How to play
The big pad in the middle of the page is the whole game. Click it once and the session arms itself: the pad turns amber and holds a message telling you to wait. Somewhere between roughly one and a half and three and a half seconds later it flips to green, and the clock starts on that exact frame. The number you get is the gap between the pad turning green and your click landing, rounded to the nearest millisecond.
The delay before the green light is drawn fresh every time from your browser's cryptographic random number generator, so you cannot learn a rhythm and ride it. Five valid attempts finish a session. Each result is added to the row of chips under the pad, and the boxes above the pad update as you go, so you can watch your average settle instead of waiting for a single verdict at the end.
An early click — any click while the pad is still amber — cancels the countdown, marks the attempt as a false start and puts it back in the queue. You are never punished with a score for it, but you never gain from it either, which is the point. When the five attempts are done, the page stores your fastest single click and your best session average on this device only, and the Try again button clears the board for a fresh set.
Controls
- Mouse or trackpad: click anywhere on the pad, both to arm an attempt and to stop the clock.
- Keyboard: Space or Enter behaves exactly like a click, whether or not the pad has focus.
- Touch: tap the pad. A tap is registered the same way a click is, though a touchscreen usually reports it slightly later than a mouse does.
- Try again: throws away the current session, including any half-finished attempt, and starts over at attempt one.
What the number is actually made of
The figure on the screen is not a measurement of your nervous system. It is the total time taken by a chain of separate events, and only part of that chain belongs to you. First the pad has to actually appear: the browser paints the green frame, and your monitor then has to show it. A display that refreshes sixty times a second only has an opportunity to change every sixteen milliseconds or so, and the paint almost never lands exactly on that boundary, so the light reaches your eye a little after the code thinks it did.
Then the part that is yours. Light hits the retina, the signal travels toward the visual cortex, and something in there has to decide that this particular change is the one you were waiting for. That decision is not free — it is the single biggest chunk of the whole number, and it is also the part most affected by how tired you are, how much you expected the change, and whether you were looking directly at the pad or slightly past it. Only after the decision does a motor command travel down your arm to a finger that then has to physically move far enough to close a switch.
Finally the hardware on the way back in. A mouse that polls at one hundred and twenty-five hertz can sit on your click for up to eight milliseconds before it tells the computer about it, and the operating system, the browser and the page's own event loop each add a little on top. That is why the same person gets noticeably different numbers on a gaming setup and on an office laptop, and why a wireless mouse on a busy machine can read slower than the person using it. It is also why this test is a rough instrument: it is measuring you and your equipment together, and it cannot tell you which part of the result came from which.
The practical consequence is that comparison is only meaningful against yourself, on the same machine, in similar conditions. Median matters more than best for exactly this reason. Your fastest single click of five is partly luck about where the green frame fell relative to your screen's refresh; the median throws away the lucky one and the unlucky one and tells you what your typical attempt looked like.
Every millisecond figure on this page is an estimate from our own play testing, not a verified statistic. In our sessions, attentive adults on a mouse tend to land in the low to middle 200s, touchscreens usually read a little slower, and a session average below 200 almost always means the player was anticipating rather than reacting. This is a browser toy for curiosity and practice. It is not a clinical or diagnostic tool, it does not assess anyone's health or fitness to do anything, and no result here should be read as medical information.
A very old experiment
Measuring how long a person takes to respond to a signal is one of the oldest experiments in psychology, and simple visual reaction tests have been in use since the nineteenth century. The early versions were mechanical and fussy: a stimulus was presented, a subject released a key, and the apparatus recorded the interval. The interesting finding was not any single number but the structure behind it — that responding to a light takes measurably longer when you first have to decide which light it was, which is how researchers began separating the raw response from the decision inside it.
What we have built here is the simplest possible version of that idea, with one signal and one response, written from scratch for this site. It keeps the shape of the classic experiment while being completely honest about the fact that a web page in a browser tab sits several layers away from the careful conditions a laboratory would insist on.
Questions players ask
Why is my result slower here than on another reaction test?
Different tests start their clock at different moments and end it on different events, and your own hardware adds delay that no page can subtract. Compare your numbers against your own earlier sessions on the same machine rather than against a number from a different site.
What counts as a good reaction time on a mouse click?
From our own testing, most attentive adults on a mouse land somewhere in the low to middle 200s of milliseconds. These are estimates from our play testing, not verified statistics, and this page is not a clinical or diagnostic tool.
Can I get a good score by clicking constantly?
No. A click that arrives while the pad is amber cancels the pending timer and repeats the attempt, so a click can never be sitting there waiting for the green frame. A short lockout follows each false start, and any result under 80 milliseconds is discarded as anticipation.