Why speedrunners love broken physics

Ask a speedrunner what their favorite moment in a game is, and there's a decent chance the answer involves something the developers never intended to happen: a wall that can be clipped through at a specific angle, a jump that launches a character three zones ahead of where the story expects them, a frame-perfect input that turns a loading transition into a shortcut. To outsiders, this looks like the game breaking. To the community that's built an entire culture around it, it's the game's real hidden layer, and finding it is the point.

A bug is just an undocumented feature

The speedrunning community's relationship to glitches starts from a simple reframe: a game's physics engine doesn't know the difference between "intended" and "unintended" behavior, it just applies the same rules consistently to everything that happens inside it. Rocket jumping, famously, started as players noticing that explosion knockback could be aimed at themselves for extra height and distance, and developers eventually leaned into it rather than patching it out, because it turned out to be fun on its own terms. Most modern glitch-hunting follows that same logic in reverse: find the edge of the rules, then find out what the edge actually lets you do.

This is why the best speedrunning communities treat glitch discovery less like bug reporting and more like a shared research project. Dedicated routing teams will spend months mapping exactly which wall geometries in a given game allow a clip, cataloguing the precise pixel or frame windows where an input needs to land, and building community wikis that read less like strategy guides and more like engineering documentation for a system nobody designed on purpose.

Consistency is what makes it beautiful

The games that develop the richest speedrunning scenes tend to share one trait that has nothing to do with genre: a physics engine that behaves with total, predictable consistency, even when that consistency produces strange results. A game where the same input sequence reliably produces the same outcome, frame after frame, run after run, gives a community something to build mastery against. A game with inconsistent or randomized physics, by contrast, usually struggles to develop the same depth of competitive routing, because there's nothing stable enough to optimize toward.

• Rigid, deterministic physics let a glitch be reproduced reliably, which is what turns a fluke into a technique.
• Sequence breaks reward players who understand a game's internal logic better than its map was designed to allow.
• Frame-perfect tricks become a shared skill ceiling the whole community trains toward together.
• Developer reactions vary widely, from quietly leaving glitches in to actively building new content around community-discovered tech.

When developers start listening back

Some of the more interesting recent history in this space involves developers actively engaging with what their speedrunning communities find, rather than treating every discovery as a patch target. A handful of well-known sequence breaks have survived years of subsequent patches specifically because the developers decided the trick was more interesting to leave in than to fix, and a few studios have gone as far as adding developer commentary or in-game acknowledgment of community-discovered techniques. It's a small but genuine sign of games being treated as living systems that a community can meaningfully shape after release, rather than fixed objects a studio simply ships and walks away from.

That relationship cuts both ways, though. Plenty of glitches do get patched, sometimes to the visible frustration of a community that had built years of competitive history around a specific trick, and the tension between "preserve the game as a historical object" and "keep patching for a healthier general audience" is a genuinely unresolved argument in most long-running speedrunning scenes.

The appeal, stripped down

At its core, the appeal of broken physics in speedrunning isn't really about breaking anything. It's about discovering that a game's rules are more specific, more mechanical, and more discoverable than the surface-level experience suggests, and that mastering those rules is its own distinct skill from simply playing the game as intended. The wall you clip through isn't a failure of the level design. It's proof the level had edges in the first place, and someone finally found them.

The social layer underneath the technical one

What outsiders often miss is how collaborative this discovery process actually is. The fastest routes through most heavily-run games weren't found by a single player in isolation; they emerged from years of community theorycrafting, shared frame-data spreadsheets, and a kind of open peer review where one runner's new discovery gets immediately stress-tested by dozens of others trying to break it, improve it, or find an even faster alternative. Watching that process unfold on a community Discord server, with dozens of people arguing over whether a half-frame window counts as "consistent" enough to add to the official route, is its own strange kind of collaborative science.

That culture also produces a distinct kind of credit system, separate from the leaderboard times themselves. Finding a genuinely new technique, even one that only saves a fraction of a second, tends to carry more community prestige than simply having the fastest overall time using tricks other people found first. It's a reminder that the appeal here isn't purely competitive; a meaningful part of it is closer to puzzle-solving, where the solution itself is the reward independent of where it lands you on a leaderboard.

Why this keeps spreading to new games

As more games ship with built-in replay tools, frame-by-frame capture, and active modding communities, the barrier to this kind of glitch archaeology keeps dropping, and newer releases are developing serious speedrunning scenes faster than older games ever did. It's not unusual now for a dedicated routing community to form within weeks of a major release, sometimes before the game's story has even been fully spoiled online, which says something about how much of the appeal has shifted from "exploit something rare" to "be first to map a system everyone knows is there." The instinct hasn't changed. The tools, and the speed at which a community can organize around them, clearly have.