I would like to share my experiences with two principles I have used in selecting movies to watch: radical diversification and high quality. This was my main focus in 2016 and will probably remain important for some time to come.

I also tried to gain a solid movie knowledge. An example would be films I watched only because they were in the Sight and Sound top 250. Although it is an interesting objective, I would not necessarily recommend it.

1. I experienced a very rapid expansion of movie knowledge.

2. Even with 500 films watched in total you can have much knowledge. As long as you experiment with different kinds of films.

3. I find it hard to focus on one director. My four favorite directors are Ford, Ozu, Tarr and Kieslowski. I haven't seen all the major works of any of them.

4. I often know which movies will be my favorite without having seen them, or without even having seen another movie by the same director. I anticipated that The Human Condition would be my favorite film and had equally high expectations of Satantango.

5. I always have many films I want to see as soon as possible. Often when I see an especially great movie I think about another movie I will probably like even more.

6. I often don't take enough time to reflect upon the movies I have seen, because the next film is demanding all the attention. I think this is a disadvantage.

7. Incredibly good films are just average for me. It's just that your standard for average changes. Nevertheless, the experiences can be very rich.

8. Even though right now I know more about my taste in movies, this doesn't stop me from exploring new horizons. If anything, it does the opposite.

I think it would be interesting to investigate a little into what constitutes diversification. In order to do that, I will compare two lists, IMDB's Greatest Films and The Great Unknown on Letterboxd.

I took the first 25 films of each list and will try to quantify the level of diversification. I have included one entry for both the Lord of the Rings trilogy and the Human Condition Trilogy.

First I'd like to take a look at the represented countries. I divided t productions among the involved countries.

I would like to invent and introduce a few percentages, in order to quantify geographical diversity of films.

The US Percentage (US) gives the share of films made in US. The US t Percentage (USJ) includes t US productions, with t productions divided among the involved countries. The non-US Percentage (NUS) states the share of films made outside the US. The non-English-speaking Countries Percentage (NESC) gives the share of films made in non-English-speaking countries.

The reason I introduce these specific percentages is because in my experience films from US and English-speaking countries are often overrepresented in English-speaking countries. Maybe if I would be Japanese, I would have introduced a Japan Percentage.

I would also like to introduce a ratio I think is interesting: the Dominance Ratio (DR), defined by dividing the average of the number of films of countries 2,3 and 4 by the number of films by country 1. This indicates the dominance of the most represented country.

With these ratios in mind, let's analyze IMDB's top films and The Great Unknown (TGU).

IMDB List (first 25 films)

USA: 17.75
Italy: 1.25
Brazil: 1.00
: 1.00
Japan: 1.00
UK: 1.00
Australia: 0.50
: 0.50
New Zealand: 0.50
Spain: 0.25
West : 0.25

US: 60.0%, USJ: 71.0%, NUS: 16.0%, NESC: 16.0%
DR: 1 : 16.4

TGU (second 25 films):

Japan: 8.00
: 3.50
USA: 3.00
India: 2.00
Taiwan: 2.00
Iran: 1.50
Soviet Union: 1.50
Brazil: 1.00
Philippines: 1.00
Sweden: 1.00
Cuba: 0.50

US: 12.0%, USJ: 12.0%, NUS: 88.0%, NESC: 88.0%
DR: 1 : 2.8

The difference is obvious. If the IMDB list would be a representative sample of the movies someone watches, we might argue their watched movies are not properly geographically diversified.

I like the dominance ratio. If we wouldn't pay attention to it, we might think Japan is overrepresented in the TGU list, just like the US in the IMDB list. But the DR nuances this view. The US in the IMDB list is about 5 times "as dominant" as Japan in the TGU list (16.4/2.8).

I think it is interesting to note that if we count t productions, the total number of represented countries is the same (that is, 11). That's why I don't think this measures diversification adequately.

However, if we look at the number of countries represented without ing for t productions, we get 5 countries for the IMDB list and 10 for the TGU list. This seems accurate.

Let's now look at chronological diversification. I will first introduce the number of Major Decades Represented (MajDR). I consider the first quantitatively important decade the 20s, so the maximum is 10. The Minor Decades Represented (MinDR) is the number of decades represented between the 1870s and 1910s (5 in total).

I will denote the Total Decades Represented (TDR) as MajDR+MinDR, in order to give insight into the nature of the decades in which the films were made.

We do have to note the importance of considering the amount of movies over which we compute TDR. That's why I also want to look at the Average Films per Decade Ratio (AFDR), that is, TDR divided by the number of films considered.

Let's now compute TDR and AFDR for both lists.

IMDB: 8+0 (over 25 movies), TGU: 8+0 (over 25 movies).
Both have an AFDR of 3.1.

Let's now take a closer look.

IMDB: 1990s: 12, 1970s: 4, 2000s: 3, 1950s: 2, 1940s/1960s/1980s/2010s: 1.

TGU: 1960s: 7, 1970s: 5, 1950s: 4, 1980s: 3, 1930s/1990s: 2, 2000s/2010s: 1

We could make an Decade Dominance measure (DD). This time I will define it by dividing the films from the most common decade by the AFDR.

IMDB: 3.84 (1990s) - TGU: 2.24 (1960s).

Only now we can see a difference in chronological diversification. The 1990s are more dominant than the 1960s in their respective lists. Still, both have a somewhat dominant decade.

Let's take a look at Color Films/B&W-Color Films/Black-and-White films/Silent Films.

IMDB: 21/1/3/0
TGU: 9/0/17/0

I am not entirely sure about all the TGU films, but this is my best guess. If we would make a Color/B&W Ratio (CBR) (with B&W-Color counting as Color) we would get

IMDB: 1 : 7.3
TGU: 1 : 0.53

Here is one of the biggest differences between the lists.

I would also like to look at the length of the movies. Let's start with the average length. If we use the data on Letterboxd we arrive at the following:

Average duration: IMDB: 144 minutes, TGU: 134 minutes.

There is an average difference of 10 minutes. No big deal, right? Now we also compute the standard deviation.

IMDB: 144 +/- 30 minutes, TGU: 134 +/- 57 minutes.

This is interesting. We see that the spread in duration of the TGU films is significantly greater than the deviation of the IMDB films. This is a good indication of diversity in duration.

We can also compute the ratio between the shortest and the longest film as a measure for diversity, although it is more sensitive to outliers. If we would do that, we'd get:

IMDB: 207/96 = 2.2, TGU: 340/83 = 4.1.

Incidentally, both measures delineate more or less the same picture about the spread in duration.

I think all in all this kind of analysis does a good job of an initial comparison in diversity.

One of the most interesting findings is that mass popularity of a film is perhaps not only dependent on the film, but also on the proximity and conformity to an homogeneous ideal. A 140 minute 1990s US film might have a better chance at a IMDB top 25 spot than a 70 minute 1960s French film, regardless of the quality.

There remain interesting other measures of diversity to be found and defined. For example, we have not looked at genres or directors (the female director/writer : male director & writer ratio seems especially promising). We have not looked at documentary/film diversity.

I hope this gives some ideas about the ways in which we ourselves can diversify. For example, I know nothing about China's films, Nollywood or Bollywood (geographical diversity). I have not watched many silent films (we counted those). And I probably have a high DD for the 1960s.

I will give some advice for those who might want to try to adopt the diversification principle. I think many dichotomies (e.g. US/non-US) can be overcome by equanimity with regard to origin, release date, etc. Personally, though I am biased in favor of and Japan, I don't care where a movie is from.

Secondly, many people think twice before committing to a 4+ hour movie. I think this hesitation can be overcome by watching one incredibly long movie. After watching Out 1: noli me tangere (12.5 hours) I could handle almost anything with relative ease, including those 3 hour Tarkovsky masterpieces.

I think it is worth noting that according to the high quality principle it is not a "waste of time" to watch a long movie instead of more shorter movies, as long as the average quality is sufficient.

On the other hand, many people often don't watch movies under 10 minutes. I think Don Hertzfeldt would be an excellent starting point in that regard.

With that I will conclude this list. I will now go and probably watch a movie.