Lights Out

What this one actually measures

This is the cleanest test in the battery: one signal, one response, nothing to decide. The lights come on one by one, hold for a moment, and go dark — and the going-dark is your cue. Real racing starts work the same way, which is why drivers obsess over it: the start is the only part of the race where pure reaction time decides anything.

Because there is nothing to think about, there is also nothing to blame. Whatever number comes out is close to the floor of how fast you can turn seeing into moving.

Why the hold is random

If the gap between the last light and lights-out were fixed, you would stop reacting and start counting. Everyone does, within about three attempts, without noticing. So the hold is drawn fresh each time, between 0.2 and 3 seconds. The only strategy that survives that is the honest one: wait, watch, go.

Press early and the attempt is simply thrown away. We do not end your run over it — a jump start already tells us the number would have been a guess, and guesses are not worth punishing twice.

Where jump starts come from

A jump start is not a judgement error, because it happens below the level where judgement operates. While you wait, your motor system is not idle — it is charging up, leaning on a door it is about to open. The longer the hold drags, the more certain your body becomes that the signal must be due, and sometimes the door just opens. That is why most early presses come on the long holds, not the short ones, and why they feel so weirdly involuntary. They are.

Which points at the only fix that works: stop predicting. An extra half-second of red carries no information here — the delay is drawn fresh every round — so treat it as none. People who make peace with the wait jump less than people who try to outsmart it. And if you want to feel what stakes do to that peace, the same instinct with someone across from you is Showdown — there an early press does not cost you a redo. It costs you the round.

What a decent number looks like

Most people land somewhere between 220 and 300 ms on a computer. Phones tend to read slower — not because your reflexes drop when you pick one up, but because touchscreens add their own delay before we ever hear about your tap. That is why results are ranked per input device and never mixed.

Track and field settled a related argument years ago: sprinting rules treat any reaction inside 100 ms of the gun as a false start, because nobody genuinely hears and moves that fast. We use a similar floor here — anything under 80 ms is discarded as a guess, however good it looks.

Can you actually get faster?

A little. Mostly not where you expect. Your floor — the fastest press you are physically capable of — is set by things practice does not reach: nerve length, muscle, age. Decades of lab work agree that training moves simple reaction time by maybe a few hundredths, not the big satisfying chunks people hope for.

What does move is everything around the floor. Sleep moves it. Caffeine moves it, briefly and honestly. A proper warm-up — three or four throwaway rounds before the ones you count — moves it more than people admit. And practice, while it barely touches your fastest press, visibly cuts the spread: fewer 350s scattered between your 250s. That is why the number worth chasing on this page is your average, not your best. A best is one lucky alignment; a tight average is a skill.

One more honest note: a session on a website measures seeing-to-moving today, on this screen, with this much sleep. It is a snapshot, not a verdict — which is exactly why the shape across all the tests says more about you than any single number in one of them.

One tip that actually helps

Do not stare at a single lamp. Soften your gaze so the whole row sits in view, and react to the change rather than to any one bulb. And once you have a number you like, try Free Fall — same reflexes, but with eight places to watch instead of one. The difference between those two scores is the cost of choice.

Further reading: the 100 ms rule is covered under false starts; for how reaction time is measured at all, see mental chronometry and our own methodology page.