Earlier this week, The Athletic released a list of the Top 100 sports movies of all time as voted on by their staff. The list was a fun look at which films their writers like and generated some good arguments over which movies were under or overrated.
If you follow me on Twitter, first of all, you’ve made a terrible mistake. But also, you’ve probably seen me messing around with horror movie data. When I saw the list at The Athletic, I felt compelled to make my own attempt at ranking the best sports movies using the same approaches I use when working with other genres.
So while The Athletic’s list represents which movies a group of sportswriters think is best, this list is an attempt to present a consensus list of the top 99 sports movies according to the internet. Why 99? Because I like multiples of three and because it fits nicely in the format I use to visualize lists as you’ll see below.
Generating the list
Like anything that’s popular, the internet provides some data we can use to try to understand it better. For movies, the Internet Movie Database (IMDb) provides public data sets that include average ratings from their community of users. Depending on your perspective, you might have varying opinions of those ratings. If you’re a serious movie critic, you probably think the ratings are mostly junk because they represent the opinions of anyone with an internet connection and not just film buffs. But I would argue the opposite. I’m more interested in which movies the average movie watcher likes and for that purpose, a sampling of opinions from millions of movie fans is a powerful measure.
IMDb also does us the favor of listing genres for each movie in their extensive database. I’ve had good success using these genre identifiers for other projects but this is where things started to go haywire on this one. IMDb only lists a maximum of three genres for each movie and while it does have a “Sport” genre, it isn’t used as often as other genres and appears to be further down the priority list. Hence, as an example, Mighty Ducks isn’t classified as a sports movie, which is obviously nonsense.
That meant I had to seek out another source. MovieLens is an open source movie rating project that publishes large data sets for anyone to explore. One of the things they offer is an extensive system of tags. The tags are more granular than genres and are related to the content of the film. For example, a sports movie wouldn’t just be tagged as “sports” but also the specific sport involved.
In addition to the list of tags, MovieLens provides a relevancy score between 0 and 1 for each tag for each movie. So a movie like Back to the Future II that has a small baseball related portion of its plot would get a lower baseball relevancy score than A League of Their Own, which is centered around a baseball team.
The most subjective part of this process was determining how to use these tags. For reference, the list of sports related tags I explored were “sports”, “baseball”, “football”, “boxing”, “soccer”, “racing”, “olympics”, and “bowling.” Since this is a hockey blog, I’ll note that “hockey” does not exist in the data as a separate tag from “sports.”
Without getting more granular we have to this point, I set the relevancy minimum at 0.7 meaning that a movie had to have a score of at least 0.7 in one of the tags listed above to be considered. I also had to eliminate the “racing” and “olympics” tags because they pulled in too many non-sports movies. For example, Munich from 2005 is not really a sports movie although it’s closely related to the Olympics. And unless we think the Fast and the Furious franchise count as sports movies, the “racing” tag had to go as well. Probably the biggest loss in this decision was Talladega Nights, which is I think is a sports movie. But it only scored a 0.47 relevancy score on “sports” and without the “racing” tag, it got chopped.
At this point, we have our list of movies. The next step is coming up with a method to determine what “top” means. For me, it doesn’t just mean quality. It has to also include a measure of popularity. If we only care about quality, we’ll get a list with lots of unknown films that are rated highly by the small number of people who have seen them.
To account for that, I used a blended rating based on the number of reviews on IMDb as a proxy for popularity and the average rating as a measure of quality. To calculate popularity, I only compared a movie to those that came out within five years of its release date to as older movies will typically have less reviews than ones that came out in the last 20-30 years. The final score for each movie is based equally on popularity and average rating.
In aggregate, I think this is a reasonable list. Raging Bull is a legitimate pick for the best sports movie ever. The ranking does deviate heavily from The Athletic list though. Starting at the second pick, Warrior was rated in the 50s at The Athletic. I’m sure people would scoff at the idea that Warrior would be above Rocky and I certainly wouldn’t rank them that way. But each movie has over 400k reviews and Warrior has a slightly better rating. Some might say that’s an indication of the problem with IMDb reviews. I say that’s the people making their voice heard. Sly boxing is great. But Tom Hardy fighting is a little bit better.
Another thing that sticks out to me on this list as that by taking using a data set that includes voters from all over the world, we get some films that I wasn’t as familiar with. For instance, we have several entries from Bollywood. Dangal leads the pack landing in 6th place. That might seem high but again, it has a great rating and tons of reviews. According to Wikipedia, Dangal is the highest grossing Indian film of all time and the 5th highest grossing non-English movie. In that context, it deserves to be high on this list.
Several movies mentioned by commenters on The Athletic article appear here. She’s the Man, Bring it On, and 42 all make an appearance. For the Love of the Game was another commonly mentioned movie that sneaks into the top 93 here. Slap Shot is the top hockey movie at #37 followed by The Might Ducks at #66.
Making lists is fun. I enjoyed the one The Athletic released and I think this is an interesting companion to that. As a closer, I’ll offer my five personal favorites. Apologies for my bad taste.
- White Men Can’t Jump
- Happy Gilmore
- A League of Their Own
- Rocky IV