Why this cannot be an eye test
Visual acuity has a specific meaning: the smallest angular detail you can resolve, measured with optotypes of a standardised size at a standardised distance under standardised lighting. Every part of that definition is a physical control, and a web page has none of them. It does not know the physical size of your display, only its pixel count. It does not know the pixel density, so it cannot convert pixels to millimetres. It does not know your viewing distance, which changes the angle subtended by everything on screen and is the single largest variable. It does not know your screen brightness, your ambient light, or whether you are wearing correction.
The practical consequence is easy to demonstrate. Run this at 0.3 seconds sitting normally, then lean six inches closer and run it again. Your score will improve and your eyes will not have changed at all. Any page that converts a browser result into a 20/20-style figure is inventing the distance and the physical size, and the number it prints is fiction. This page does not print one.
If you have a reason to think your vision has changed — new blur, headaches, trouble with night driving, one eye noticeably different from the other — that is a reason to see an optometrist, not to run a browser game. A full sight test also looks at eye pressure and retinal health, neither of which has anything to do with whether you can read a chart.
What it does measure, roughly
Brief-exposure recall: how much of a display you can extract and hold long enough to type back after it is gone. That is a real thing and it has been studied since Sperling's partial-report experiments in the 1960s, which showed that the visual system briefly holds far more than people can report, and that the report is limited by how fast the contents decay rather than by how much was seen. At 0.3 seconds you get roughly one fixation, which is why longer numbers become disproportionately harder — there is no time to move your eyes across them.
The dot-counting rounds probe something different. Up to about four items, people report the count immediately and without counting; this is called subitizing and it is fast and near-perfect. Past four or five, accuracy and time both degrade sharply because counting has to start, and a third of a second is not enough time to count. The rounds here span three to seven dots deliberately, so they straddle that boundary and most people find the 6 and 7 rounds much harder than the 3 and 4 rounds regardless of exposure setting.
Reading your own runs
| Comparison | Worth anything? |
|---|---|
| Your run today against your run yesterday, same machine, same setting | Yes, modestly — mostly reflects attention and fatigue |
| Your 1.0 s score against your 0.3 s score | Yes — the gap shows how much you rely on eye movement |
| Your digit score against your dot score | Yes — different mechanisms, and people are often lopsided |
| Your score against a friend's on a different device | No — you are comparing screens and seating positions |
| Your score against any published norm | No — there are no norms for an uncontrolled browser task |
Strategy, since you will look for one
At the shortest exposure the winning approach for digit rounds is to fixate the centre of where the number will appear and take the whole string as a shape rather than reading left to right. Reading left to right guarantees you lose the last digits, because the display is gone before your eyes get there. For dot rounds, resist the urge to count; try to see the arrangement as groups of two and three and add them, which is what people who are good at this are doing whether they know it or not.
None of this improves your vision. It improves your score at this particular task, which is precisely the reason a score from this task should not be read as a statement about your eyes.
Questions people ask
Can I use this to check whether I need glasses?
No, and it would be a bad idea to try. The task depends on your screen size, resolution, pixel density, brightness, ambient lighting and viewing distance, none of which the page can detect and all of which move the result more than a mild refractive error would. Someone with uncorrected vision sitting close to a large bright monitor will outscore someone with perfect vision squinting at a dim laptop. If you have any reason to suspect a change in your sight, book a sight test.
Why do the dot rounds feel so much harder above five?
Because a different mechanism takes over. Three or four items are apprehended in one go without counting, a phenomenon called subitizing, and it is fast and accurate. Above four or five you have to count, and counting takes far longer than the exposure allows. The step in difficulty between four dots and six dots is a real and well documented feature of human vision, not a flaw in the test.
Does the exposure time actually match what I chose?
Approximately. The page asks the browser to hide the display after the chosen interval, but the actual time on screen is quantised to the display refresh — about 16.7 milliseconds on a 60 Hz monitor — and can stretch if the browser is busy. At the 0.3 second setting that error is a few percent, which does not matter for comparing your own runs but is another reason the result is not a measurement.
Is anything stored?
No. The rounds and the score exist in the page while it is open and are lost on reload. Nothing is sent anywhere and there is no history to clear.