The physicist Leonard Mlodinow has looked into the issue and concluded -- it's just plain luck.

http://articles.latimes.com/p/2006/jul/02/magazine/tm-random27

Anonymous said...

"nothing the americans did in world war 2 was particularly difficult relatively speaking. of course this sends hardcore america first redneck types into keyboard pounding mode."
B.S. Tell that to a Pacific-island hopping U.S. Marine veteran.

Anonymous said...

Apropos of boobies, as I understand it actress's contracts are extremely detailed when it comes to nudity: exactly what naughty bits will be shown and for how long, whether the shots will be close-up or distance, how brightly lit the scene will be, and so on. Nothing is left to chance.

Peter

Anonymous said...

their own fission bomb efforts were physically sabotaged by fins or swedes if i recall. - Jody

The sabotage was conducted by Norwegian resistance orchestrated by the British.

Mr. Anon said...

"The only use of movie statistics I can recall involved the lawsuit against Oliver Stone that was aided by legal novelist John Grisham over Stone's movie "Natural Born Killers," a vile film about a couple who conduct thrill kill murders of random people. A friend of Grisham's was randomly murdered by a couple who had just watched "Natural Born Killers" a whole bunch of times in a row while taking drugs."NBK" was made, right after "JFK", near the height of Stone's considerable powers of cinematic razzle-dazzle."

Natural Born Killers was indeed an arresting film - the kind that grabs and holds your attention. It was brilliant, and yes, as you said, vile.

In light of the law suit brought by John Grisham, I wonder why Michael Mann was never sued for by one of the victims of the Hollywood shootout. I find it hard to believe that the perpetrators of that bank robbery weren't at least partly inspired by "Heat".

Svigor said...

Not exactly OT, but probably my best chance to ask this here; why is every movie I watch anymore desaturated all to hell and back?

It's like 60 years on from The Wizard of Oz; where'd all the color go?

Anonymous said...

Apropos of boobies, as I understand it actress's contracts are extremely detailed when it comes to nudity: exactly what naughty bits will be shown and for how long, whether the shots will be close-up or distance, how brightly lit the scene will be, and so on. Nothing is left to chance.

Almost all Hollywood actresses nowadays use body doubles.

Anonymous said...

There was one scene in Natural Born Killers that made me despise Stone forever, and it slipped most people's minds. It had to do with feminism believe it or not.


A "dorky" man was in a diner, right before Woody Harrelson and Julianna Whatsherface shoot everybody therein. She was wearing a very provacative lack of clothing and walking down the diner aisle. The geek surreptitiously snuck a glance at her legs as she walked by without raising his head or making any above-the-waist eye contact and Stone not only filmed it, but put it in the center of the frame. She eventually shot him and somewhat nodded to him openly recognizing his "crime" against her, of sneaking a glance.

I found this to be enraging. A half dressed woman can expect men to steal some glances if half their ass is hanging out of a dress so short it barely covers the necessities. Its an act akin to wearing a thong in public........she can expect to be stared at. To say that men aren't supposed to look at that because its sexual harrassment would be like saying men could use socks to cover their genitalia instead of pants and not expect women to openly stare................of course they are going to.

Stone had this guy shot up on screen because he represented some of the repression of Julianna Whatsherface, and why he was part of her turning into a serial killer was "societies fault". I thought, "gimme a break" Stone, as much as he is known for whoremongering throughout Hollywood (and using call-girls for years if the rumors are correct, just like Keith Olberman and Bill Mahr do). m

TGGP said...

Speaking of coup detat stats, I put up data from Edward Luttwach's Practical Handbook here.

PSGInfinity said...

Jody,

Do some stats on the number of casualties/acre the Marines suffered in the Pacific. Then repeat that for Normandy. Now consider that the vast majority of the gear the Allies used was made in America. Now add the industrial effort required to build the diffusers required to separate and enrich the U235 need for fission. Simultaneously.

Remember, they lost: of course it looked harder. beating us was more than they could handle.

Curious said...

Can Testy write a post anymore without typing "SWPL"?

SFG said...

"Steve you have written about "Game" a couple of times,and lots of those guys read your blog.If u can care about movie cuts,let alone sports statistics,to this degree,I can just imagine what wouldve happened had you dedicated yourself to the Venusian arts. You'd be like the love child of Style crossed with Tyler Durden!! But what a blog!"

Common misconception. Geeks can't be as successful with women as our mastery of trivia would imply because it requires a different and diametrically opposed skill set. It's sort of like saying if a track star put that kind of effort into studying biology, he could invent a cure for cancer. To a first approximation, most of this information is explicit and quantifiable (number of cuts per minute, etc) whereas most of the information you need to seduce a woman is implicit and subjective (how she feels about you at any given moment).

All men are not created equal. Some of us got it, and some us don't.


"Stone had this guy shot up on screen because he represented some of the repression of Julianna Whatsherface, and why he was part of her turning into a serial killer was "societies fault". I thought, "gimme a break" Stone, as much as he is known for whoremongering throughout Hollywood (and using call-girls for years if the rumors are correct, just like Keith Olberman and Bill Mahr do)."

I'm not saying you're wrong about the feminist angle as far as Oliver Stone goes, but Roissy could explain this in terms of female disdain for betas as well.

Anonymous said...

Curious said...
Can Testy write a post anymore without typing "SWPL"?


Sure, but he'll just go back to ranting about the WASPy Harvard Elites and those snooty bitches who won't sleep with beta males. Choose your poison, Curious.

rob said...

Testing99,

You know what would make a great movie? An alliance of MUSLIMS and SWPLEs in Mexico puts a NUKE on a speedboat, boat it to NYC, and set of a NUKE in NYC.

Plus, it's so true to life!

Anonymous said...

Svigor said...

Not exactly OT, but probably my best chance to ask this here; why is every movie I watch anymore desaturated all to hell and back?

It's like 60 years on from The Wizard of Oz; where'd all the color go
?

That's because the more Hollywood tries to be "realistic" the more they decide to lower their production values. The great irony of it all is that as Hollywood has tried to be more realistic, it has lost touch with reality.

James Kabala said...

Actually, 1925 was the only year Ruth lost to partying. When he hit 40, though, his career did collapse pretty quickly.

testing99 said...

No.

Anonymous said...

I worry that the future will be devoid of males high on the autism spectrum making interesting quantitative observations. I fear that future generations will be lost to the number crunching and complex labarynthe of cause and effect for MMORPG mechanics.

rob said...

testing99 said...
No.
The shortest testing99 comment EVAR!

Fine, we'll make the movie better. The SWPLs will have tall, leggy, blonde girlfriends who will steal your job and never, ever, sleep with you.

Truth said...

"Anonymous testing99 said...

No.

Grasshopper, that is the most logical, most well-thought-out, interesting and effective post you have ever written. Kudos to your improvement.

Svigor said...

That's because the more Hollywood tries to be "realistic" the more they decide to lower their production values. The great irony of it all is that as Hollywood has tried to be more realistic, it has lost touch with reality.Hmm, my knee-jerk reaction was, "yeah, I could buy that if the reality Hollywood creates wasn't pretty much a direct inversion of actual reality wherever the latter needs some Tikkun Olam."

But maybe we're on to something here. Maybe there's a correlation. Say, the further Hollywood travels from the aforementioned unsatisfactory actual reality, the further it has to travel toward "realistic" production values by way of compensation? Ah, now I'm just talking out of my ass.

Either way, realism isn't on Hollywood's menu. Anyone who knows the history of any given period piece knows that (my personal favorite: Medieval people wore FANTASTICALLY BRIGHT clothing by modern standards. Full-on make-your-eyes-bleed dying was the norm, and muted colors were expensive and out of reach. Medievals (even the peasants) did not shop at the Gap).

Svigor said...

Blogger Truth said...

"Anonymous testing99 said...

No.

Grasshopper


How come you never call me "grasshopper"? With me it's always "sport," or "champ" or "my friend."

Racism?

albertosaurus said...

testing99 wrote
... most movies ... lose money domestically, and only break even or make a small profit via foreign sales ...No. It's much more interesting and complicated than that. Read Edward Jay Epstein's blog.

For example:
http://www.edwardjayepstein.com/Demyst3.htm

Anonymo said...

You know what would make a great movie? An alliance of MUSLIMS and SWPLEs in Mexico puts a NUKE on a speedboat, boat it to NYC, and set of a NUKE in NYC.

Plus, it's so true to life!

LOL. Thanks, Rob. Don't forget the dastardly role played by the HARVARD WASP MAFIA and its control of media, banking, and NYC law firms.