Stop at the right second
Hold, and let go when you think you have reached the target. Nothing is counting for you.
Hold, and let go when you think you have reached the target. Nothing is counting for you.
You get a targetBetween 4 and 12 seconds
Hold and waitNo clock, no counter, no bar
Five goesTotal error is the score
You are shown a target — 7.384 seconds, say — and you press and hold. From that moment the target dims and the screen gives you nothing: no counter, no progress bar, no ticking. Let go when you believe you have arrived. The gap between your guess and the target, in milliseconds, is your error for that go.
There are five goes and the score is the total error, so lower is better. Targets run from four to twelve seconds and are never round numbers, which is deliberate: 7.384 cannot be counted out in whole beats, so the last fraction has to be felt.
Everyone in the world gets the same five targets today, and they change once a day. Knowing them in advance is no help — the whole difficulty is in the body, not in the number.
A total under 40ms across five goes is rejected: that is a stopwatch, not a sense of time.
Estimating an interval of a few seconds without counting is a well-studied task, and people are systematically bad at it in a particular way: short intervals get overestimated and long ones underestimated, an effect steady enough to have a name — Vierordt's law. The crossover for most people sits somewhere around two or three seconds, which is why the targets here start above it.
Counting out loud helps, and everybody discovers that within two goes. It also caps you: a count is accurate to about a tenth of a second, while a practised feel for a specific duration can beat it. Whether to count is the only real decision in the game.
A progress bar or a running clock would turn this into a reading test — you would watch a number approach and release on it, which measures reaction time, something the site already has a page for. Removing every cue is what leaves the estimate where it belongs.
The server does not take your word for the wall-clock time either. It knows when the round started and when the result arrived, and a run that claims to have held for thirty seconds inside a three-second round is not counted.