One way this expresses itself is by keeping track of my movements on Flickchart, partly so that if their server ever dies, I'll still have my rankings in an Excel file. But partly also for analyses like the one I am about to do.
I've just reached the milestone of 6,000 movies added to Flickchart, which is still 187 shy of my actual number of viewings. I've gotten behind on what was once a real-time ranking of new movies seen, but right now I'm as close to being caught up as I have been in some time, only about six months behind.
And having just crossed that milestone, I decided it was time to analyze my own Flickcharting to see how consistent I've been. But first I need to explain how it was even possible I could do that.
I'll start by explaining my method of tracking my Flickcharting offline. I have an Excel spreadsheet which gives a real-time snapshot of where my movies are on Flickchart. When a new movie gets added, I go to the appropriate row on that sheet and insert it, moving all the other movies down by one, in keeping with what actually occurs on the site. When a movie wins a duel against another and moves up that way, I'll cut it and paste it from its lower rung on the spreadsheet to its new position of prominence. In this way, I will always have an exact backup of my Flickchart in Excel.
But then I also have a sheet in the same workbook where I keep track of the movements as they occur. I tend to alternate adding new movies with dueling random movies, so you will get a pair of line items on the sheet that look like this, my most recent two "transactions" completed:
A quick explanation of this.
The top line shows that The Golden Compass beat Copland in a random duel, which moves it from 929th on my chart to 884th, pushing Copland down to 885th. I recently rewatched Copland and was not enthralled with it, so I expect more duels with outcomes like this.
The second line shows my addition of C'mon C'mon, a film I was not a big fan of. There are only about a thousand films I've ranked so far that I liked less. As you can see, it landed at 4923 on its initial entry into the chart, losing to Skyscraper but besting The Last Unicorn. (Seeing as how I have not seen The Last Unicorn since the early 1980s, it's dubious how accurate this is anyway.)
If my screenshot had been extended just a little to the left, you'd see that this was row 4243 in the spreadsheet, which gives you an approximate idea how long I've been keeping track of these movements.
So what to do with all this data? Well I'll tell you.
One would be to track a particular film's movements over time. Did I mention that I also take snapshots of the main page every 10,000 rankings, saving them as their own sheet, so I can look back in time to see where I once was? Well I do that too. However, that's on as-needed basis when I'm curious about a particular title, and so far, the need has not arisen.
Instead, I decided to use this milestone to see how consistent my dueling decisions have been over the years. By searching titles -- assuming I spelled them correctly in my haste to record them -- I can see how many movies beat another movie at one point in time, then lost to the same movie at a different point in time. Theoretically, the implicit goal of working toward a perfect Flickchart means that all your decisions are sacrosanct and internally consistent with one another -- an impossible standard, but one all we Flickcharters secretly like to believe is hypothetically attainable.
This little trawl of my data would show me just how consistent I've been.
Now, if I wanted to spend all my time analyzing this data, I'm sure there's a lot I could find out. But I'd really need an algorithm to help with something like that. I'm not some Howard Hughes who has all the time I want to parse my Flickcharting decisions while wearing boxes of Kleenex on my feet.
So I decided to just go from the top of the list and see how many inconsistencies I could find in my top, say, 100. Due to the way I duel -- which I won't explain right now, just trust me -- those top 100 movies duel more than any other on my chart, so they are the easiest to track if I want the most number of results.
Now, it should be no surprise that the top ten did not contain many movements. I've not had a lot of changes in my top ten during this time. In fact, since Raising Arizona has been my #1 the whole time, it doesn't appear at all on this page of the workbook -- having never lost a duel nor needed to win a duel to move up.
But as we get closer to the bottom of that top ten, some interesting results are revealed.
For example, I discovered that my #7, Fargo, has both beaten and lost to my #8, Toy Story. Since they are only separated by one on my chart, one might assume something like this would be the case. A rewatch of either one or the other could easily be responsible for such a shift, though in this case, I doubt that's a factor -- I happen to have not rewatched Fargo since 2012 and Toy Story since 2013. I haven't been keeping this spreadsheet for that long. (In fact, Excel says I created it on May 28, 2014.)
So while we said this result is not a surprise, it may be a surprise just how frequently this has occurred after these two.
The Iron Giant, my current #10, has both beaten and lost to Wargames, my current #21. I think this movement probably occurred not long after my most recent Iron Giant viewing in late 2018, when it ultimately moved into a spot in the top ten that it has not yet relinquished.
Tangled, my current #14, has both beaten and lost to Jesus Christ Superstar, my current #15. But then Jesus Christ Superstar has also beaten and lost to Schindler's List, which is now down at #43. JCS used to "slum it" with Schindler's List down in the 40s but has since gotten a big boost.
As we continue, more movies have switched places with more than one other movie. Do the Right Thing (#20) has had trouble deciding whether it's better or worse than both Run Lola Run (#22) and Unforgiven (#24). Wargames not only traded places with Iron Giant, as listed above, but also with Big (#33). The same can be said for Run Lola Run, which adds Donnie Darko (#23) to its list of dance partners in addition to Do the Right Thing, and then there's Shawshank Redemption (#26) being fickle with both A Fish Called Wanda (#30) and Time Bandits (#31).
I abandoned this exercise at this point, as it was becoming pretty clear that these sort of results would continue to bear themselves out down the chart.
I suppose what I would really need is to see if two movies that are far apart from each other on the chart have actually mutually beaten each other, but a) that's much harder to track, because I would have to have my suspicions of which titles these might be in advance, and b) Flickchart doesn't really work that way, except for over a long period of time with a lot of intervening movements. Once one film beats another, they are next to each other on the chart, at least for the immediate future -- so it's not so easy for them to be at a great distance from one another again, unless you purposefully re-rank a movie because it seems out of place.
I did find one interesting result from within this top 26, which was that Bound (#19) actually beat Big (#33) at two different times. How is it possible for one film to leap over another film twice? Shouldn't it be ahead of that film after the first win? Yes it is, but then the second film can leap frog the first by beating a film that's ahead of both of them. It's nice to know that both times that Bound and Big dueled with Big in the lead, I saw it fit to reverse their respective positions. That's the sort of consistency I'm looking for. Of course, I don't and can't keep track of how many times one of the films might have beaten the other without their positions changing -- and in fact, Big might have won a duel against Bound in this scenario when it was in the higher slot.
There's one other takeaway I want to discuss before I let you go.
I noted that Shawshank had beaten A Fish Called Wanda and vice versa, but what I didn't say is that A Fish Called Wanda has actually leaped past Shawshank two different times, with Shawshank getting the win in the middle duel.
Now I don't keep track of the dates these duels occurred on this spreadsheet, but I can tell you that on line 1888 -- which means after I'd added 944 new movies and had 944 lower movies win a duel over a higher movie -- Wanda defeated Shawshank to move from 30th to 21st. Five hundred twelve "transactions" later, on line 2400, it was Shawshank upending Wanda to go from 24th to 23rd. Then a mere 142 lines later, which was probably only a couple months depending on how frequently I was charting at the time, Wanda claimed its revenge on Shawshank by moving from 25th to 24th. It seems unlikely that I watched either movie during that time, though I have seen both movies in the past ten years.
But that's not nearly the shortest time period to change my mind. Sticking with Shawshank, on line 2602, Time Bandits bested Shawshank to move from 26th to 25th. Only 64 transactions later -- in other words, only 32 new movies added and only 32 movies repositioned by the outcome of a duel -- Shawshank took the position back by jumping from 27th to 25th. And yes, that two-slot difference suggested there was an intervening transaction related to Shawshank, which was the aforementioned Big insinuating itself between the two of them.
The only real conclusion to draw from this is what Flickcharters have acknowledged from time immemorial, time immemorial being the 13+ years Flickchart has been available to the public: that human beings, perhaps especially cinephiles, are fickle, and they might select differently between two movies depending on what day of the week it was and what mood they were in.
And when you have a chart with more than 6,000 movies on it, these relative positions are anything but absolute -- even when your thoughts don't change as the result of new information, like a repeat viewing. These are fine differences indeed, imperceptible levels of preference and variations in quality.
So yeah, I've written all this to conclude something you likely would have surmised before you even started reading.
Hey but at least I used my data.
No comments:
Post a Comment