The emperor died. Hours of careful progress, gone. The player had built that character, chosen those skills, watched the numbers rise. And then a battle went wrong and the screen showed a succession ceremony. A new emperor took the throne. The skills remained. The person did not. Some players closed the cartridge and never returned. Others kept playing, confused at first, then curious. This was Romancing SaGa 2, released in 1993. It was designed to feel like no role-playing game before it.
Akitoshi Kawazu joined Square in 1985, a year or two before the company released the game that would save it from bankruptcy. His first assignment was graphics for a racing game. The following year he was moved to battle design for a project called Final Fantasy — a title chosen because its creator, Hironobu Sakaguchi, planned to leave the industry if it failed. It did not fail. The game sold enough to keep Square alive. Kawazu stayed on to design battles for Final Fantasy II in 1988. Those two games taught him the architecture of turn-based combat. They also showed him where the walls were.
Most role-playing games at the time followed a single path. The player controlled a party, followed a story, leveled up through predictable stages, and reached an ending. The structure was stable and proven. Kawazu had learned how to build it. But he had also started to wonder what lay outside it. In 1989, he was given a chance to find out. Square asked him to design a game for the Game Boy, a portable system with a tiny screen and limited processing power. The result was Makai Toushi SaGa, released in North America as The Final Fantasy Legend.
The game was experimental by design. Instead of an open world, it was structured as a vertical tower. Characters could die permanently and be replaced mid-game. The leveling system behaved differently depending on whether a character was human, mutant, or monster. It broke the expected rhythm of role-playing games, and it sold over a million copies — Square's first title to cross that threshold. Kawazu had created a new series. He followed it with SaGa 2 in 1990, refining the systems. Then he moved on to something larger.
Romancing SaGa, released in 1992 for the Super Famicom, introduced the structure that would define his career. Eight playable protagonists, each with their own starting point and intersecting stories. A non-linear quest flow where player choices determined which events became available. A battle system built on weapon durability and techniques learned through repeated use. The game was commercial and creative. It was also divisive. Some players felt lost without a single narrative path. Others felt liberated. Kawazu had built a game that resisted being played the same way twice.
Romancing SaGa 2, released in 1993, pushed further. It introduced a generational system. The player did not control a single hero across the entire game. They controlled a lineage of emperors, spanning centuries. Each ruler lived, fought, and eventually died or abdicated. Skills and territory carried forward to the next emperor. The individual did not. It was a game about legacy, not triumph. About what remains when the person who started the journey is gone. Critics called it brilliant. They also called it obtuse, unforgiving, visionary, and frustrating — sometimes all at once.
Kawazu continued building games this way. Romancing SaGa 3 in 1995. SaGa Frontier in 1997, with seven protagonists and separate scenarios. SaGa Frontier 2 in 1999, experimenting with watercolor visuals and interwoven generational narratives. Each game tested a different structural idea. Each one divided its audience. In interviews across the years, Kawazu has repeated a consistent philosophy: every game should try something new. Limitations are creative opportunities. Players should have the freedom to fail. He has said he was influenced by tabletop role-playing games like Dungeons & Dragons and Wizardry, where the rules created possibility rather than a fixed path.
As of 2026, Kawazu remains active at Square Enix. The SaGa series celebrated its 35th anniversary in 2024. In a recent interview, he said he wants each game to feel like something the player has never experienced before — even if they have played every previous entry. That commitment to unfamiliarity is what makes his work difficult to describe and difficult to forget. He has spent four decades building games that value discovery over guidance, where the path forward is not always clear and the destination is not always the same. Not every player chooses that kind of game. But for those who do, his work has been a quiet light — not the brightest in the room, but steady, and never asking to be anything other than what it is.
The road you are walking now, where no one is holding your hand and the map does not show every turn — is that cruelty, or is it trust?