Click here for the spreadsheet
The American Film Institute's theater outside Washington, DC -Photo by Kate Mereand |
Twenty years ago on June 16th 1998, I was on vacation in the Outer Banks, and I couldn't tear myself from a show on CBS. Even as a relatively young person, I already had the priorities of a octogenarian. While the people I was vacationing with were busy doing things like partying in the sand, relaxing by the pool, or playing cards and games in adjacent rooms, I was glued to the television. Needless to say, I was not the coolest guy on the trip. I even grabbed my artist sketch pad I had brought to do drawings, so I could keep notes on the proceedings. My priorities may not have been yours, but, dammit I was committed to them.
The American Film Institute had put together a TV special based on counting down the one hundred greatest American films of the 20th Century. It was a broadcast so specifically designed to appeal to all of my key interests that to this day I worry I may have been used as a guinea pig in some scientific experiment that designed it. I loved movies. I loved ranking things. I have an inherent subservient tendency that love when authority figures tall me what is valuable in art. That's why I love when the Oscars declare with autocratic confidence what was best in film in a given year. Well this special, titled AFI's 100 Year... 100 Movies, felt like an Oscar ceremony for a whole century of movies!
A love affair began between me and the American Film Institute that continues to this day. The next day I all I wanted to do was talk about the list with people. I want to debate it, and analyze it, and break it down to the nth degree. Instead of wasting the wall space of my bedroom on photos of supermodels or rock stars I clipped a newspaper article with the AFI's list and taped that up. Years later I would rejoice at the opportunity to move close to the AFI Silver theater in Maryland. I was so excited to take part in their many events celebrating the past and present of filmmaking. Then, like any good creep, I would then try to form a relationship with the good folks there over the internet.
In the years following my initial introduction to their work AFI created more rankings shared across more specials. Like Waylon Smithers with Mr. Burns, I would dutifully knowledge each one, copy down everything they said, and nod in agreement. Over the years the AFI ranked the 100 greatest comedies, the 100 greatest romances, the 100 most thrilling movies, the 100 most inspiring movies, the 50 greatest heroes and 50 greatest villains, the Top Ten of ten different genres of film, and several more lists denoting cinematic greatness. I recorded all of these carefully in notebooks before moving on to spreadsheets on a computer. While the AFI released their final list ten years ago, I have gone back to those lists over and over again for my own purposes.
Eventually, I began to consider the idea of compiling all of those lists into a single resource. In a move that my therapist called "an obsessive's dream" I took it upon myself to create one comprehensive List of Lists ("Because YOU demanded it!"). Thankfully AFI was generous enough to share tons of very useful information on their website. There I could find the lists from their 13 special events and the background information of all the lists of nominated films they used to build their rankings on their website. From those raw materials I crafted a single spreadsheet to summarize everything that was great from over 100 years of American film history.
If you think my work with AFI lists is obsessive, you should see my Oscars research |
Now here comes the part of the piece where I describe my methodology in excruciating detail that almost no one will care about. However I have spent too much time in the company of academics to skip this step. So if you want to insight into my obsessive mind, read on! Everyone else can skip to the next picture. After that I will talk about fun things you can do with the List of Lists.
The first step in making a List of Lists is determining all the lists you have to list. Over 11 years the AFI produced 13 events announcing a set of rankings. They were, in chronological order:
- June 16, 1998- AFI's 100 Years...100 Movies, ranking the 100 best American movies to that point all-inclusive (all other lists will also be limited to films produced in the United States or with major involvement from Americans)
- June 16, 1999- AFI's 100 Years...100 Stars, ranking the 25 greatest male and 25 greatest female movie stars who had made their film debut before 1950 or who had died prior to the creation of the list (so more like the 50 most popular stars you'll find on TCM)
- June 13, 2000- AFI's 100 Years...100 Laughs, ranking the 100 greatest comedies
- June 12, 2001- AFI's 100 Years...100 Thrills, ranking the 100 most thrilling, or "pulse-pounding" in the words of the telecast, movie; there was not a strict genre restriction on this list like there were on others
- June 11, 2002- AFI's 100 Years...100 Passions, ranking the 100 greatest love stories in films both dramatic and comedic (but disappointing some, none that were pornographic)
- June 4, 2003- AFI's 100 Years...100 Heroes & Villains, ranking the 50 greatest heroes and 50 greatest villains portrayed in films
- June 22, 2004- AFI's 100 Years...100 Songs, ranking the 100 greatest musical moments, not limited to those in movie musicals or even songs written for films, rather any instance of music being used in the body of a movie
- June 21, 2005- AFI's 100 Years...100 Movie Quotes, this one should be self-explanatory
- September 23, 2005- AFI's 100 Years of Film Scores, ranking the 25 greats film scores, this was presented in a special one-night only event at the Hollywood Bowl instead of a broadcast
- June 14, 2006- AFI's 100 Years...100 Cheers, ranking the 100 most inspiring films; again this list was not confined to any specific genre
- September 3, 2006- AFI's Greatest Movie Musicals, ranking the 25 greatest musicals, this was also presented at a live performance rather than over television
- June 20, 2007- AFI's 100 Years...100 Movies – 10th Anniversary Edition, revising the original list of the 100 greatest movies, both to incorporate films since released and changing opinion regarding certain films (so goodbye to 1915's The Birth of a Nation)
- June 17, 2008-AFI's 10 Top 10, ranking the 10 greatest films of 10 genres: animated films, court room dramas, epics, fantasy, gangster movies, mysteries, romantic comedies, science-fiction, sports movies, and westerns.
My main purpose for this spreadsheet is to create a database of movies judged to be of holistically high quality. That means that some of these lists would not be conducive to my goal.
The list of 50 stars was the most difficult to deal with as I could not figure how to account for it. The AFI surely did not mean it as an endorsement of every film in the combined body of work of over 50 actors. I determined that The Plastic Age did not merit inclusion in the spreadsheet by virtue of an uncredited appearance by Carole Lombard. Had the AFI based inclusion on that list for specific films starring the actors as they did with the heroes and villains then I could use it. Instead I did not include that list at all.
Also I divided the 10 Top 10 into ten separate lists. Since each top ten was based a specific list of nominees that occasionally overlapped, it seemed appropriate to treat them as distinct entities.
The list of heroes and villains had some unusual features. Heroes and villains were both given the same number of ranked spots, but were not equally represented among the nominees. Further the nominees were not categorized as heroes or villains. Thus I treated it as one list of characters with split rankings.
This left me with a total of 21 lists to consider. The lists represented a combined 1,949 films ranging chronologically from 1912 to 2006. Some where represented in multiple rankings, while a majority of them were only nominated on a single list. Simply listing all the nominated films would leave me with a useless catalog that would not communicate the various ranking the AFI went to so much trouble to create. To try to capture the relative merits of each movie according to the lists, I would need a system to account for the frequency and placement of a film among the combined rankings.
First I decided to divide the rankings into tiers. I did not want to simply rank films by their ordinal number for a few reasons. There was no clear way to deal with the hundreds of nominees which had no ordinal ranking. Since I cannot be privy to the specific mechanations that determined the rankings, I couldn't be sure how much value to assign the difference between the 53rd and 52nd film on a list. However tiering could assign proper weight to the kind of intuitive ranges that the juries creating this lists likely had in the back of their minds. In a list of 100 films distinguishing the top 50 from the top 25 or top 10 makes sense. Tiers also made it easier to compare scores across lists.
From there I would assign each film a score for appearing on a list in any position weighted relative to it that position. I awarded a movie a number of points equal to the number of films nominated for that list divided by how many other films fell in the same tier (or higher) as that movie ( #Nominees/# in Tier=Points). If that quotient didn't result in an even number I rounded to the nearest integer. Thus each film that was nominated and not ranked earned 1 point (#Nominees/#Nominees=1). Then the higher up a list a movie ranked the more points it would earn. Lists of 100 would be have tiers for their top 50 and top 25. All lists had a tier for the top ten. Then the top film on each list was awarded as many points as there were nominees in the list ( by the logic that it was in a tier of 1 and #Nominees/1= #Nominees). Isn't this all sounding more interesting than anything you could do on a subtropical beach?
Ranked films that were drawn from larger pools of nominees (like the 500 nominees considered for the greatest comedies) earned more points than those from smaller pools (the 10 Top 10 lists each came from only 50 nominated films). That gave appropriate weight for films that stood out among crowded fields versus those selected within narrow genre restrictions.
The heroes and villains list again posed some challenges. Since I had decided to treat it as one list of 100 characters, I chose to split the points evenly between the two sub-categories. So I awarded the top 25 heroes and top 25 villains the number of points they would have earned for being in the top 50 of another list. Then I created a top 20 tier instead of a top 25 to include an even number of both goodies and baddies while also observing a number that seems like an intuitive division within such a list. The top ten and number one of both lists were treated as ties and shared the respective points (i.e. #Nominees/2).
Because the 10th Anniversary was openly an attempt to improve upon the original list by altering the rankings of some of the films, I also included an adjustment based on movement from one list to the other. The formula was designed to reflect both upward and downward changes by subtracting or adding a number of points equal to the difference between the movie's position on the 1998 list and its average position on both lists.
In keeping with my goal to create scores that were reflective of a film's totality, I grouped the lists into different two groups. Most lists ranked films as a whole or for a major component (e.g. the hero or villain of the story). I added all of those together for one score. In the List of Lists I kept those on the sheet labeled "List Set 1". For the lists that only ranked relatively minor aspects of a movie (the score, musical moments, and quotes), I included those points in a secondary score, the sheet labeled "List Set 2". You may choose to rely on either score as you see fit.
After that I could begin using the List of Lists to explore any number of research areas. I tried to treat the points like Bill James treats Win Shares. I keep their limits in mind, but I use them as a tool when answering certain questions. If creating the list was an obsessive's dream, then this was the sort of extended fantasy sequence for an obsessive that Hollywood might portray in a lavish musical number.
Unlike a beach vacation the fun never ends with a database like this! |
For example, Casablanca, was ranked on eight lists and was nominated for another. The Wizard of Oz reached the top 10 in six different lists including topping two of them. Perhaps most impressive is Schindler's List which was not the top movie on any list yet had the 11th highest primary score and 14th highest secondary score, topping many movies that had claimed a number one ranking! (I included that exclamation point so you can understand how much more fun I find this than whatever fun and games other people were having in the Outer Banks all those years ago.)
Another use for the list is to group films together in various ways. I included the year each film was released in my data and combined this with their cumulative scores to compare what films stood out most from each year. You can compare this to contemporary attempts to adjudicate the best movie each year, say for example the winners of the Academy Award for Best Picture. Those comparisons help us understand some of what goes into helping a film stand the test of time.
Given the amount of shade that gets tossed at Oscar and all his friends at AMPAS, you might expect hindsight to reveal the Best Picture winners to frequently underperform. Yet the AFI and the Academy actually agree on the best picture quite often (but of course they don't always agree). In the first year the Oscars were awarded based on the calendar year a film was released, 1934, both picked It Happened One Night as the top film. The last year of movies that were released before the majority of AFI lists were determined was 2002. In that year we again have a consensus pick in Chicago. In between those two extremes the Best Picture and highest score film line up may other times, e.g. Titanic, Unforgiven, Out of Africa, The Deer Hunter, The Godfather, The Sound of Music, On the Waterfront, Casablanca, etc.
However I also found plenty of disagreements that were pretty revealing. The Greatest Show on Earth is frequently cited as one of the weakest picks to win Best Picture. The AFI certainly supports that judgement. Across all 21 list it only scored 1 point. The highest scoring film of that year, 1952, was Singin' in the Rain has one of the ten highest point totals for any of the nearly 2000 films listed. And it wasn't even nominated by the Academy! Perhaps the happy and poppy musical seemed light weight compared to the big scale production by Cecil B. DeMille. However hindsight has allowed critics to recognize that beneath the song and dance numbers is an interesting story about the end of the silent movie era with a nuanced message about artistry and authenticity.
Two of the also-rans from 1987 accumulated more than ten times the points of that years Best Picture winner. While The Last Emperor is the sort of historic epic that people tend to associate with the Academy Awards, both Moonstruck and Fatal Attraction are better remembered today. Maybe the Academy overlooked them because they are stories primarily driven by women. Remember that #OscarsSoWhite campaign? This helps demonstrated that #OscarsSoMale too.
Conversely this method can identify some instances where the Oscars must have been very competitive. In some years multiple movies will have very high point totals. The race in 1977 may have come down to the Star Wars movie that started it all and eventual winner Annie Hall. They both racked up lots of points with the AFI with the winner depending entirely on which list set you favor. The Sound of Music had to face Doctor Zhivago in 1965, and according to my primary set of scores they are exactly tied! The year before may have been even harder though. Three of the nominees for Best Picture -Mary Poppins, My Fair Lady, and Dr. Strangelove: Or How I Taught My Eyes to Skim Past the Inordinately Long Subtitle- all rank among the 100 highest scoring films of all time.
I even tool breaking things down by year one step further. I assess each year based on how many films from that year appear in the List of Lists, how many total point were earned by the films from that year, and the average number of points earned by the listed films of that year. You can find that data in the sheet labeled "Year by Year", including a set of charts to help visualize the numbers.
Measuring years as a unit tells an interesting story of cinema history. The number of listed films each year is low through the early years of the 20th century reflective of an art form still establishing itself. Things pick up in the late 1920's as the silent film era peaks, then really take off with the introduction of the talkies. After that noteworthy films appear each year at a fairly steady pace with the exception of a brief drop in the mid-1960's marking the gap between the "Old" and "New" Hollywood eras. The numbers only tail off in the 21st century because a decreasing number of the AFI lists cover those years. This suggests that the film industry has basically been in good health since the introduction of sound.
The point totals and averages provide some more nuance to the story. The total points show some distinct peaks in filmmaking. The end of the 1930's and start of the 1940's stand out far above all other years. 1939 has frequently been described as the greatest year in American movies, and this data certainly bears that out. It has the most total points by far and the best average to a less extreme extent. This marks the peak of the studio system in Hollywood when movie making enjoyed a level of artistic and financial success it will likely never see again. Unfortunately the effects of World War II damaged the film industry taking away talent and resources for a higher purpose. The movies regained their footing, but they wouldn't hit another height like that until the period from 1959 to 1962.
The graph showing the average points has a shape very similar to a wave pattern. This suggests that the art form rises and falls as every few years new styles, approaches, and technologies revitalize it. While it appears as though the wave is ebbing with the peaks decreasing, I believe this is a natural effect of human retrospection. Those movies that have had longer to influence others and being studied and appreciated will of course have higher standing. However as they fade from living memory and more recent films have their time in the sun the shape of the wave will shift. Doubtlessly some folks in the mid-20th century would have argued that the silent films they remembered were far superior to the musicals and special effect laden blockbusters that followed. I conclude from this that the art of movie making is as alive and vibrant as ever with new peaks still to come.
You can do so much more with this data. I have already tried a few different experiments that I may share in later posts. Hopefully you will be inspired by the 20th anniversary of the AFI 100 Years... series and pursue your own ideas. If you do please tell me about them in the comments. Until then thank you for joining me in celebrating the wonder of the movies and strangeness of a guy who would rather watch a TV show about movies than party on the beach.