What constitutes a 'glitch' in a speedrun, how are they discovered and executed, and what is their technical impact on the game engine? Additionally, what are your thoughts on the ethical boundaries and community rules surrounding glitch usage? In which cases are glitches acceptable, and when do they cross the line into spoiling the challenge of the run? If anyone has experience with this, I’d love to hear your approach and any advice you might have.
What constitutes the use of 'glitches' in speedrunning, and what are the ethical boundaries?
👁️ 13 views💬 5 replies❤️ 0 likes
5 Replies
Thanks for opening this topic, it helps me better understand how glitches work in speedruns. Has anyone tried using any memory capture tools to detect glitches without breaking the game's integrity?
A few months ago, I decided to tackle the "Portal 2" co-op speedrun record, and that's when I discovered the infamous "crouch-boost," which lets you jump higher than intended. I stumbled upon it by accident while trying to speed up the exit of the first test chamber. I noticed that if I crouched just before the boost pad, the character gained an extra push. After testing it multiple times and logging the physics values, I incorporated it into my route, shaving off nearly 3 seconds—enough to go from an average runner to a top 10 contender.
Back then, I debated with several members of my clan whether to include it in the official category. Most agreed that while the glitch is documented and reproducible, it goes against the level’s intended design, so they’d classify it as "illegal." In the end, I posted it under the "Any% Glitchless" category as a side note, keeping the "Any%" version without the jump to avoid contaminating the official leaderboard. My advice is always to check the community rules before using a glitch—if the forum has a list of allowed techniques, stick to it. When a trick is widely accepted (like the "sokoban" in "The Legend of Zelda: Ocarina of Time"), there’s no issue. But if the advantage comes from an unintended exploit the developers didn’t plan for, the most honest approach is to separate it and keep it out of "clean" rankings.
I'm still the type of player who thinks a glitch is just a typo in the code, so I probably just mash buttons and hope the physics break themselves 😂. If I ever accidentally wall‑clip, I'll be the first to claim it's a brand‑new speedrun category! 🎮🤦♂️
Speedruns' glitches can, in a way, be compared to the "debug shortcuts" we use in VR development to bypass engine bottlenecks. In both cases, the goal is to exploit an undocumented behavior or vulnerability in the game engine: in speedrunning, the player discovers and incorporates it into their route; in development, the programmer temporarily enables it to test a scene without going through the entire pipeline. The key difference lies in the context: while in development the objective is optimization and the "glitch" is an internal tool, in speedrunning the audience judges the final time as a competitive merit, so the use of these exploits is regulated like any other category rule.
From an ethical standpoint, accepted glitches are usually those any player could reproduce with the same knowledge and without modifying the binary, much like classic cheat codes that only require pressing a button combination. On the other hand, exploits that need external patches or additional hardware (such as script injection or memory modification) break the level playing field and are considered disqualifiers, similar to how the VR community disallows "cheat engines" in public demos. My recommendation is to draw a clear line: if the glitch is part of the game as it shipped and is documented by the community, it’s acceptable; if it requires altering the code or execution, it’s best to run it in a "glitchless" category or avoid it altogether.
In my case, the first time I encountered a glitch that truly altered the pace of a speedrun was during a *The Legend of Zelda: Breath of the Wild* marathon. While testing different routes to the *Master Sword*, I discovered that throwing the shield just before entering an enemy-heavy area would cause the game to "reboot" the zone without a full reload, saving me several seconds in transition. Technically, the game world engine reloads assets via a soft-reset of the collision bucket; the zone becomes "empty," and the character reappears instantly. It took a few attempts and a couple of community videos to refine the timing, but once mastered, the advantage was undeniable.
From an ethical standpoint, the speedrunning community generally accepts glitches that are "intentional" or "discovered" without altering the game's code, as long as they don’t violate event rules (e.g., using external tools or modifying the binary). In my case, the shield glitch stayed within bounds because it didn’t require file editing—just precise execution. However, any exploit involving RAM manipulation or external scripts is considered cheating and, in my experience, gets immediate backlash. My advice is to always check tournament rules and, if in doubt, ask in the forums before including a glitch in your run; transparency keeps the sport’s integrity intact and prevents unnecessary division in the community.