Image courtesy of Castle Rock Entertainment

Can IMDb’s Ranking System Be Trusted?

The myth of democracy in The IMDb Top 20

The Internet Movie Database — or IMDb — is one of the most-visited websites in the world, (no. 50 overall, says Alexa). Hollywood relies on it for record-keeping; industry professionals use it to check on production details or to message various film and TV workers. For that, it’s truly useful. But IMDb isn’t merely industry-facing: movie and TV-watchers use it, too. There’s even a 10-star rating system that tells IMDb what your favorites are — data they use to generate their list The IMDb Top 250 Movies. With IMDb’s 65 million registered users and industry elite crawling its pages daily, this list would seem to be an important representation of their taste.

But it’s not. Dig deeper and you’ll discover that The IMDb Top 250 is completely and irrevocably rigged.

Both IMDb and The IMDb Top 250 have roots on early-internet USENET groups — initially called rec.arts.movies, beginning in the late ’80s as a series of lists of actors, filmmakers and trivia that would be passed around by film nerds. The first post, made by a British film enthusiast named Col Needham, was called “Those eyes” and was about actresses with beautiful eyes. Within these groups, the Top 250 list emerged — and its first best movie was crowned: Star Wars: A New Hope. Over the past 25 years, the list has evolved in real time to reflect its users’ overall aggregate change in taste.

As of writing, here are the top 15 and their respective ratings:

1. The Shawshank Redemption (1994) — 9.2
2.
The Godfather (1972) — 9.2
3.
The Godfather: Part II (1974) — 9.0
4.
The Dark Knight (2008) — 8.9
5.
Pulp Fiction (1994) — 8.9
6.
Schindler’s List (1993) — 8.9
7.
12 Angry Men (1957) — 8.9
8.
The Lord of the Rings: The Return of the King (2003) — 8.9
9.
The Good, the Bad and the Ugly (1966) — 8.9
10.
Fight Club (1999) — 8.8
11.
The Lord of the Rings: The Fellowship of the Ring (2001) — 8.8
12.
Star Wars: Episode V — The Empire Strikes Back (1980) — 8.7
13.
Forrest Gump (1994) — 8.7
14.
Inception (2010) — 8.7
15.
The Lord of the Rings: The Two Towers (2002) — 8.7

If you’re wondering why the top 15 of the IMDb Top 250 is missing oft-lauded classics like Casablanca, Dr. Strangelove, Citizen Kane, Psycho, Vertigo and 2001: A Space Odyssey — we only have IMDb to blame.

Their overall goal is to wring as much engagement as possible from each user, from time spent on the site to visits per day, to boost ad revenue. Thus, the secret IMDb algorithm is rigged to keep its most loyal users rewarded: your votes only count if you vote a lot. Meaning, you only get to rank a film once (on a scale of one to 10, though you can change your grade of a film at any time), and if you vote on enough movies, your grades are factored into the Top 250.

But IMDb doesn’t disclose how many votes you need in order to qualify as a “regular” voter. (The FAQ says, “To maintain the effectiveness of the Top 250 Rated Charts, we deliberately do not and will not disclose the criteria used for a person to be counted as a regular voter.”) In some ways, this maintains some integrity in the Top 250, as it’s much harder to create duplicate dummy accounts and get away with ballot-stuffing. But, it also means the votes of casual fans or people who are too busy to engage as much (including the busy industry professionals who have better things to do) aren’t included in the rankings, even if they make a few of them.

Like many other websites’ algorithms (anyone remember the failed Netflix algorithm challenge?) the company doesn’t want to let you know how the sausage gets made (and declined to comment for this article), creating mystery and mystique around the list and the brand. Thus, there are plenty of online threads (like here and here) dedicated to debating how the list works on a technical level.

What do we know about the inner mechanics of The IMDb 250? They call their algorithm “a true Bayesian estimate,” which is a complicated system that increases the importance of the average rating the more votes are cast. IMDb is, essentially, trying to keep new movies from having artificially high ratings.

That doesn’t mean they’re succeeding, which might explain why Deadpool currently has a higher ranking than Citizen Kane — no joke.

But the bigger problem might be the severely uneven breakdown of its voting core, which IMDB makes publicly available on the site. Consider the breakdown in votes for the current number-one film, The Shawshank Redemption:

via IMDb

As you can see, these voters — the ones that manage to break through the algorithm — are overwhelmingly male and between the ages of 18 and 29. It’s also interesting to note that the breakdown is mostly international, and these high-scoring movies seem like what you might be into if you were perhaps, a young person in South America or the Middle East who was trying hard to show that you had “good” American taste. Quora engineer Adrien Lucas Ecoffet noticed that, as well, explaining on a Quora thread: “There is a wide gap between the average American score and the average foreign score. What this suggests to me is that IMDB might be (correctly or incorrectly) detecting some ballot stuffing because of this and other undisclosed information although, unlike the Top 1000 voter thing, this is pure speculation.”

MTV film critic Amy Nicholson thinks that, even though the list has millions of voters, there is a dominant type — a college freshman trying to put on airs. “When I look at this chart,” Nicholson said, “I see a college kid trying to assemble a wall of DVDs he thinks will impress other people. There’s a lot of bluster in it. It’s the idea of liking a film to reflect who you are as a person. You wanna tell people that your favorite movie is [The] Shawshank Redemption because you think that makes you sound tough but sensitive and smart.”

It’s also the bro with tons of expendable time, she adds: “Who is the person who takes the time to go to IMDb, login, figure out how to vote? I’ve never voted on an IMDb film in my life. The person who does that is a person who really wants to feel like their opinion matters but maybe they haven’t been that exposed to movies. If you are fairly exposed to movies, you may be spending your time writing a book about it.”

IMDb voters see things differently. The list “represents diversity in movies which award shows like the Oscars, Golden Globes, BAFTA don’t have,” says 24-year old Khaled Kalache from Algeria — who votes enough to make it fair to assume that he counts as a “regular” IMDb voter.

He says these award shows only focus on “special movies they chose,” but, he argues the IMDb list includes films with a more global appeal (despite not necessarily being made by filmmakers outside of the U.S.). “The list is a door to cinema in general, if [you] want to go beyond popcorn movies and mainstream movies and watch movies like The Shawshank Redemption. If they love it, they go deeper [into] the list.”

The work of Christopher Nolan — a director popular with a young, male crowd — is the most debated filmmaker to crack the list. His film, The Dark Knight, ruffled feathers in 2008 when it overtook The Godfather as the No. 1 film of all time on The IMDb 250. In fact, almost every movie Christopher Nolan has ever directed is one of the 250 best films of all time — missing 1998’s Following (a micro budget project that never got wide distribution) and 2002’s Insomnia. Unsurprisingly, Nicholson thinks that Nolan is vastly overrepresented at the expense of other great films and filmmakers. “I think Christopher Nolan’s entire body of work might be on the top 250, but When Harry Met Sally isn’t? There’s no way that’s reflexive of the world. But I guess the person who would say he loves Christopher Nolan movies would probably never say he likes romantic comedies.”

Dibyayan Chakravorty, a 24-year-old assumed IMDb voter from Calcutta, disagrees: “I don’t think gender discrimination is an issue here. Movies about men — The Godfather, The Godfather: Part II, 12 Angry Men, Goodfellas — have better story, suspense, technique than movies about women, for example, The Iron Lady, Nights of Cabiria, La Strada.”

IMDb likes to promote the IMDb 250 as a corrective for those stuffy, invite-only awards. Because anyone can freely sign up for an IMDb account to vote — it’s therefore more democratic, right? Not exactly. Zero films in the top 20 of the IMDb 250 have a female protagonist — in fact, there are zero films directed by women in the entire list. “There’s this supposed idea of democracy, but it ends up being more white male than the Oscars,” says Nicholson. “If you want to talk about representation, there’s almost no female films in the list’s top 100. The most you get is something like [The] Silence of the Lambs, which is not a female picture. There’s nothing on that list with a female sensibility. No one’s gonna stick up for anything Reese Witherspoon’s ever done. You wanna pick movies that nobody will make fun of.”

Jonny Coleman is a Los Angeles-based writer. He recently reported on Tijuana’s cultural renaissance.

More entertainment on MEL:

For even more MEL, follow us on Medium or subscribe to the newsletter.