Free time perception test โ stop the clock at 2, 3, 5, 7 and 10 seconds using nothing but your internal sense of time. Your error is measured in milliseconds, averaged over five rounds and ranked worldwide. A one-shot bet mode lets a group take turns on the same target
How to use it
- Press START. The pad shows the first target, "2s". Tap once and the clock starts running.
- The pad turns blue, the stars speed up and "Tap when it feels right!" appears. Count in your head and tap again the moment you believe the target time has passed.
- The reveal shows the real elapsed time to two decimals, a line saying how many milliseconds early or late you were, and an error bar โ centre is the target, left is early, right is late.
- Tap for the next round. Targets grow from 2 to 3, 5, 7 and finally 10 seconds; the dots and bars at the bottom track your progress and per-round accuracy.
- After five rounds the result screen shows your average error in ms, a grade line, accuracy in percent, your best error and each round's figure, with buttons for Try again, Share result and a nickname field to Submit score.
How the score is calculated
Each round's error is the unsigned gap between real and target time, in milliseconds. The average error across the five rounds is your leaderboard score โ lower is better. Accuracy is computed per round as 100 minus the error as a percentage of the target, floored at zero, then averaged: missing 10 s by half a second is a 5 percent error and scores 95, while missing 2 s by the same half second is 25 percent and scores 75. A round chip turns green when the error is within 3 percent of the target (0.3 s on the 10-second round), and the mid-round message "Almost perfect!" appears when you land within 2 percent.
| Accuracy 97% and up | โฐ Human clock |
|---|---|
| 93โ97% | ๐ต Human metronome |
| 85โ93% | ๐ Pretty good sense of time |
| 70โ85% | ๐ Roughly right |
| Under 70% | ๐ Time works differently for you |
Because the ranked figure is an absolute error, the long rounds dominate it: a 5 percent slip costs 100 ms at 2 seconds but 500 ms at 10 seconds. If you are chasing the leaderboard, the 7 and 10-second rounds are where the points are.
What is being measured, and why devices matter less here
Time runs from the instant your first tap is registered to the instant of the second, using the browser's high-resolution timer. Both taps happen on the same device in the same way, so touchscreen input lag and display refresh delay are added to both ends and largely cancel out. That makes this test far less hardware-sensitive than a reaction time test โ what remains is almost entirely your own timekeeping. The drifting stars and the rotating dots on the pad are deliberately non-periodic, so watching the screen gives you no beat to count along to.
Bet mode
The ๐ป Bet mode button on the title screen opens a one-shot bet. Pick 3, 5, 7, 10, 15, 20 or 30 seconds or type any target from 1 to 120 seconds in half-second steps. Each player gets exactly one tap-to-tap attempt; the target, actual time and error are revealed immediately, with "nearly perfect" for an error under 30 ms. Enter a name (up to 10 characters, or leave it blank for "Player 1", "Player 2"โฆ), add the result, and hand the phone to the Next player. The round table sorts everyone by distance from the target. Share result copies the ranking as text together with a link that opens the same bet โ the recipient sees a challenge notice and the bet dialog pre-set to your seconds. Bet results are never posted to the global leaderboard.
Leaderboard and saved records
Your best average error is stored only in this browser. The Global Top 10 on the title screen is sorted by smallest average error; submitting requires a nickname of up to 12 characters and returns your current rank. The server accepts whole-number averages from 1 to 5,000 ms and at most five submissions per minute from one connection. When a friend opens your shared result link, your average appears on their title screen as a "beat it" challenge.
Frequently asked questions
Am I allowed to count in my head?
Yes โ that is the whole point. The test measures your internal clock, and silent counting is part of it. Looking at a real clock or timing to a song's beat would measure something else entirely.
I am consistently early (or late). Why?
Your counting tempo is biased in one direction. Use the error bar and the early/late message after each round to adjust the next one. Nervousness tends to speed the count up; boredom slows it down.
Why is the 10-second round the hardest?
Tempo error accumulates. Counting 20 ms fast per second costs 40 ms at 2 seconds but 200 ms at 10 seconds, and since the ranked score is an absolute millisecond figure, the long round weighs most.
Do bet mode results count for the ranking?
No. Bet mode is for the people in the room; it never touches the global leaderboard or your personal best. Only the standard five-round run with 2, 3, 5, 7 and 10 seconds is ranked.
Related tools
- Reaction Time Test โ tap the instant the screen turns green
- Rhythm Repeat โ hear a four-beat rhythm and tap it back
- Steady Hand Test โ trace a winding path without touching the walls
- Perfect Circle Test โ how round is your freehand circle