Top 1,000 Films of All Time (Calculated)

Created by gokhan.

Favorited 3 times, disliked 0 times, added to 5 watchlists.

SOURCE: https://letterboxd.com/prof_ratigan/list/top-1000-films-of-all-time-calculated/

There are many statistically-minded lists available. Generally, they are either based on averages or cumulative presence in lists. This list attempts to bring the two together, with a composite score that uses Letterboxd, IMDb, Metacritic, and Rotten Tomatoes scores as well as a quantification of the movies' presence on several lists (enumerated below). I would describe the results as a consensus opinion of the films someone "should" see. [You can also see the list of 5,000 films here.]
I started with a list of about 30k titles gleaned from many Letterboxd lists, IMDb data, and other sources and crunched the numbers. I limited composite scores to those above 500 IMDb votes. Votes were normalized against the maximum votes on that website to account for the differing traffic, so the Highest Voted movie would be a 1.00 and the lowest 0.01. I weighted the differing website rating according to my own sense of importance with Letterboxd about 10x Rotten Tomatoes and Metacritic about half of IMDb. I also included weighting by year to account for the significant recency bias in online vote tallies. That weight was Votes x (1+((2000-Year)/100).
The way that I quantified movies on lists like Sight & Sound was to give those films a sort of fictional rating as if it were from IMDb. So, for instance, Vertigo, at #1 on Sight & Sound's Top 250 got 100, Citizen Kane got 99.96. Lists where there was no ranking all received 100. Sight and Sound, Palme D'or, and Oscar Winners received the greatest weight, which was about 10x a Oscar nomination outside Best Picture and about twice the Golden Bear or AFI Top 100.