A little over a year ago, I did some data mining on the IMDB Top 250 and AFI Top 100 lists of movies in an attempt to highlight changes in movie quality over time. The resulting analysis spoke more to the flawed nature of these lists, and lists in general, and less to any actual trends in movie-making over the years.
With that in mind, I thought it’d be interesting to revisit the IMDB Top 250 list to see how it had changed over the past year. Was it even more weighted towards recent movies? Is “Pirates of the Carribean” still in the Top 250? Who else made the cut? Find out, after the jump.
Here’s a reproduction of the original analysis from the IMDB Top 250 list as of September 30, 2008. This graph shows the count of movies in the IMDB Top 250 list for each year, plus a 4th order polynomial trend-line.*
*This is just what Excel has handy. The same disclaimer from last year still applies: IANAS (I Am Not A Statistician).
When compared to the AFI Top 100 film lists, the IMDB seems to exhibit a bias towards newer films. You can see this in the graph through the cluster of red bars towards the right. The median year for the distribution is 1975, and the best year for movies is 2004, which had 9 movies make the Top 250.
So what happens when we redo the analysis for the list as it stands on October 18, 2009, roughly one year later?
This distribution isn’t radically different, but there are a couple of differences worth noting. First, the median year moved up to 1977 from 1975; in other words, over the course of only one year, the composition of the Top 250 list advanced by two years. Second, the best year for movies is now 2008 instead of 2004. 2008 had nine films crack the Top 250.
These changes would suggest that the IMDB list’s bias towards new movies is intensifying. So what exactly is driving these changes? Let’s take a closer look.
Additions to the List
|Bringing Up Baby(1938)||245||8||1938|
|Sweet Smell of Success(1957)||238||8||1957|
|Harold and Maude(1971)||242||8||1971|
|Good Will Hunting(1997)||226||8||1997|
|The Curious Case of Benjamin Button(2008)||215||8||2008|
|Låt den rätte komma in(2008)||220||8||2008|
|(500) Days of Summer(2009)||197||8||2009|
Of the nineteen additions to the 2009 list, twelve came from the last two years (6 from 2008, 6 from 2009). These include big 2008 Oscar winners “Slumdog Millionaire,” “The Wrestler,” and “The Curious Case of Benjamin Button” as well as the cream of the 2009 summer movies crop: “Star Trek,” “Up,” “Inglourious Basterds,” “District 9,” and “(500) Days of Summer.” And the other 2009 addition?
“Zombieland,” at #164. Huh? Okay, that movie was pretty good, but was it Top 250 good?
To me, all of this is pretty solid evidence that the Top 250 list is increasing in its bias towards newer movies. As further evidence, let’s take a look at the movies that dropped off to make way for these newcomers.
Subtractions from the List
|Cabinet des Dr. Caligari., Das (1920)||224||8||1920|
|Nosferatu, eine Symphonie des Grauens (1922)||198||8||1922|
|Bride of Frankenstein (1935)||221||8||1935|
|Belle et la bête, La (1946)||191||8||1946|
|Ace in the Hole (1951)||250||7.9||1951|
|The Searchers (1956)||239||8||1956|
|Dolce vita, La (1960)||242||7.9||1960|
|The Man Who Shot Liberty Valance (1962)||232||8||1962|
|In the Heat of the Night (1967)||225||8||1967|
|Young Frankenstein (1974)||235||8||1974|
|Barry Lyndon (1975)||245||7.9||1975|
|Tonari no Totoro (1988)||247||7.9||1988|
|Lola rennt (1998)||243||7.9||1998|
|Mou gaan dou – Howl’s Moving Castle (2002)||233||8||2002|
|Pirates of the Caribbean: The Curse of the Black Pearl (2003)||238||8||2003|
|Hauru no ugoku shiro – Infernal Affairs (2004)||249||7.9||2004|
|Shaun of the Dead (2004)||248||7.9||2004|
|Iron Man (2008)||205||8||2008|
Of the nineteen drop-offs, only five were from this decade. The rest were fairly evenly distributed between 1920-1998. Here, IMDB’s bias actually helps self-correct itself, at least a little bit. Did “Shaun of the Dead,” “Iron Man,” and “Pirates of the Caribbean” really deserve to be in the Top 250? Last year, I had actually argued that “Shaun” and “Pirates,” though they were excellent movies, probably didn’t deserve their status in the list, and sure enough, they fell off.
Unfortunately, that bias works across the board and also affects movies that (arguably) should have stayed. I was shocked to find that three classic horror movies from the 2008 list–“Cabinet of Dr. Caligari,” “Nosferatu,” and “Bride of Frankenstein”–didn’t make the new cut. Though perhaps it’s fitting that in this age of “Twilight,” the original genre-defining vampire classic no longer has a place in the Top 250.
Again, IANAS, but my interpretation of the movement in data from 2008-2009 shows pretty clearly that IMDB’s bias towards newer movies has gotten worse. The fact that twelve movies from 2008-2009 that muscled their way into the list after one year seems pretty egregious to me.
Or then again, I could be wrong. The raw data for the 2008 and 2009 lists are available here if any of you want to take your own cracks at crunching the numbers. And as we all know, there’s a lot of room for debate when it comes to IMDB’s rating methodology. I more than welcome others’ statistical expertise and insights into this data.
What do you think? Does “Zombieland” deserve a spot in the IMDB Top 250 at the expense of “Shaun of the Dead” and “Nosferatu?” Is the IMDB Top 250 inherently flawed? Do you still use the IMDB Top 250 as your go-to list for best movies of all time? Sound off in the comments!
Newer movies almost always get to spend a little while in the list, as initial viewers slap a bunch of 9 and 10 ratings on it. Given time though, you can pretty well bet that most new movies will fall off – I would definitely expect to see Zombieland (and maybe Star Trek) fall off the list sooner or later.
For the reasons DaveW said, it may be more applicable to throw out the last year (or x months) of movies from the list before analyzing.
What I’m curious about is how those 80-year-old movies make the jump ONTO the list. Safety Last? Sherlock Jr.? I can only assume that those movies’ IMDB ratings haven’t changed much – it’s just that the cutoff ebbs and flows. They’re always on the bubble.
@Dan V, DaveW, that would be interesting to do, but part of the point of my analysis was that the brand new movies that hop on due to overinflated ratings doin’t fall off the list at the same rate.
In other words, the rate of self correction lags behind the rate of inflation of new movies on the list.
So, anyone want to take the raw data and try to lop off the last X years?
On a side note, are there any Nosferatu-Twilight mashups out there on teh internets? Seems like a winning combination. Like, OMG, Nosferatu.
Also, isn’t it amazing how much better the ‘subtractions’ list is than the ‘additions’ list?
And re: how do old movies get added–especially with the silent movies from the ’20s, could it just be a function of them finally getting enough votes to be counted? I’m not sure how many people decide to go watching silent movies and then voting on them in the internet, but once you get beyond the more famous ones, I would think that it would be a small number.
I also know that the IMDB algorithm is somewhat wonky, partially designed to keep someone from signing up for a million accounts and voting a bunch of ’10’s and ‘0’s and doing nothing else, and also involves a bunch of statistical weighting. It’s far from a simple average of votes.
Actually the Nosferatu and Twilight have more in common than you might think. One of the big emotional beats in Twilight is the idea that the Edward/Bella relationship is a fatally stupid for both of them. If you think about this too carefully, it’s really stipod. Obviously Bella could get killed, but what’s the worst that’s going to happen to Edward? He falls off the no-human-blood wagon for a day? (Or even a year, or a couple hundred years?) Cry me a river. I guess he’d feel bad about destroying his soul mate, but all he has to do to avoid that is make her a vampire too, which we all know is going to happen anyway, so who gives a monkey’s?
But this is a case where the average Twilight fan is a little smarter than people might think. Sure there’s no reason to think that Bella poses a threat to Edward based on Twilight alone. But there’s a STRONG tendency in vampire stories for the monster to be destroyed by the love/sacrifice of a pure woman. And even though you can see this narrative at work in the Gary Oldman ‘Bram Stoker’s Dracula,’ it’s actually something that appears for the first time in Nosferatu, as far as I know. And the two versions of Nosferatu are still the only ones where the vampire actually seems to be somewhat afraid of the woman even as he’s drawn to her… other than Twilight. So yeah. Point to Stephanie Meyers.
The upshot is that a Nosferatu-Twilight mashup would actually be hella easy, because Nosferatu is all about doing shot/reverse-shot sequences between the vampire and his love interest. All you’d need to do is replace the footage of Max Schreck looking like a rodent with footage of Robert Pattinson looking like a nubile god.
Ha! I wrote “really stipod.” Who’s stipod now, internet?
@Stokes, re: stipod: At first I thought Apple had come out with a new trendy product I hadn’t yet heard of.
@Everyone else: It’s funny to me that The Curious Case of Benjamin Button is now on the top 250. That movie came up on my Netflix account today and had a ranking of something like 3.5 out of 5–decent, but definitely not “the top films of all time” material. (Also, I thought that movie was stupid and boring and trite, and my opinion is gospel so nyah! But, seriously, any universe in which that movie is considered better than Young Frankenstein or La Dolce Vita is a universe I don’t want to live in.)
Mark, I hope you keep tracking these figures for as long as OTI is running. It’s fascinating stuff.
Hey, you know what else is interesting? Of the movies that fell off the list, nine were foreign-language (although I admit I’m stretching the definition a bit with the silent movies). Of the movies that were added, one, maybe two were foreign-language (depending upon where you place Slumdog Millionaire).
@Stokes: Yup, I plan on making it an annual ritual.
Another fun thing to do with the statistics is analyze changes in rankings.
Top 2 losers (without falling off the list):
“Battle of Algiers” (1966), -56
“Crash” (2004), -49
Top 2 gainers:
Sunrise: A Song of Two Humans (1927), +42
In Bruges (2008), + 40
Also, “Terminator 2” advanced 16 spots in the rankings. :-)
Fascinating analysis, thank you! I come across a surprisingly large number of people who have no patience for older movies (which would include movies from the median year of 1977). I suppose IMDb’s top 250 is a reflection of what movie viewers of all kinds are enjoying the most, rather than a careful analysis of the best movies of all time.
Will anyone remember Zombieland in 5 years? No. Will anyone remember 500 Days of Summer? Or even Star Trek, of which I’m a huge fan?
All of these “best movies of all times” list should institute some kind of waiting period. Maybe a movie has to be out for at least 5 years before it can go on the list, to prevent the newest and coolest from stealing the spots of more deserving fare.
Unless the point of the IMDB list is to service fanboys. “What’s a really good movie, according to broad pop culture, current generation thinking? Oh, Shaun of the Dead is good, huh? I’ll go see that.” If that’s the point of the list – what’s a good movie, subjectively, right now – then everything is fine.
Somewhat on topic, I liked Bloomberg’s Top-Grossing Films of All-Time list that came out back in July, Gone With The Wind is number one after adjustments for inflation: http://www.bloomberg.com/apps/news?pid=20601088&sid=aLzh3gmCoWdo
not that I care but you switched infernal affair and holw’s moving castle in your last table.