Speedrunning looks impenetrable from the outside — world records built on frame-perfect inputs and glitches that seem to require reverse-engineering the game. But the distance between watching a run and completing your own first one is far shorter than the leaderboards suggest.
Choosing your first game
Pick a game you already know and genuinely love. This matters more than any other decision, because speedrunning means playing the same game hundreds of times — running something you merely tolerate becomes a chore within a week, while running something you love feels like deepening a relationship you already had.
Shorter games make better starting points, simply because a full attempt takes minutes rather than an entire evening. A game with a thirty-minute run lets you make ten attempts in a session and feel real improvement; a four-hour run means a single mistake costs you the whole evening. Many runners start with a short indie game precisely for this reason.
Understanding categories
Every speedrun exists within a category, and reading the rules before your first attempt saves a disqualified run later:
- Any% — reach the ending as fast as possible, using any glitch or skip available. Usually the fastest and most spectacular category.
- 100% — complete everything the category defines as completion. Longer, more methodical, and often more approachable because it relies less on difficult glitches.
- Glitchless — finish quickly without exploiting bugs, which appeals to players who want the challenge to be execution rather than exploit knowledge.
- Category extensions — community-invented challenges for games whose main categories are crowded, and frequently the friendliest place for a newcomer to set a respectable time.
Speedrun.com hosts leaderboards, rules, and community guides for essentially every game with an active scene. Reading your chosen game’s rules page is the mandatory first step.
The tools you actually need
Far fewer than expected. A timer is essential — LiveSplit is the community standard, free, and splits your run into segments so you can see exactly where you lost time compared to your best attempt. That segment data is what turns aimless repetition into deliberate improvement.
Beyond a timer, you need a route: the sequence of actions a run follows. Watch a couple of runs by mid-tier players rather than only the world record holder, since their routes are usually simpler and their commentary more explanatory. Many communities also maintain written route guides, which are easier to reference mid-practice than a video.
How improvement actually happens
Your first runs will be slow, and this is not a sign you chose wrong. Every runner’s first attempt looked like yours. Improvement comes in a predictable sequence: first you learn the route, then you stop making major mistakes, then you begin optimising individual segments, and only much later do frame-perfect tricks become relevant.
The habit that accelerates all of it is practising segments in isolation rather than always running from the start. If a particular section costs you thirty seconds every attempt, practising just that section twenty times improves your runs far faster than twenty full attempts would.
The community is the actual draw
Most people who try speedrunning stay for the community rather than the times. Speedrun communities are unusually welcoming to newcomers, partly because every member remembers being slow, and partly because a game’s scene only survives if new runners keep arriving. Joining the game’s Discord and asking questions is normal and expected, not an imposition.
Submitting a run that lands nowhere near the top of the leaderboard is also entirely normal. Leaderboards have many pages, and the satisfaction of seeing your own time recorded — then beating it — is the actual hobby. The world record is somebody else’s project.
Frequently asked questions
Do I need to record my runs?
For casual practice, no. For submitting to a leaderboard, most games require video proof, particularly for higher placements. Free capture software handles this easily, and recording also lets you review your own mistakes, which is valuable regardless of submission.
Is using glitches cheating?
Not within a category that permits them — the categories exist precisely to separate different rule sets. Using glitches in a glitchless category is cheating; using them in Any% is the point. The rules page defines the game you are playing.
Can I speedrun on a controller, or do I need keyboard and mouse?
Whatever the game supports. Some communities track input methods separately, but most do not care. Use what you are comfortable with — comfort over hundreds of attempts matters more than a theoretical advantage.