Did Wonder Woman really cost too much for you to love it?
I was trying to get a group together to see Wonder Woman, but I ran into opposition I didn’t expect. First: a friend revealed his policy against movies on the weekend. Simultaneously, a boyfriend revealed his policy against big-budget Hollywood movies.
I wanted to prove them wrong wrong wrong.
This post is about being lasso-of-truthed by data. Uff.
Here’s the text thread from my pseudonymised boyfriend, Sigmund.
I won’t see any Hollywood movie with budget over $80 mil
Well actually I haven’t calibrated the figure
I think it lives somewhere between 50 and 100 mil
Wonder Woman was like 120 mil, so it’s a def no
Big budget films kept disappointing me, so I nixed them
Oh! I think I found my calibration point!
Arrival was 47 mil
So … any film above 30–40 mil
There are rare exceptions of course — Gravity, for example (100 mil)
So what we have is a probability function that goes from about 0 at 30 mil and climbs steeply so that it’s asymptoting to 1 at around 80 mil. 50% probability around 50 mil
Sigmund loves the films of Quentin Tarantino and Alfonso Cuarón. So I went off in search of other people who loved Reservoir Dogs, Pulp Fiction, Kill Bill (both Vol. 1 and 2), Inglourious Basterds, Django Unchained, Y Tu Mamá También, Children of Men, and Gravity.
The MovieLens data set has 78,037 users that rated at least one of these 9 films. I normalized each user’s ratings for all the films they had rated (not just these 9) so I knew who loved/hated/was relatively indifferent to Tarantino and Cuarón. I count someone as a proxy for Sigmund if they had rated at least 3 of the 9 films and loved almost all of them. I allowed them a single “ehn, about middle of the pack” but didn’t allow them to dislike any. There were 6,049 of these people who had given 2,641,491 ratings to 20,018 different movies.
For the contrast set, I chose people who didn’t like the Tarantino/Cuarón — I included folks who had just rated one of the nine and didn’t like it. For people who had seen multiple Tarantino/Cuarón films, I let them have one that they were “ehn” about but they couldn’t love even a single one. There were 17,095 of these folks who gave 2,334,676 ratings to 16,482 movies.
Then I grabbed movie budget info from The Numbers, leading me to our real consideration set of 3,475 movies that have both budget and Sigmund-like and unSiggy-like ratings. I also got IMDB movie ratings, which go all the way from movies of 1915 all the way up to Wonder Woman.
Okay, so is Sigmund right that there’s a 50–50 chance of him liking something with a budget of $50 million and basically no chance of liking something with a budget of over $80m? No. But kind of yes.
All the scores are normalized — in online reviews, there is a prominent skew towards high numbers of stars. So a “zero” here doesn’t mean zero stars and it doesn’t even really mean 3 stars. It’s more like “a bunch of people rated this right in the middle of all the various things they rated”. The MovieLens data lets me normalize per person, then I averaged this for each budget-bin. The IMDB ratings are normalized across all movie ratings, not per individual.
But for what it’s worth, Siggy, the budget for Wonder Woman was more like $150m, so you have the same probability of liking it as a $30m film.
If you like Tarantino and Cuarón what else might you like?
I also did the analysis above by normalizing production budgets for what was typical of the year of release — $30m today means something quite different than it did in, say, 1956. That year, the average movie budget was $4.3m and The Ten Commandments was a huge deal for having a $13.5m budget. But honestly, the results aren’t much different.
But let’s cruise through some more of the stuff that Sigmund look-alikes liked. We’ll start with the tiny budgeted movies. Some of these are very surprising to me. But others are really the usual suspects.
- The Horse Boy — Mongolia! Autism! Documentary!
- Big Parade — did you know Wonder Woman is ALSO about World War I (she fights to end all war; this 1925 movie was actually filmed pretty close to the events depicted in WW)
- The Usual Suspects — I can’t imagine you haven’t seen this? But maybe it’s overdue for a rewatch.
- Kevin Hart: Let Me Explain (not on Netflix, but let’s watch one of his other shows)
- Memento — Mnemosyne was the goddess of memory, a Titaness (that’s a word, it could also be Titanide), mother of the 9 muses, including your favorite one, Siggy: Terpsichore
- 12 Angry Men — Honestly, my favorite version of this is when Angela Lansbury ended up in a jury on Murder, She Wrote. I mean, sign me up for empaneling with Jessica Fletcher.
- Before Midnight: guess where they’ve been living for 9 years? GREECE. Just saying.
- Monty Python and the Holy Grail: it may be time to watch this again.
- Housebound: comedy horror. I’m tentatively in.
- Casablanca: Maybe not today, maybe not tomorrow, Siggy.
For small budgets, check out:
For normalishly budgeted movies, check out:
Interestingly, there are no “Big Budget” ones you really need to see. Maybe Raging Bull, The Shining, or The Prestige.
Finally, here are the ones Tarantino/Cuarón lovers also loved that were Huge Budget movies. Let’s look at them in terms of gender.
- The Godfather: Diane Keaton gets tenth billing
- Fight Club: Eugenie Bondurant gets eighth billing as “Weeping Woman”; a total of 3 women in the top ten billed actors
- The Godfather: Part II: Diane Keaton gets third billing, there’s one other woman in the top ten characters
- The Dark Knight: Maggie Gyllenhaal gets fifth billing, there’s one other woman in the top ten
- The Matrix: Carrie-Anne Moss gets third billing, there are two other woman in the top ten
- Apocalypse Now: Uh, guess how many women are credited in this (answer: zero)
- Modern Times: Paulette Goddard gets second billing, there are no other women in the top ten
By contrast, Wonder Woman has Gal Gadot in top billing, with three other women. That’s 40%, which does feel low for a movie about a character who grows up on an island with 0% men. Fwiw, three of the top five billed actors are women. I would normally report the Bechdel Test results, but these seem to have been presumptive instead of based on the film. Gosh, how could we get that data?
One other thing I was wrong about
You and Mnemosyne may remember that at outset there was one other problem I ran into, a friend who said:
I don’t like seeing movies on the weekends. Not social enough! I wanna talk, interact!
The MovieLens data gives timestamps but no information about what time zone reviewers are on. Nor do I know how close to watching a movie someone in the data writes the review.
HOWEVER, I can use a proxy and compare people writing reviews on Saturday compared to Wednesday. Because there are basically 500,000 reviews for each day, it’s easy to get to statistical significance. So the difference is tiny and probably meaningless, but yes, you could say that people are giving higher ratings…on weekdays compared to weekends. Merciful Minerva!