What reaction time actually measures
Reaction time is the delay between something appearing and your response to it. In simple lab-style tests, where a single signal appears and you press one button, healthy adults commonly land somewhere around 200 to 300 milliseconds for visual signals. Individuals vary, and results shift with age, alertness, practice and the testing equipment, so any single figure should be read as a rough guide rather than a fixed number.
The picture changes when you have to choose. If the right response depends on what you saw (dodge left or right, block high or low), the time rises, because recognising and deciding takes longer than merely noticing. Most in-game reactions are choices of this kind, so the simple-test figure is closer to a best case than a typical one.
Converting milliseconds to frames
Games are built on frames, so it helps to translate. At 60 frames per second, one frame lasts about 16.7 milliseconds (1000 divided by 60). A 200 millisecond reaction is then roughly 12 frames. A 300 millisecond one is around 18 frames.
This is where design comes in. In fighting games, an attack whose startup is 10 frames lasts about 167 milliseconds before it can hit. That is shorter than a typical simple reaction time, so a player cannot reliably see it begin and then respond. They must anticipate it or already be defending. Slower attacks, say 20 frames or about 333 milliseconds, can be reacted to in principle, which is part of why designers and players talk about moves being “reactable” or not. The details of how startup and recovery are counted are covered in frame data, and the same logic underpins frame advantage.
Why designers care
Human response speed sets a floor that designers work around. Some examples of how it shows up:
- Telegraphs. Boss attacks and special abilities often include a visible or audible wind-up that lasts longer than a typical reaction, so that dodging is fair.
- Hit timing and time to kill. In shooters, time to kill interacts with reaction time: if a weapon kills faster than a player can respond after being spotted, the first shot decides the fight, which changes how the game feels.
- Overheads and fast attacks. Attacks that cannot be reacted to turn a defence into a guess, which connects to mix-ups and the idea of guessing games in mixup guessing.
Designers also take into account that the game sees your input only after display, input device and network delays. Total latency, from the moment an event occurs to the moment your action registers, adds to your own reaction time and can be a noticeable fraction of it. Online play adds more delay, and the way a game handles that, for example with rollback netcode, shapes how fair responses feel.
Why anticipation beats raw speed
If reactions are slow compared with the pace of play, then strong players must be doing something other than reacting. In most games, they are predicting. They know which events usually follow which, so they begin the response before the signal arrives.
Examples include:
- positioning to cover the likely approach instead of chasing it
- recognising the early cues of an attack, such as a character animation or a sound, rather than waiting for the hit
- remembering what an opponent tends to do in a situation, a point developed under opponent habits
- using scouting information and knowledge of timings, such as when a cooldown returns, to be ready at the right moment
This is why a veteran with slower measured reflexes can beat a younger player with faster ones. Reaction time tests single responses, while expertise reduces the number of situations where a response must be a pure reaction.
Can reaction time be trained?
Somewhat, with limits. Practice on a specific task usually improves speed on that task, partly because recognition gets quicker and partly because responses become more automatic. Gains in raw, general reflex speed from games or training are smaller and less certain, and claims about large improvements deserve scepticism. Sleep, alertness, and avoiding fatigue also affect speed from one session to another.
Because the transferable gains are mostly in recognition, the useful kinds of practice are specific:
- drill the particular cue and response you struggle with, as in deliberate practice
- review moments where you reacted late and ask what cue you could have used earlier
- learn what the common options are in a situation so that fewer choices must be made on the spot
Testing yourself sensibly
Online reaction tests are fine for curiosity, but their results depend on your screen refresh rate, input device and browser, so they are not precise. Compare yourself to your own earlier results on the same setup, not to other people. Take several attempts and look at the typical value instead of the best single one, since single attempts vary a lot.
The practical takeaway
Treat reaction time as a constraint, not a skill to obsess over. Know roughly how fast you respond, understand what that implies for the game you play, and put most of your effort into reading situations early. Faster hardware and lower latency can help at the margin, but better prediction usually gives larger returns.