Everyone copied permadeath and procedural rooms. Almost nobody copied the part that actually mattered.
The rogue-like genre had a moment. Then the moment became a decade. Then every third indie game on the market had procedurally generated rooms and permadeath slapped onto whatever core loop the developer had already built. The genre exploded because a handful of games proved that randomness plus stakes could be wildly compelling. What followed was a wave of games that understood the ingredients but missed the recipe entirely. And right now, players are starting to feel it.
The best rogue-likes aren't compelling because they're random. They're compelling because every run is a complete arc: a beginning, a string of consequential decisions, an ending. The death is the period at the end of the sentence, not a punishment for failing a test. Players go again not because they want to try harder, but because they want to try differently. The run gave them something to chew on: a build they almost pulled off, a decision they'd make differently next time.
That's the real engine. Not procedural generation. Not permadeath. The feeling that each attempt was a story worth running again from a different angle. When a game hits that, it's almost impossible to stop playing. When a game misses it, no amount of randomized room order fixes the problem.
Procedural generation is a tool, not a feature. It's useful for extending replayability, for making sure no two runs are structurally identical. But a randomly generated dungeon full of interchangeable encounters isn't variety. It's just a shuffled deck of the same cards. Players figure this out fast. Once they've seen every enemy type and every room layout, the "randomness" stops feeling dynamic and starts feeling thin.
Permadeath has the same problem. Death only has weight if the run had meaning. If the core loop is shallow, if every room is just a damage-sponge fight and the decisions between rooms are trivial, dying doesn't feel like the end of a story. It feels like a loading screen you didn't consent to. Copy the mechanics without building the loop that gives them weight, and you get a game that's technically a rogue-like and spiritually just annoying.
The smartest thing the genre figured out over the last decade is meta-progression, the layer of advancement that survives your failed runs. Done well, it solves permadeath's biggest friction point: the feeling that your time was wasted. When failure still pushes something forward, the run keeps its value even if you never beat the final boss. You unlocked something. You learned something. You're slightly different than you were when you started.
The nuance that most games get wrong is that meta-progression should change how you play, not just how strong you are. Unlocking a new starting item that enables a build you've never tried before is design. Unlocking a flat 5% damage bonus is just grinding with extra steps. The former changes the run. The latter changes a number. Players feel the difference, even if they can't articulate it.
The market is oversaturated, and players have gotten good at spotting the formula. A new rogue-like launch used to carry a promise: this might be the next great one. Now the default assumption is skepticism. Players have been burned too many times by procedurally generated time-wasters dressed up in a striking art style. The genre has a credibility problem it didn't earn in its best years but absolutely earned in its worst ones.
The breakout games still break out, the ones with a distinctive core loop, a fresh theme, a genuinely original take on run structure. But the middle of the market is getting squeezed hard. A competent rogue-like with no identity used to be a reasonable bet. In 2026 it barely gets wishlisted. The bar for "give me your time" just got a lot higher.
If you're building in this space, the question isn't "should I add permadeath?" The real question is what your core loop actually rewards. If the answer is expressing a build identity, or making consequential decisions under pressure, or learning a system well enough to exploit it, then the rogue-like structure might genuinely serve you. If the answer is killing enemies and getting loot, it probably won't save you.
The genre isn't dying. The best versions of it will keep getting made and played. What's dying is the assumption that the format is a shortcut to replayability. It never was. It was always just a frame, and what goes inside that frame is still entirely on you. Design the loop first. The structure is just the container you pour it into.
The Guy in Cube13
Playable Ideas · Cube 13 Gaming