It started with Stripey’s post about what your top 5 says about you. And in my comment, I realized that of my 5 favorites, 4 came from 2007 (Code Geass, Gurren Lagann, Seto no Hanayome, Higurashi Kai), and the other candidates aired in 2006 (Haruhi season 1, Higurashi season 1). And then I realized all of those were shows that I had marathoned. So it got me thinking. Is there something about watching a show weekly that takes away from my enjoyment of it? Alternately, was I subconsciously boosting the scores and memories of shows I did marathon? I hypothesized watching a show on a weekly basis might take away from the enjoyment and the memories a slight bit, because I got to stew on each episode for a week, whereas when I marathon a show, I tend to be more bigger-picture focused, and not dwell on minor issues.
The numbers get worse, I’m afraid to say. Source
So I thought I’d have a little fun with numbers, but taking a look at my MAL scores. I discounted all the movies, ONAs and OVAs, since almost all of them are one-shot deals, and focused on just the series that aired at least 12 episodes (or in Nodame Cantabile Paris’ case, 11). I counted 86 shows completed or near complete. 37 are shows I marathoned (defined as having started and finished in 2 weeks or less – although over 90% were started and finished within 1 week). 49 were shows I watched on either a weekly basis or over an extended period of time. This group includes shows that I may have watched in "mini-marathon" batches along the way. It also includes Bakemonogatari, Zan Zetsubou Sensei, and FMA Brotherhood, all series I haven’t technically completed, but felt confident enough about my rating to include them in this group.
Some interesting numbers came to light.
The average score of my marathon group was 8.43, while my average score of my weekly group was just 7.04. But to be fair, the marathon group largely consists of many "greats." So the difference in raw scores is to be expected.
So I took a look at the variances vs. the global score instead. My marathon group had a variance of +.16, while my weekly group had a variance of –.92. Interesting, but again not surprising, because the weekly group was much more likely to include shows that I disliked. And lower deviations tend to have heavier weight in my rankings, since even a bad show on MAL typically won’t get a global rating under 7, whereas I’m willing to rate as low as 4 (anything under that I wouldn’t stick around to watch). I wondered if the strong negative outliers would have a big impact on the difference in variances, but when I took out the top 5 positive and negative variances, the marathon group still had an overall variance of .12, while the weekly group had a variation of –.8. Better, but .92 is still a big difference in what amounts to a 7 point scale.
The numbers were interesting, but it still didn’t tell me whether or not I had a ratings bias for shows I marathon vs. shows I follow weekly.
My last ditch effort to try to quantify the bias was to compare the variances of shows that I rated an 8+ between both lists. Once a show reaches 8+ level, I think my rating system becomes a little more consistent. So I figured if there was any sort of bias, it would either show or not show in these cases.
I rated 32 of the 37 shows I marathoned, and 22 of the 49 shows I followed weekly, an 8+. Of the marathon group, I had an average rating of 8.64 and +.3 variance vs. the global score. Meanwhile, of the weekly group, I had an average rating of 8.21, but interestingly, there was a +.07 variance, the first positive variance for this weekly group. On first glance, it still seems like a significant difference though. However, when you take away my 5 favorite shows (of which I have a predictably high rating and variance), the average rating and variance for the marathon group drops down to 8.18 and –.02. Very comparable to the ratings of the weekly group.
These are pretty simple calculations. But in the end, I think the numbers showed that I rate shows pretty consistently, regardless of whether I marathon them or watch them weekly, particularly when a show reaches a very good level (once you account for the favorites bias). It may just be that my memory is skewed to thinking that I enjoy marathoning more, simply because the shows I did marathon were fantastic shows to begin with. However, given a choice between marathoning and following a show weekly, I’d take the marathoning route. Simply because I enjoy experiencing that sense of accomplishment of striking a show off my list. There’s nothing worse than having a long "currently watching" list that seems to never shrink.
So what say you, people of the world? To marathon or not marathon? Do you have a preference? Or am I just foolish to think that anyone but myself (and Lambdadelta I guess) would be interested in pointless numbers. 😉