Measured in level reached ยท higher is better
Visual Memory Test
Squares flash, then vanish. Click the ones that lit up. The grid grows as you go.
Before you start
The median on mouse is 9 level. Run the test and we will place you against it.
Compared against
Grid size adapts to your screen, so a phone and a monitor show the same pattern at different sizes. Ranked separately.
The other device classes
Trackpad
Touchscreen
Curve is a fitted baseline plus 9 first attempts recorded here. The baseline is a curve, not a sample: no part of it is a real person. Baseline is an estimate: centred on a plausible span for a growing grid recall task. No published dataset was found for it. Per-device offsets are extrapolated, not measured. The baseline is dropped once a device class holds 5,000 real first attempts. See the full distribution.
Remember the pattern
Squares flash for about a second, then vanish. Click the ones that lit up. The grid grows every time you get it right.
Understanding your score
| Band | level | What it means |
|---|---|---|
| Exceptional | 16 and above | Rare. Usually a deliberate chunking strategy. |
| Very good | 13 to 15 | Well above average. |
| Good | 11 to 12 | Comfortably above the middle. |
| Average | 8 to 10 | Where most results land. |
| Below average | 5 to 7 | Common when tired or rushing. |
| Low | 4 and below | Slow down. Most low scores here are haste. |
What is being measured
This is visual working memory: the small amount of visual information you can hold in an active, usable state for a few seconds. It is famously limited. Most people can hold three to five distinct items reliably, and performance falls off sharply past that.
The reason results here reach level 9 or higher rather than stopping at 5 is chunking. A pattern of eight squares is impossible to hold as eight separate positions and quite manageable as two shapes. What this test really measures is how quickly you convert a scatter of squares into structure.
What actually improves the score
Look for shapes, not positions. Three squares in a line is one thing to remember. Three coordinates is three.
Take the full flash time. The pattern is visible for nearly a second, and most people stop looking after a third of it because they think they have it.
Do not rush the reproduction. The clock is not running during your answer, and hurried clicking costs more lives than a bad memory does.
Expect the first two or three rounds to be easy and the difficulty to arrive suddenly. The grid grows on every success, so the point where you fail is genuinely near your limit.
Why this does not predict much
Working memory span correlates with some measures of reasoning, but the relationship is weak and it does not travel well. Being good at remembering flashing squares does not make you good at remembering faces, names, or where you left your keys, which are all different systems.
Practice improves your score on this specific task quickly and then stops transferring. That pattern, large gains on the trained task and none elsewhere, is the consistent finding across the brain training literature.
Why the flash duration matters
The pattern stays visible for nearly a second, which is far longer than it needs to be to see it and roughly the time it takes to organise it. Those are different tasks. Seeing eight squares takes a fraction of a second. Converting them into two remembered shapes takes most of the rest.
A very brief flash would test iconic memory, the high-capacity but extremely short-lived visual buffer that fades within about half a second. That buffer holds far more than you can report, which is precisely why it is measured differently and why this test does not use it.
What this test measures instead is what survives the transfer from that buffer into working memory. The bottleneck is the transfer, not the seeing.
How this relates to the clinical span tests
The closest established measure is the Corsi block-tapping task, where an examiner taps a sequence on a set of blocks and the subject reproduces it. The sequence grows until it fails, and typical adult spans land around five. It is used clinically alongside digit span, and the pair are informative because they load different systems: one visuospatial, one verbal.
This test differs in one important way. Corsi is sequential, so order matters and it recruits the same serial rehearsal machinery that digit span uses. Here the squares flash together and order is irrelevant, which makes it closer to a change-detection task: how many items you can hold simultaneously, not how long a sequence you can replay.
That is why the numbers here run higher than a Corsi span, and why they are not interchangeable. Neither is a substitute for the other, and neither is a clinical instrument when run in a browser with no supervision, no practice control, and no way to know whether you were distracted.
What a low score does and does not mean
The single most common cause of a low result is going too fast. The pattern is visible for nearly a second and the answer is untimed, and most poor scores come from people who looked for a third of the flash and then clicked in a hurry. Run it once more, deliberately slowly, before reading anything into the number.
The second most common cause is state rather than capacity. Working memory is unusually sensitive to sleep debt, alcohol, stress, and background conversation, and swings of several levels between a good day and a bad one are ordinary. A single run is a measurement of this attempt, not of you.
Working memory does decline with age, on average and gradually, and it is affected by a long list of conditions and medications. None of that can be diagnosed from one browser test. If your memory has changed noticeably in daily life, that observation is worth far more to a clinician than any score from this page.
Questions
- Would a shorter flash make this harder?
- It would make it a different test. A very brief flash measures the short-lived visual buffer, which holds more than you can report before it fades. The current duration measures what survives into working memory, which is the more meaningful limit.
- What is a good visual memory score?
- Level 9 to 10 is typical. Above 13 is well above average. Above 16 usually means a deliberate strategy such as grouping squares into shapes rather than a naturally larger memory.
- Can visual memory be improved?
- Your score on this test can improve a lot with practice. Your underlying capacity barely changes. Most of the gain comes from getting faster at converting scattered squares into remembered shapes.
- Why does it get so hard so suddenly?
- Because the grid grows on every success, so it converges on your limit quickly and then sits there. The sharp wall is the test working correctly.
- How does this compare to the Corsi block test?
- Corsi presents a sequence and requires you to reproduce the order, which recruits serial rehearsal and gives adult spans around five. Here the squares flash simultaneously and order does not matter, so it measures how many items you hold at once. The numbers run higher and the two are not interchangeable.
- Does a low score mean something is wrong?
- Almost never on its own. Most low results are haste, and working memory swings by several levels with sleep, stress, alcohol, and background noise. Run it again slowly. If your memory has changed noticeably in daily life, that observation matters far more than any score from a web page.
- Why do my results vary so much between attempts?
- Because working memory is state-dependent to an unusual degree, and because the grid grows until you fail, so your score always lands right at your limit where small differences in attention decide the outcome. Take the best of three if you want a stable number.
Where does the curve behind this page come from, and when does it get thrown away? The methodology page names the source of every reference distribution on the site, the sample size behind it, and the point at which real results replace it.