Written by: [jomifisi](https://comuniq.xyz/profile?u=jomifisi_17) <br>Between 1991 and 1999, the ending of a big Japanese RPG almost always looked the same: dozens of hours of walking, a final dungeon with no breathing room, and a boss whose health bar seemed calibrated on a different scale of values. In *Final Fantasy VI* (Super Famicom, April 2, 1994), the party had to split into three teams to climb Kefka's Tower before facing a villain with multiple layers. In *Chrono Trigger* (March 11, 1995), Lavos appeared in three consecutive forms. In *EarthBound* (August 27, 1994), the last adversary only yielded to something the attack menu didn't offer. This wasn't an accident or a lazy team: the difficulty spike was the predictable result of an economic, technical and dramatic design that the genre had inherited and didn't yet know how to question. ## The easy excuse: "it was the arcade era" The explanation that travels fastest is the one about coins: old games were hard because they wanted our money. That works for *Ghosts 'n Goblins*. It doesn't work for a forty-hour RPG on a cartridge that had already been paid for in the shop. An RPG didn't charge coins, it charged **time**, and that's where the problem gets interesting. Someone who has invested that much time instinctively demands an ending that feels earned. The final boss became the point where design tried to convert hours spent into a feeling of accomplishment, and it did so in the crudest way available: by making the number on screen bigger. The lineage matters too. *Dragon Quest*, released in 1986 by Yuji Horii, was born looking at *Wizardry* and *Ultima*, two computer RPGs famous for showing no mercy. When Square and Enix started competing for the same audience in the Super Famicom era, the idea that an easy ending would be a betrayal was already in the genre's DNA. And no studio had much incentive to challenge it. ## The unsolvable problem: what level is the player who reaches the end? In an action game, the designer controls the curve: whoever reached the last level went through everything the game put in front of them. In an RPG with random encounters, not so. Two players can arrive at the last boss's door twenty levels apart. One followed the story straight through. The other spent an afternoon deliberately getting lost in the plains of the Veldt, or stood in a corridor earning experience while watching TV. The designer has to pick a target, and almost always picked the diligent player: the one who explores, does the side quests, stockpiles. Whoever rushes hits the wall. That explains the cruelest aspect of the design. The final boss didn't only measure firepower; it measured **inventory discipline**. *Final Fantasy VI* is the textbook case. Kefka's Tower forces you to manage three teams, each with their own equipment and spells, and at the top waits an adversary whose *Fallen One* attack reduces the whole party's health to 1. The fight presupposes a reserve of revival and restoration items that the player didn't burn through in trivial fights along the way. Whoever did arrives at the final battle with an empty bag, and the game never told them that was a resource for the end. The wall wasn't about level. It was about planning, and nobody had taught it. ## Layers, phases and the grammar of spectacle There's a second explanation, less economic and more theatrical. 16-bit hardware didn't allow cinematic cameras, and a final boss had to look bigger than anything the player had ever seen, built from sprites alone. The answer was stacking: forms that break apart, transform, grow with each phase. Kefka's final form is a vertical construction of figures that dominates the screen. Lavos, in *Chrono Trigger*, goes from shell to core to true form. And every transformation reopens the resource ledger. The real spike isn't in the last phase's hit points, but in the fact that the player spends magic and items on the first phase without knowing two more are waiting. The difficulty comes from a calculated ignorance. Curiously, the most-cited example of the decade is also the one that least fits the cliché. *Final Fantasy VII* (January 31, 1997) closes with Safer Sephiroth, a long and spectacular fight, but surprisingly soft for anyone who obtained the *Knights of the Round* summon, usually through an optional chocobo quest. It's the flip side of the coin: if the ending is hard for those who rushed and trivial for those who stockpiled, it isn't measuring skill, it's measuring how much time the person had to spend. And then there's *EarthBound*, which inverts the logic. Giygas can't be beaten with numbers. The fight becomes a question about what the player does when the usual tool stops working, and the answer, Paula's prayer, is one of the few times a 1994 final boss uses difficulty to say something instead of merely blocking the way. ## Information was part of the wall too There's a layer that rarely shows up in retrospectives: who knew what. In 1994 there was no guide a click away. There were magazines, printed guides and, in the American case, phone counseling lines run by Nintendo. The design of RPGs from that period implicitly assumed the player had access to that ecosystem or could live with very slow trial and error. For someone in Portugal or Brazil playing an imported cartridge in Japanese or an amateur translation, trial and error was the only route. That changed the meaning of the spike. A boss that took three attempts wasn't a calibrated challenge; it was an entire night, with no way of knowing whether the failure lay in the strategy, the equipment, or a decision made twenty hours earlier. ## Difficulty migrated to the optional Near the end of the decade, the balance shifts. *Final Fantasy VII* already had the Weapons, Ruby and Emerald, as optional adversaries much harder than the final boss. *Final Fantasy VIII* (February 11, 1999, in Japan) took the logic further with Omega Weapon, whose defeat demands a degree of preparation the mandatory ending never asks for. The story's conclusion becomes cinema with a small obstacle; the real challenge lives in dungeons and superbosses that only those who seek them will face. It was a clever solution and, from a certain angle, a surrender. Instead of calibrating a single exam for everyone, the industry started offering several and let the player pick theirs. The price was what the ending stopped being: the mandatory climax lost its threat, and many of those endings began to feel like ceremonies. ## What nostalgia forgets ### The last tenth of the game was the least tested In my reading, there's a factor that rarely enters the conversation: the calendar. The final dungeon and the last boss get built at the end of the cycle, when teams are most tired and deadlines tightest. The first ninety percent of the game goes through many rounds of testing and tuning; the last tenth, through far fewer. I don't have production documents to prove it; it's an inference, not a fact. But there's a hard-to-ignore clue: in 1991, Square itself released in Japan a version called *Final Fantasy IV Easy Type*, and the North American edition of that game, then titled *Final Fantasy II*, was also softened. If the spike were just the genre's noble design, the company wouldn't have needed to tone it down for half the world. ### Where I disagree with nostalgia Nostalgia likes to wrap all this in honor: the games were hard and we were patient. I disagree. Many spikes were simply bad: walls with no clues, built for people who had a guide at their side or a phone number at hand. What deserves remembering is the minority that solved the problem better. *Chrono Trigger* lets the player face Lavos from a certain point in the story, instead of forcing them to wait for the end. Whoever feels ready goes in; whoever doesn't keeps exploring. The ending stops being a wall and becomes a choice. There's also *New Game+*, which turns the second playthrough into a way of tuning the difficulty yourself. That, not the hit-point count, was the smartest thing the genre had in the 90s. ## What the wall left in today's games Today, almost every major JRPG offers a difficulty selector, and many let you retry a fight without losing progress. It seems obvious, but it's a late admission that the 90s design asked the player to guess what the game wanted. At the same time, a part of the audience, the one that grew up with *Dark Souls* (2011), rediscovered the value of the wall, this time with a clear contract: the game warns, the player chooses, defeat teaches. What changed wasn't the hardness. It was the honesty about it. One question remains that I can't close. When the screen asks us today whether we want to play on "Easy," is it correcting an old injustice, or removing the only moment when those forty hours were truly put to the test?