In April of 2021, after more than a decade, one of the most popular and influential manga and anime series finally came to a close: Attack on Titan or Shingeki no Kyojin. Over the course of this decade-long saga, Attack on Titan became the single most popular and influential manga anime series of the 2010s. When the first season of the anime aired in 2013, it quickly engulfed the anime community. I’d argue that Attack on Titan was actually a major factor in the increasing “mainstreaming” of anime because of its ability to appeal to people beyond typical anime fans. Unlike a lot of other long running anime and manga, the Attack on Titan hype train didn’t lose steam or peter out, in fact the hype only kept building and building to titanic proportions as more and more people got invested. I distinctly remember during the airing of Season 3 part 2 after the famous Erwin speech, the show brief held the mantle as the highest rated anime on My Anime List. It currently occupies the second spot, behind Fullmetal Alchemist Brotherhood.
I personally was actually an anime only watcher until Season 4 started airing, at which point I gave in and just read the remaining chapter in the course of one night. When it was announced that after over a decade, the final chapter would release in April 2021, people started going bananas. After all the twists, all the trials and tribulations over the many years, it was finally going to end. The internet was ablaze with people speculating where the story would go and how it would end, especially after the twist of the rumbling. However, as the leaks started to come in, it was quickly becoming apparent people would not be getting the grand conclusion they wanted. But still, people held out, hoping that the finished product might be different.
On one fateful day, April 9th 2021, the 139th and final chapter of Attack on Titan came out. Not to a thunderous and rapturous applause for one of the biggest anime manga sensations, but to a collective puzzlement and confusion. “Is this really it?” seemed to be the dominant thought of fans across the world. Without wading too far into the debate over Attack on Titan’s ending, the work’s ending provides an interesting lens from which to view narrative endings more generally, specifically the endings of manga, anime and light novels. To understand Attack on Titan’s ending, it’s necessary to look at what makes an “ending good” and why it seems to be so difficult to write these good endings.
At its most basic, for the ending of a story to be “good” or at least narratively successful, it really only needs to do two things. It needs to: one, wrap up the most important plot points and conflicts and two, provide a compelling conclusion to the arcs of the story’s major characters. It seems simple enough, but executing on these things well ends up being much harder in practice. A “good” ending can elevate an otherwise mediocre work while a “bad” ending has the potential to undermine an otherwise compelling story. To understand how these two factors work in practice, it’s helpful to analyse some works that are widely considered to have “good” endings.
Perhaps the platonic ideal of a well done ending to anime or manga is the ending of Fullmetal Alchemist, specifically the ending of the manga and the 2009 Brotherhood version of the anime. FMA is widely regarded as one of, if not the best, anime and manga series ever created. After all, it does hold the number spot on MAL The work is often lionised for its top notch approach to character development, world building and storytelling. But what is most interesting for our purposes is the ending of FMA. FMA succeeds where many other long running narrative works fail, in providing a compelling ending that satisfies the two criteria listed above. Brotherhood succeeds in wrapping up most of the important plot points. Father is defeated and Amestris is slowly able to rebuild. Most of the secondary characters get satisfying conclusions. Greed sacrifices himself to cripple father. Even though Edward doesn’t let Hoenheim sacrifice himself, he is able to die in peace, knowing that his work is over. Mustang regains his sight and Winry is finally able to be together with Edward.
But where the ending of FMA truly shins is in its ability to bring the journey of the Elric brothers to a satisfying and logical conclusion that is consistent with the themes of the show. The whole inciting incident of FMA is the brothers attempting to restore their bodies after a misguided attempt at human transmutation to revive their mother. Along the way, their quest to find a philosopher’s stone to fix themselves turns into a broader conspiracy over the government of Amestris and father’s plan to turn the whole nation into a giant philosopher’s stone. In the end, we see Alphonse sacrifice his soul to give Edward his arm back so that he can defeat Father. Edward, for his part, then sacrifices his ability to perform alchemy to bring back his brother.
Despite the almost sickening sweet optimism of FMA’s ending, it presents a fantastic example of how to deliver on your show’s principle characters. Edward sacrificing his ability to perform alchemy is such an incredible demonstration of his growth as a character. FMA begins with the Elrics’ dedicating everything to alchemy, believing it can fix all of life’s problems. This, of course, is what gets them into trouble as they believe they can bring back their mother. So for Edward to sacrifice his alchemy to bring back his brother is both a demonstration of his growth of a character but also an excellent way to bring the story full circle and deliver on one of the story’s key themes of sacrifice. FMA’s ending serves as the pinnacle of how to write a conventionally “good” ending that most fans will enjoy. I could go on and give more examples but most other popular and critically acclaimed anime like Steins;Gate or Code Geass also have endings that deliver in the same way FMA does.
So what went wrong with Attack on Titan? Fundamentally, it was a failure of the two key points of concluding the plot and finishing Eren’s character arc. For one, the ending of the story feels unsatisfying because it seems unable to commit to a truly dark and fatalistic ending, instead trying to straddle this middle ground where Eren only kills 80% of humanity and all of his friends thanking him for his sacrifice, even though they spent the past like twenty chapters doing everything they can to stop him. Furthermore, the ending seems to retcon many previously established elements of the narrative and leaves many plot points unfinished. Like when it’s revealed that actually Ymir was really in love with King Fritz even though he was massively abusing her. Perhaps the biggest culprit here is how Historia is treated in the final part of the story. Despite playing a major role in the events leading up the time skip, she does literally nothing afterwards. She quite literally could have just died off screen and it wouldn’t have made any difference which was rather bizarre for a character who was ostensibly so important to the plot. On the subject of bad characterisation you also have Armin’s bizarre “Thank you. You became a mass murderer for our sake.” I know some people thought this was a bad translation because it seems almost comically dumb, but I looked into the original Japanese and what Armin says is “僕達のために殺戮者(さつりくしゃ)になってくれて” which does basically mean something to the effect of “you become a killer for our sake.”
But the biggest tragedy with Attack on Titan’s ending is the way Eren’s character was handled. This is most apparent when it’s revealed that Eren was the one who directed Dina to eat his mother. This seems like a really innocuous statement, it really undermines Eren’s whole motivation for fighting the titans. Throughout the series, it is constantly reiterated that Eren blames Reiner for the death of his mother, so for it to suddenly be revealed that it was Eren who orchestrated it seems to be in conflict with the most basic parts of Eren’s characterisation. Even worse is the insinuation that Eren is being controlled by Ymir, the founding titan. The whole build up to the rumbling is that Eren knows that unless he wipes out the rest of humanity, the people of Paradis will never be able to live in peace. Therefore, he commits to committing mass genocide to allow his people to live free from the retribution and hatred of the outside world. For it suddenly be revealed that Eren might have been controlled and that he doesn’t actually want to die and see his plan through seems like a great injustice to his character. None of this is even to mention the really bizarre part at the ending of the chapter where it appears that Eren has been turned into a bird.
You often hear the ending of Attack on Titan compared, for good or bad, to the ending of Code Geass. But there are key differences that explain why Code Geass’s ending has basically been canonised in the anime fandom where the Attack on Titan ending seems rushed and contrived. What really resonated with Code Geass’s ending was its ability to deftly conclude Lelouch’s character arc. He starts off wanting to protect his sister and create a better world for her, to slowly realise that his goal is more important than just Nunnally. He wholeheartedly embraces becoming a villain and is willing to make any sacrifice necessary to see his plan, the Zero Requiem, through to its end. That’s what is symbolically important that Lelouch uses his Geass on Nunnally in the final episode to control the FLEJAS. It’s representative of how he’s changed from someone who is willing to abandon the whole Black Rebellion to fly off and save Nunnally to someone who is willing to make Nunnally hate and despise him if it means he can carry out the Zero Requiem. Lelouch’s death in the end is a fitting end for his character. From the very first episode, we get his recurring belief that “the only ones who should kill are those who are prepared to be killed” so it’s quite fitting that he ends up dead. Contrast this with Attack on Titan where Eren’s characterisation is not brought to a narratively consistent end and actually seems to undermine his development as a character.
To be clear, I’m not trying to say that for an ending to be “good” it needs to be universally popular with fans or even popular at all. There is a distinction between endings that are satisfying for fans and endings that are narratively successful or well written in the context of their respective stories. Of course there is a degree of subjectivity in what makes an ending good, after all there is no shortage of works with controversial or mixed endings. But as a gross oversimplification, most “good” endings tend to also be narratively successful while most “bad” endings tend to be narratively questionable. In reality, most endings tend to occupy some position in the middle of exceptionally well constructed and absolute garbage. I’m trying to make the assertion that when people talk about endings being “good” or “bad” it’s often in relation to the two points I detailed earlier. Endings are near universally panned, think like the final season of Game of Thrones, are usually perceived so because they either fail to conclude major narrative conflicts and plot points or they have characters act in ways that are out of touch with their previous characterisation.
If ending a story competently, if not well, only boils down to two fairly obvious concepts, why is it that good endings are so rare and seemingly hard to write with bad or mediocre endings being a dime a dozen, after all there are no shortage of dispossessed Attack on Titan fans opining about how they could’ve written a better ending. There are, of course, narrative reasons that gracefully ending a series, especially a long running one is hard. At the end of the day, making good art is hard and requires a plan, a plan that many writers probably don’t have. But in the end, I think the problem of endings often boils down to a combination of economic factors paired with the psychological and sometimes physical toil that working on one story for years brings. I find that, in many cases, these factors are amplified in the anime manga industry.
The economic factors are probably easier to understand and intimately tied to how the manga industry operates. The manga industry has a sort of two-tier production mechanism of producing manga magazines and individual collections. First are the manga magazines that introduce various series to readers for dirt cheap. The weekly Shonen Jump, for instance, is sold for $2 because of the intense competition within the industry. These regular releases in manga magazines is what gives rise to mega-long running series like One Piece or Dragonball. The fact that a lot of popular manga is serialised weekly in anthologies like Shounen Jump is problematic for authors because underperforming works face the risk of cancelation. This creates an incentive for authors to dream up lots of different and intricate plot details as they write their story to keep readers engaged and numbers high. The longer a work runs and the more characters and plot points are introduced, the more and more difficult it becomes to conclude them all effectively. There is often this negative feedback loop where as a series’s writing starts to decline, the work naturally becomes less popular. This creates pressure from publishers to improve or risk cancelation, which often has the effect of making the writing even worse. In many cases, authors find out that their work is getting axed without much notice so they need to dream up an ending far sooner and in much less space than they originally intended.
For publishers, manga magazines act as a form of advertisement to draw readers to the more profitable secondary market of individual collections. Once a manga proves popular in a magazine, the chapters are repackaged into Tankōbon, or volumes that are sold individually. However, even when you have a massively popular franchise like Attack on Titan, there are still diminishing returns for the publisher, especially as a story reaches its end. It hardly matters to the publisher if the ending is good or not because they have already made boat loads of cash selling these independent volumes over the years. Publishers are financially better off investing in a new and upcoming series than painstaking milking out the last bits of revenue from a nearly finished saga.
In TV original anime, you end up with a similar problem where a show’s initial season can be incredibly popular, leading to producers wanting to make more of it without necessarily having a narrative thought out. I think the best example of this is Psycho Pass. The show’s second season is quite notorious for being terrible, likely because it was just made to fill space until the movie came out. In fact, Gen Urobuchi, the lead writer of Season 1, didn’t even work on it. But ultimately a lot of these economic factors are responsible for bad writing more generally, not necessarily bad endings in particular.
To understand bad endings, it’s necessary to understand psychology, specifically both the psychology of fans and the psychology of what it’s like to be an author of a massively popular work like Attack on Titan. When you have something like Attack on Titan, that has cultivated such a titanic following, there’s virtually no way to deliver something that will meet the monumental expectations fans have placed on it. Despite how harsh I was on the Attack on Titan ending earlier, in reality if it was attached to something less popular, I doubt the public outrage would be that bad. In most other cases, the ending would probably have been considered passable, if not relatively decent.
I think manga anime fans, in particular, make writing endings difficult due to the extreme amount of tribalism that seems to permeate the fanbase. This is most apparent with the harem elements or “waifu wars” that pervade a lot of modern anime. I’d argue that anime fans often have a tribal attachment of characters that goes far beyond that of other mediums like films or books. I don’t think any other medium has “best girl” contests to rank characters in the same way the online anime community does for instance. So when you write a series that has any element of competing romantic interests in it, something that seemingly every anime does these days, you’re put in this position where one half of your fan base is going to be extremely pissed off if you have the main character conclusively get together with one girl. The author of The Devil Is a Part-Timer literally got death threats because people didn’t like the girl who Maou ended up with. Even in something like Oregairu, where there a clear logical romantic choice in the context of the narrative, that doesn’t change the fact that diehard fans of the other girls are going to get upset, leading to endings like Haganai, that cop out by not really delivering on the romance at all. Authors are put in this lose-lose situation where they either majorly piss off parts of their fanbase or slightly piss everyone off by writing a narratively inconclusive ending.
The extreme toxicity of some elements of the anime fanbase leads me into my next point about the stress of being an manga or light novel author. I imagine that no one really starts writing or drawing with the expectation that their work is going to blow up withs thousands, if not millions of people reading and watching. I honestly cannot imagine the extreme degree of stress that must put on a person to know that there are legions of people out there who will scrutinise every little detail of your work and ruthlessly critique it. When you spend years of your life dedicated to a story, I can’t imagine you’d want to intentionally write a subpar ending. But when you combine the pressure internally, from fans and from your publisher, I can see how that would be almost impossible to handle. I think that’s why in some cases authors just straight up don’t end their stories and just stop part way through. The author of the Haruhi Suzumiya light novels is a good example of this. The light novels were on seemingly indefinite hiatus starting in 2011 until it was abruptly announced in 2020 that they would continue. This is someone who went from a nobody to one of the biggest names in anime seemingly out of nowhere. Having thousands of fans eagerly awaiting your next book is pressure that I’d imagine few people are really able to deal with. This intense pressure often results in physical problems in addition to the mental stress. Drawing manga every week for an extended period of time is very physically draining and harmful to your health. There are a lot of examples of manga or webtoons artists causing physical harm to their bodies by drawing too much. Think like the Hunter x Hunter mangaka’s back problems or the author of Tower of God’s wrist problems that have put these series on indefinite hiatus.
Bringing it back to Attack on Titan, we’ll probably never know what actually happened. As it currently stands, the ending seems uncharacteristically rushed for a series that came to fame for its tight and impactful writing. Even though good endings are deceptively easy to write, there are a whole host of factors behind the scenes that make these good endings hard to achieve. After working on this story for over a decade, it’s quite possible that perhaps Isayama was starting to feel burned out or even tired of the series as it no doubt came with a whole host of pressure and stress. It’s also possible that Kodansha, the publisher of Attack on Titan might have somehow interfered and pressured Isayama to end the series sooner or faster than he would have liked. But ultimately we fans have no real way of knowing about any number of factors that went on behind the scene. So I think it’s worth giving Isayama the benefit of the doubt, after all I don’t think he wanted to write an ending that massively disappointed fans. To be clear, I’m not saying that you have to like the ending of Attack on Titan, rather I’m saying that we should give Isayama a certain degree of grace and respect. Even if the ending may not be what everyone wanted, it has undeniably been an unparalleled experience, for better and for worse, aboard Isayama’s wild ride for the past eight years.
