Visual Memory Test

Order does not matter, and that changes everything

This is the counterpart to the sequence memory test, and the difference is the whole point. There, squares light one after another and you must reproduce the order. Here they all light at once and you only have to say which.

Most people are markedly better at one than the other, and which one tends to be consistent for a given person. Set memory and serial memory are not the same capacity, and comparing your two scores is more informative than either on its own.

How the difficulty is built

Two things rise together: the number of squares to hold, and the number of places each could be. That second one is why level 10 feels so much harder than level 7 despite only three more squares — the search space has nearly doubled.

Lit cells are capped at 45 per cent of the grid on purpose. Beyond about half, the pattern becomes easier to remember as its negative — the empty squares — and the test stops measuring what it is supposed to.

What people do to get further

• See shapes, not squares. Three cells in a row is one thing to remember. Four cells forming a square is one thing. This is chunking again, and it is the difference between holding twelve items and holding four.
• Split the grid in half. Memorise the top and the bottom separately; two small patterns are easier than one large one.
• Use the edges and corners. They are distinctive and anchor everything else. The middle of a 6×6 grid is where patterns dissolve.
• Count. Knowing you are looking for nine squares tells you when you have finished guessing, and stops you clicking a tenth out of momentum.

Why the flash does not get shorter

Some versions shorten the display as levels rise. This one does not. Shrinking the window turns the test into a measure of how fast you can scan a grid, which is a perceptual speed task rather than a memory one — and it makes the late levels fail for a reason that has nothing to do with what is being measured.

The difficulty comes from the pattern instead, which is the honest place to put it.

Why most people are better at one than the other

Comparing your score here against your sequence memory score is more informative than either number alone, because the two tasks load different systems.

This one is closer to a snapshot: the pattern arrives all at once and either survives or does not. Sequence memory is a chain, where every link has to hold and one broken link ends the run. People who rehearse verbally — turning positions into a spoken list — tend to do better on the chain. People who see and hold shapes tend to do better here.

Neither is better, and the gap between your two scores is mostly telling you which strategy you default to rather than anything about capacity.

What the grid growth actually does

Every three levels the grid gains a row and a column, so it goes from 9 cells to 16 to 25 to 36. That is not a gentle increase: by level 10 there are four times as many places a square could be as at level 1, while the number of squares to remember has only risen from three to twelve.

The consequence is that late levels fail differently. Early on, people miss a square. Later on, they remember roughly the right region and click a neighbour — the pattern is held approximately rather than precisely, which is a recognisable feature of spatial memory rather than a failure of attention.

On phones, and on accessibility

The grid scales to the space it has, so a pattern takes up roughly the same proportion of a phone screen as a desktop one, and the test is designed to work at 375 pixels wide. Every cell is a real button with a visible focus state and a label naming its position, so the whole thing is operable from a keyboard.

As with every test here, results are stored in this browser only. No account, no upload, and nothing that syncs anywhere.