This is my 14th year doing a version of Blindspot, so perhaps my customary introduction to the concept is totally unnecessary… but just in case there are still people who haven’t heard of it, this is a challenge in which you pick twelve films you’ve never seen but feel you should have (your blindspots) and watch one per month throughout the year.

In chronological order, this year’s films are…

	1910s Anders als die Andern
1920s The Phantom Carriage
	1930s Dr. Jekyll and Mr. Hyde
1940s Suspicion
	1950s Hiroshima Mon Amour
1960s Belle de Jour
	1970s Sorcerer
1980s Poltergeist
	1990s The Prince of Egypt
2000s The Royal Tenenbaums
	2010s The Great Gatsby
2020s Poor Things

For many of the previous 13 years, my Blindspot selection process has been tortuously complicated. I don’t think I’ve ever simply picked twelve films I feel like watching (which is how I’ve seen other people make their selection — perfectly reasonably). Instead, I usually compile various “great films” lists, rank and weight them in various imaginative ways, and thus concoct some kind of ranking-of-rankings to generate 12 new picks. I long ago ruled out relying on the same methodology every year, because that way it’s never surprising and never refreshed — I’d just be working down a very long list, year by year.

Last year, I made things a lot simpler: I looked on Letterboxd for the most popular film I’d not seen from each decade since the origin of feature films, and there was my list. I enjoyed that ‘history of cinema’ approach so much, I decided to repeat it this year. But, as I said, I don’t like to just take the next film on any given list, so I mixed it up slightly. By default, the decade search on Letterboxd displays 18 films (three rows of six). So, having filtered it to the ones I’d not seen, I randomised the selection by rolling a d20 (because rolling dice is fun).

If you’re curious, the rolls were as follows…

“But, hold on a minute,” you might say, “a d20 has 20 sides, and your selection lists only had 18 films — what happened on a 19 or 20?” A perfectly reasonable and well-observed question. And, as you can see, a 20 was indeed rolled. Of course, I’d thought of a solution in advance, because it was likely the situation would arise (there being a 1-in-10 chance of rolling 19 or 20 on a d20, and there being 12 rolls). The solution was… more dice rolls (because rolling dice is fun).

Or, rather, one more roll, as it only happened once. As a 19 or 20 is a high success, I limited it to the top row (i.e. the six most popular films) and rolled a d6*… and on that, I got a 1 — meaning the 1950s is the only decade for which I’m watching the most popular unseen film, i.e. the one I would’ve watched if I had just picked the next film on the list. Which is fine — the point of randomising the choice wasn’t to stop it ever being the next film on the list, just prevent it being definitively and only that.

* (I appreciate that this system would make no sense in a game (why is 1–6 the optimal result but 19 or 20 results in another 1–6?!), but this isn’t a game, it’s a random number generator, so it’s fine.)