Fun With Numbers: Kishi Seiji’s Rough Road

Way back in 1960, the NBA’s Cincinnati Royals drafted future hall-of-fame basketball player Oscar Robertson with the first overall pick. In his first season, he was named rookie of the year. In his second season, he became the only player in NBA history to average a triple double (i.e. putting up ridiculous stats in 3 separate historical categories). In his fourth season, he was named the league’s most valuable player. In his fifth through seventh seasons, he never made it past the first round of the playoffs. In his eighth through tenth seasons, he didn’t even make the playoffs despite putting up consistently great personal stats. In his eleventh season, on a new team with the man who would become Kareem Abdul-Jabbar, his team won a then-record 20 games in a row and eviscerated opponents in the playoffs, winning him his first-ever NBA title.

Robertson was great for basically his entire career, and it’s not like he lost those skills when his teams weren’t winning. So then why didn’t they win? Because in team sports like basketball, teams matter. When his team’s second-best player is a 41-year-old coach coming back into the game as a publicity stunt, how good a player is doesn’t much matter. It still takes good players to win championships, but great players not named Michael Jordan don’t win championships alone.

Anime production is not very much like basketball, but it’s a similarly complex process where circumstances can contribute as much as individual skills do to the net result. Before work on a show can even start, a producer has to successfully pitch an idea to sponsors and justify the business side of operations. A capable cast and staff have to be assembled. Those staffers then have to both have to develop an clear vision for the series and adequately communicate that vision with the hundred-plus animators who typically work on a modern TV anime. And for the project to be a success, that vision then has to resonate with its target audience, something which just doesn’t always happen.

I mention all this because it pertains very much to the discussion of director Kishi Seiji, one of only four directors in the history of anime to helm 3 non-sequel 10k+ hits, and the only one to do so at three separate studios. In spite of having set a career milestone that puts him on the same spreadsheet as Tatsuyuki Nagai and Yoshiyuki Tomino, Kishi has been a constant target for all sorts of fan ire. Taking a quick look at his career, it’s fairly easy to see where this sentiment originates. After a barely-notable start to his career, Kishi spent the years between 2007 and 2010 knocking off three straight winners (Seto no Hanayome, Astro Fighter Sunred, and Angel Beats) and making a bit of a name for himself. Angel Beats, for all its success, has its fair share of detractors, but the majority of bad mojo Kishi has generated comes from the next 3 years of his career, the stretch from 2011 to 2013, that made him one of the many to earn the nickname “the Uwe Boll of anime”. I categorically reject this label, not because all of the shows he directed over that stretch were good, but because the stretch was a daunting one in a way people rarely think about (and included some impressive achievements regardless).

What follows is a defense of the man’s record based mostly on the extenuating circumstances he’s faced and the results he’s produced regardless, rather than the aesthetics of any particular show he’s directed. This take leans primarily on sales numbers as a metric of success because, while they’re an imperfect measure of show quality, they’re at least a bit more solid than the nebulous idea of a unidirectional popular consensus, a concept which is typically a side effect of people tending to spend the most time in communities that agree with them. It’s worth being clear here that, in the long term, good sales aren’t the only effect makes something notable. If you implement a visual technique that people imitate for decades to come, especially if you get to name it ala Dezaki’s postcard memories, then money tends to matter a bit less as far as notability goes. But great sales figures do make a show notable. They mean a show has strong appeal with at least a specific group (not the same thing as universal appeal).

The most basic reason why Kishi is sometimes see as being all over the place as a director is that, from 2011 to 2013, he was literally traveling all over the place to make his shows. During the aforementioned stretch, Kishi directed a total of 8 non-sequel series at 4 studios (AIC ASTA, Lerche, Brains Base, and Sanzigen). That is a lot of different shows, and is relatively rare among notable directors. Very often, directors big enough to have a recognizable name take years between projects, either waiting for a “good” opportunity or taking more time to prepare a better product. To get a quantitative idea for exactly how uncommon that was among notable directors, I took a look at the CVs other directors who had directed multiple non-sequel 10k+ hits, and also tossed in directors with 1000+ myanimelist member favorites as well, to get a first-order comprehensive list of notable anime directors. I then looked looked at how “busy” each had been during the busiest 3-year stretch of their career.* All data used can be found on this sheet.

Of the 21 non-Kishi directors on the list, only 2 put up comparable 3-year stretches directing over 6 non-sequel series; Akiyuki Shinbo (2009-2011, 8 series) and Tsutomu Mizushima** (2010-2012, 7 series). Let’s take a look at what sort of stuff they adapted/made over that stretch:

Akiyuki Shinbo non-sequels (2009-2011):
Natsu no Arashi [manga]
Bakemonogatari [20k+ light novel]
Dance in the Vampire Bund [manga]
Arakawa Under the Bridge [manga]
Soredemo Machi wa Mawatteiru [manga]
Madoka Magica [original]
Denpa Onna to Seishun Otoko [light novel]
Katte ni Kaizou [manga]

Tsutomu Mizushima non-sequels (2010-2012):
Shinryaku! Ika Musume [manga]
Yondemasuyo, Azazel-san [manga]
Plastic Nee-san [manga]
Blood-C [original]
Another [novel]
Joshiraku [manga]
Girls und Panzer [original]

Genres aside, these two lists actually have a decent amount in common. Both of those runs have pretty high-quality peaks, containing some unquestionably notable series (Madoka Magica, Girls und Panzer). They’ve also got some pretty unambiguous duds (Dance in the Vampire Bund, Blood-C). Over half of the series being adapted over each stretch were manga, which come pre-storyboarded and are thus typically easier to adapt than other media, which makes for a good point of transition into an examination of Kishi’s similarly packed 3 years:

Seiji Kishi non-sequels (2011-2013):
Kamisama Dolls [manga]
Carnival Phantasm [manga]
Persona 4 [game]
Jinrui ga Suitai Shimashita [light novel]
Devil Survivor 2 [game]
Aura [light novel]
Dangan Ronpa [game]
Aoki Hagane no Arpeggio [manga]

First and foremost, Kishi’s 2011-2013 stretch is different from the previous two in that it includes as many game adaptations as it does ones of manga. Game adaptations are often a tough order to fill because they lose player input (a fundamental part of the medium which good games make pretty extensive use of) in the transition to anime, and it doesn’t really help that two of them were 1-cour. The novels Kishi adapts are also considerably less popular pre-anime than Bakemonogatari (45k in two weeks pre-anime) or Another (50k-8k=42k by a re-release before the first episode of the anime aired) were. And there’s not an original series in sight (though Kishi would go on to direct the admittedly underwhelming Hamatora in Winter of 2014). In spite of all that, Kishi did pretty well commercially, putting 4 of the 6 TV series above 3k averages, heftily boosting the Kamisama Dolls manga, and turning P4 and Arpeggio into 10k+ hits.

The most common counter to the argument of using his big successes to argue for Kishi’s ability level is the idea that he backed into his 10k+ notches like an elderly driver backing into a tree; unintentionally, and with negative consequences. Dismissing that level of success as trivial is problematic for a pair of reasons. Firstly, the notion that an anime’s success can be either totally assured or completely impossible before it airs on name alone is a misguided one that neglects how complicated the production process and market for entertainment is.

Secondly, because the shows he got his blockbuster stripes on weren’t slam dunks by any means. In all 3 cases, he was operating with capable coworkers and good marketing, but none that guaranteed the level of success he had on each project. Let’s go through them chronologically:

-When working on Angel Beats in 2010, Kishi was partnered with Key ace writer and composer Maeda Jun, and prominent animation studio PA Works. PA works’ recent history at the time did not suggest they were a lock for 10k+ hits; their two series prior to Angel Beats, True Tears and Canaan, had averages of 2079 and 4208, respectively.  Maeda Jun’s trickier to compare directly, since he was only responsible for the source material in most cases where his name comes up with respect to anime. His works were adapted several times, though, to somewhat varying results. Averages for the Toei and Kyoto Animation versions of Kanon varied by a full factor of 2, from under 9000 to over 18,000. Clannad, the biggest Key adaptation title prior to Angel Beats, averaged 24,808 copies, a total which AB’s 34,116 handily outstripped.

-When adapting the 2011 Persona 4 series that ultimately averaged 31k, Kishi was adapting an extremely popular video game, one which had sold over 350,000 copies in Japan. That sort of adaptation probably has better odds than an adaptation of an unknown video game, but video game popularity will not get you a commercial slam dunk anime. Few video games adapted into anime are as popular as Persona 4 was, true, but there are plenty of adaptations in the 100k+ range that performed at a mediocre level as anime. Two particularly striking examples are Valkyria Chronicles (180,000 game copies pre-anime, 2620 series average) and Shining Hearts (190,000 game copies pre-anime, 315 anime average). Also of note is that 2008 Tales of the Abyss adaptation, which averaged a solid 7498, but was adapted from a game which sold over half a million copies in Japan (200k above P4). Even beyond the mechanical issues with video game adaptations, turning their adaptations into blockbusters, as Kishi did, is no cakewalk. That’s true even when the game in question has been a huge success.

-Aoki Hagane no Arpeggio, his last, most borderline, and most recent 10k+ hit, gets occasionally downplayed due to its marketing ties to Kantai Collection, a popular browser game with over 1 million subscribers at the time the anime aired. The 12,000 or so people who bought the disks represent the tiny fraction of that base, some 1.2 percent. It’s hard to find a counter to that specific hypothesis, as the shipgirl trend is a relatively recent one. But there’s a general case to be made that advertising and capitalizing on trends will only get you so far. People aren’t mindless consuming zombies who buy things they actively don’t want just because somebody suggested it to them – it has to be something they at least kind of want. That goes doubly so for something that carries a price tag north of 30,000 yen. Any given subgenre of anime, from mecha to harem fanservice to perrenial redhead stepchild yuri, contains a wide range of sales figures to match a wide range of reception. And it’s impossible to define a subgenre containing 10 or more shows where the floor is a 10k average. Capitalizing on trends is one of the oldest tricks in the book production-wise, but that’s not a guarantee of success. Execution has to fill some of that gap, and amazon reviews suggest that the battle scenes and visuals in that series swayed at least some fence sitters.

As was mentioned in the bit about 3-year stretches earlier, games occupy a weird position in the Kishi narrative. This is in part due to the position video games as a medium occupy in the western anime fandom. Unlike the Japanese sphere, fans are much more likely to have played a video game than they are to have read a given manga or light novel. People who have encountered source material before trying an anime are more likely to heavily scrutinize it, similarly to how people pick up on more symbolism and technical stuff when watching a series for the second or third time.

This is not to say that criticizing or disliking a given video game adaptation is at all unwarranted, but it has meant in practice that a lot of the commentary on Kishi in the west comes from fans of Dangan Ronpa who have cast him (rather than, say, the single cour given for the adaptation or general difficulties in adapting from games) as the culprit who ruined the adaptation of franchise. My default question is this; given the same amount of time and story to fit in, how does a replacement do a significantly better job without going in an aesthetic direction entirely different from that of the original game? I don’t see it, though I’d certainly be happy to argue the point.

Kishi Seiji is not the best anime director ever, by a long shot. But plaudits like that get tossed around too easily these days, anyway. The Mount Rushmore/Penthouse Floor of anime direction is, in no particular order, Hideaki Anno, Hayao Miyazaki, Mamoru Oshii, Yoshiyuki Tomino. And it should stay that way for at least the near future; everybody who started taking the lead last decade and is still making 1-2 shows a year has to wait their goddamn turn. What he is is a skilled one who soldiered through a tough production schedule and came out with both some impressive wins and some unimpressive losses, a guy who’s extremely fun to watch work in his zone. These next several months offer a nice opportunity to kick the tires on the man, seeing as he’s animating a well-advertised original project and adapting a half-million manga for a Friday-night late slot.*** There’s a finite possibility, however small, that he extends his own record of directing 3 non-sequel 10k hits with 3 different studios to 4 or even 5 in the next couple of months. I have been and will presumably continue to be on that bandwagon, enjoying the whole process.

*As defined by the number of non-sequel long OVAs, movies, and TV series directed over that span. You may find more 7+ non-sequels/3 years stretches if you expand the list to all directors, but likely not that many more. Prior to the 90s, anime had more episodes per series, which limited the number of different projects people could work on in one year. And in the more modern era, you’ll have a tough time finding directors who haven’t had broad-scale success repeatedly getting jobs with all sorts of franchises.

**Tsutomu Mizushima almost not having a reputation (no en wikipedia page, among other things) is as much of a downer to me as Kishi’s generally negative rep. The dude’s a fierce-ass animal who needs more attention.

***Fuji TV is presumably putting the series on late because schoolkids running through gun-related curricula are a bad-juju distraction waiting to happen. The need for that kind of self-censorship is debatable, but the time shouldn’t hurt its odds of being successful; Inu x Boku SS, Durarara, and Angel Beats all aired in similar slots (albeit on different networks).

7 thoughts on “Fun With Numbers: Kishi Seiji’s Rough Road

  1. Fascinating stuff. Shinbo is pretty well known for delegating his work. I really can’t imagine how many hours these other directors worked during their three year bout, especially when Kishi Seiji has to work with a brand new staff almost every series.

    Directing an anime, let alone 8 in 3 years, seems like an inhuman amount of work. Do anime directors even have managers or booking agents?

    • I have no idea how the technical process of staffing thoss shows went. Kishi works with one writer for a lot of his shows (Makoto Uezu), so I would imagine having some chemistry there would make the setup a bit easier.

      I doubt most directors have agents, if only because I have a hard time imagining the salaries they make are the extravagant kind that people would be gunning for a cut of. I would imagine that networking with producers would be the key part of getting large amounts of work, since they’re often the ones assembling staff/cast lists. “Free agent” directors who get that much work are fairly uncommon, so people like Kazuhiro Furuhashi/Tsutomu Mizushima/Junichi Sato/Seiji Kishi could each be unique cases. I think the easy answer, though, is that some people just work really long days. Maybe they talk about stuff like that in Shirobako at some point.

  2. Pingback: Fun With Numbers: Industry Valuation of In-Betweening vs. Voice Acting | Animetics

  3. Pingback: Fall 2014 Watch List #2 | Animetics

  4. Pingback: ICYMI: Highlight Posts of 2014 | Animetics

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s