Method

1. As mentioned in the FAQ, lists had to meet the following criteria:

   1. The list had to be made up mostly of feature films (that is, films over one hour in length that were intended to be the main attraction in theaters). It may include a few short films or made-for-television movies or miniseries. Television series and individual episodes of television series are excluded.
   2. The list had to list or rank films based on their quality or some similar characteristic such as historical importance or popularity.
   3. The criteria for a film to be included on the list could not be too restrictive. For example, it could not include only a single year or only a narrow subgenre.
   4. The list must include at least 50 films (including short films and made-for-television films but not television shows or episodes).
   5. The list had to have been made or updated since 2007.
   6. The list had to have been made retrospectively. The only exceptions are Academy Awards for Best Picture and Best Foreign-Language Picture.



2. Each list is placed in one of two categories, depending on whether the list meets at least one of the following criteria:

   1. It came from a large sample poll (at least 500 respondents), regardless of who participated in the poll.
   2. It came from a poll of professional film critics, film scholars, and/or filmmakers, regardless of the number of respondents.
   3. It was created by a single professional film critic or scholar or a team of professional film critics or scholars.
   4. It came from a print source (e.g., book, magazine, newspaper).
   5. It was created by combining other available lists or ratings.



3. Steps 4-7 below are performed separately for each category of lists. That is the steps are done for just the lists that do not meet any of the criteria in step 2 and then for just the lists that meet at least one of the criteria in step 2. To complicate matters further, the composite list derived from lists that do not meet any of the criteria in step 2 actually meets the last criterion (being a combination of other available lists) and is included in the lists of films that meet one of the criteria. (Are you confused yet?)

4. The films on each list are recorded by rank. For ties, the average of the ranks is inputted. For example, if The Rules of the Game (1939) and Independence Day (1996) are both tied for sixth place, each is listed as being of rank 6.5 (average of 6^th place and 7^th place). If the list is unranked, the films are listed at what would be the middle rank if they were ranked. For example, on an unranked list of 100 films, all films are considered to be at rank 50.5 (the middle rank).

5. Each film is assigned points for each ranking based on the following formula: (22,800 + RANK²)^½ - RANK + 50. This is a fancy formula that simply ensures that the #1 film on a list gets 200 points and that a film that appears on a list gets at least 50 points no matter how far down the list it appears.

6. For each film, all points are added together. Then the films are sorted by points from most to least. (Yes, there is one simple step.)

7. A statistical procedure called factor analysis is used to create factor scores for each film. This procedure looks at how well each list measures the concept that this site calls "greatness" and assigns more influence to lists that measure this concept better than others do. It can also assign scores to individual films based on how well they meet the abstract greatness concept. These scores are used as the first tiebreaker. The factor analysis can also be used to find clumps of similar lists. The three tastes described in this site represent three clumps that are found consistently and are easily interpretable.

8. For the analysis of lists that meet one of the criteria in step 2, the second tiebreaker is the rank of the film among the lists that do not meet any of the criteria in step 2.

9. This final list of films created from lists that meet one of the criteria in step 2 is called the "master list." The master list is cut off at the point where two or more films are tied with each other despite the tiebreakers or where a film on only five or fewer lists appears. In other words, the master list contains no ties, and all films appear on at least six lists.

10. All other film lists on this site are cut off immediately before the highest ranking film that does not appear on the master list.

11. The list of the top years in film is made by adding up all the points for all films within each year and sorting from most points to fewest points. Films contribute points regardless of whether they are on the master list. The list of years is cut off immediately before the first year that has no films on the master list.

12. For the list of the actors in top films, the Webmaster uses reviews and personal memory to determine the three most important people who appear in a film in order. The lead actor gets half the points for the film, the second actor gets three-eighths of the points, and the third actor gets one-eighth of the points. Again, actors get points from films that are not on the master list. The actors are sorted with an arbitrary cut off point, usually determined by the Webmaster's reluctance to expand the various actors pages much further.

13. The list of the top directors is made by adding up all the points for all films that each director directed. If multiple directors directed a single film (or different segments within a single film), the points are divided evenly among them. A director must be credited to receive points unless nobody is credited. In that situation, the director(s) identified by the Internet Movie Database gets the points. Directors can get points from films that are not on the master list. The directors are sorted with an arbitary cut off, albeit a more generous one as a longer list does not lead to significantly more work for the Webmaster.

This page was last modified on June 20, 2016