- Capture.PNG (34.13 KiB) Viewed 2733 times
2020 SCFZ Letterboxd Statpalooza
Re: 2020 SCFZ Letterboxd Statpalooza
SCFZ's lowest rated directors (25 ratings per director minimum)
Looks like 50 ratings is around the cutoff to make a director's poll possible, so I whittled the list down to the ones we haven't polled yet, in case anyone wants ideas:
Ritwik Ghatak
Edward Yang
Mani Ratnam
Sergio Leone
Sergei Paradjanov
Hiroshi Teshigahara
Bela Tarr
Thanks for the stats, Kanafani, they're always fascinating!
I've never been in a real fight, but I would be willing to schedule a fist fight to defend Delphine in this case. DM me Flabrezu
SCFZ's highest rated movies of the 1920s (minimum 5 ratings for a movie to be eligible)
- Capture.PNG (63.7 KiB) Viewed 2713 times
SCFZ top 40 of the 1930s:
- Capture.PNG (62.38 KiB) Viewed 2713 times
SCFZ top 40 of the 1940s:
- Capture.PNG (64.85 KiB) Viewed 2712 times
SCFZ top 40 of the 1950s:
- Capture.PNG (63.48 KiB) Viewed 2712 times
Top 40 of the 1960s:
- Capture.PNG (64.31 KiB) Viewed 2712 times
Top 40 of the 1970s:
- Capture.PNG (64.18 KiB) Viewed 2712 times
Top 40 of the 1980s:
- Capture.PNG (66.19 KiB) Viewed 2712 times
Top 40 of the 1990s:
- Capture.PNG (63.21 KiB) Viewed 2712 times
Top 40 of the aughts:
- Capture.PNG (65.44 KiB) Viewed 2712 times
Top 40 of the 2010s:
- Capture.PNG (60.97 KiB) Viewed 2712 times
Top 40 from the USA:
- Capture.PNG (64.32 KiB) Viewed 2711 times
Top 40 from France:
- Capture.PNG (62.2 KiB) Viewed 2711 times
Top 40 from Japan:
- Capture.PNG (66.22 KiB) Viewed 2711 times
Year of production for the SCFZ Top 1000 movies:
- Capture.PNG (15.94 KiB) Viewed 2707 times
Oldest movie in the SCFZ top 1000?
The Astronomer's Dream (Georges Méliès, 1898), at 657
The Astronomer's Dream (Georges Méliès, 1898), at 657
Which directors have the most movies in the SCFZ top 1000?
- Capture.PNG (26.7 KiB) Viewed 2703 times
SCFZ's worst rated movies:
- Capture.PNG (66.18 KiB) Viewed 2702 times
The SCFZ Top 100:
- Capture.PNG (163.35 KiB) Viewed 2701 times
SCFZ Thoroughly Unofficial Top 100 up at letterboxd:
https://letterboxd.com/kanafani/list/sc ... d-top-100/
I've seen 84!
https://letterboxd.com/kanafani/list/sc ... d-top-100/
I've seen 84!
when it comes to ranking our favourite directors, looking at our average rating per director is really measuring which directors we think were most consistently good. that might be the best measure, but you could look at it another way: the directors we like best might be those who made several great films, even if they also made a few duds. those directors wouldn't do so well just looking at average rating (assuming scfzers had seen the duds, since those duds would drag the director's average rating down). so i wonder what the top directors list would look like if we used a low minimum, like 25 ratings (that produced the most interesting list to me), but only took the average rating of each director's five highest-rated films.
it also might be interesting to look at two measures of variance:
- if we could, for each director, compute the standard deviation of scfzers' average rating for that director, we could work out which directors are most 'controversial', i.e. which are loved by some and detested by others. e.g. if there were five scfzers, and their average ratings for hitchcock films were 4.0, 4.0, 4.0, 4.0 and 4.0, hitchcock would have a std deviation of zero, and hitch wouldn't be controversial at all. if jean rollin had five averages of 1.0, 2.0, 3.0, 4.0 and 5.0, he'd have a standard deviation of √2 ~ 1.4, which would be very high for a five point rating scale.
- if we could, for each director, compute the standard deviation of each sczfer's ratings, and then average those standard deviations, we could work out which directors had the most controversial output overall, i.e. which directors were most inconsistent (had films we think are great but also films we think are terrible). e.g. if there were five scfzers, and looking at their ratings of hitchcock, if the first person's ratings of hitch films had a std deviation of 0.2, the second had a std deviation of 0.1, and so on, we could average those std deviations. we'd then get a measure of how much variance there is in scfzer ratings of hitchcock films, and we could compare hitch with other directors using that measure.
those might not be easy to code though! if you might want to try and if i can help at all (i'm sure you're familiar with the math but if not, i can help with that) let me know.
it also might be interesting to look at two measures of variance:
- if we could, for each director, compute the standard deviation of scfzers' average rating for that director, we could work out which directors are most 'controversial', i.e. which are loved by some and detested by others. e.g. if there were five scfzers, and their average ratings for hitchcock films were 4.0, 4.0, 4.0, 4.0 and 4.0, hitchcock would have a std deviation of zero, and hitch wouldn't be controversial at all. if jean rollin had five averages of 1.0, 2.0, 3.0, 4.0 and 5.0, he'd have a standard deviation of √2 ~ 1.4, which would be very high for a five point rating scale.
- if we could, for each director, compute the standard deviation of each sczfer's ratings, and then average those standard deviations, we could work out which directors had the most controversial output overall, i.e. which directors were most inconsistent (had films we think are great but also films we think are terrible). e.g. if there were five scfzers, and looking at their ratings of hitchcock, if the first person's ratings of hitch films had a std deviation of 0.2, the second had a std deviation of 0.1, and so on, we could average those std deviations. we'd then get a measure of how much variance there is in scfzer ratings of hitchcock films, and we could compare hitch with other directors using that measure.
those might not be easy to code though! if you might want to try and if i can help at all (i'm sure you're familiar with the math but if not, i can help with that) let me know.
-
Monsieur Arkadin
- Posts: 423
- Joined: Mon May 27, 2019 5:56 pm
Delphine Seyrig in 1975 just knocking it out of the park!The SCFZ Top 100:
Sure, flip, this is all quite straightforward stuff. I’ll work on it!flip wrote: ↑Tue Dec 29, 2020 7:50 pm when it comes to ranking our favourite directors, looking at our average rating per director is really measuring which directors we think were most consistently good. that might be the best measure, but you could look at it another way: the directors we like best might be those who made several great films, even if they also made a few duds. those directors wouldn't do so well just looking at average rating (assuming scfzers had seen the duds, since those duds would drag the director's average rating down). so i wonder what the top directors list would look like if we used a low minimum, like 25 ratings (that produced the most interesting list to me), but only took the average rating of each director's five highest-rated films.
it also might be interesting to look at two measures of variance:
- if we could, for each director, compute the standard deviation of scfzers' average rating for that director, we could work out which directors are most 'controversial', i.e. which are loved by some and detested by others. e.g. if there were five scfzers, and their average ratings for hitchcock films were 4.0, 4.0, 4.0, 4.0 and 4.0, hitchcock would have a std deviation of zero, and hitch wouldn't be controversial at all. if jean rollin had five averages of 1.0, 2.0, 3.0, 4.0 and 5.0, he'd have a standard deviation of √2 ~ 1.4, which would be very high for a five point rating scale.
- if we could, for each director, compute the standard deviation of each sczfer's ratings, and then average those standard deviations, we could work out which directors had the most controversial output overall, i.e. which directors were most inconsistent (had films we think are great but also films we think are terrible). e.g. if there were five scfzers, and looking at their ratings of hitchcock, if the first person's ratings of hitch films had a std deviation of 0.2, the second had a std deviation of 0.1, and so on, we could average those std deviations. we'd then get a measure of how much variance there is in scfzer ratings of hitchcock films, and we could compare hitch with other directors using that measure.
those might not be easy to code though! if you might want to try and if i can help at all (i'm sure you're familiar with the math but if not, i can help with that) let me know.
(and not just because I yearn for fist fights)Top 40 of the 1990s letterboxd list:
https://letterboxd.com/kanafani/list/sc ... the-1990s/
https://letterboxd.com/kanafani/list/sc ... the-1990s/
Some ideas for tomorrow or later this week:
- scfz ranks Oscar winners: rank best motion picture winners by average rating
- same as above, for a couple of shitty popular lists (imdb 250, afi...)
- one movie per year (highest ranking)
Last edited by kanafani on Tue Dec 29, 2020 10:45 pm, edited 1 time in total.