Visual Memory Test

What the level you reach is measuring

This is a spatial span task: how much of a layout you can hold accurately enough to reproduce it immediately afterwards. The capacity being tested is short-lived and small, and it is separate from the memory that stores things for later — someone with an excellent memory for faces or facts can be unremarkable here.

The score is the level at which you ran out of lives rather than a count of squares, because the difficulty rises on two axes at once. Each level adds a square to remember and periodically enlarges the grid, so level twelve is a great deal harder than twice level six.

Three lives exist so that one slip does not end an otherwise good run. A misclick on a level you had comfortably remembered would otherwise decide your score, which would make the test mostly a measure of mouse accuracy.

Why the pattern is easier to hold if you stop trying to remember squares

People who do well on this tend not to memorise positions individually. They see a shape — a diagonal, an L, a cluster with one outlier — and remember that instead. Four separate facts are near the limit of what anyone can hold; one shape made of four parts is a single item.

This is the same trick that turns a phone number into three chunks instead of ten digits, and it is why practice improves scores here quickly at first. What improves is the strategy, not the underlying capacity, and the gain arrives in the first few attempts rather than accumulating over months.

It also means a distracted attempt is much worse than a focused one. Chunking needs the whole pattern taken in at once, and a glance away during the flash costs the shape rather than one square.

On a phone, and why touch is not scored separately here

The grid is sized to the screen and the squares stay comfortably tappable on a phone, which matters because the failure mode on small targets is a misclick that costs a life for a pattern you had actually remembered.

Unlike the reaction tests, touch and mouse attempts share one distribution on this page. The reason is that the score is a level reached rather than a latency, so the few milliseconds a touchscreen adds cannot move it, and splitting the sample would only delay every chart for no gain in honesty.

What does differ on a phone is the flash being easier to take in on a smaller physical area. Whether that helps or hurts appears to vary by person, and it is not something this site claims to have measured.

The other two playlists are order, not pattern

Pattern is the default: which squares lit, any order. Sequence asks for the order they lit. Reverse asks for that order walked backwards. The three buttons sit where Flick and Track sit on the aim trainer, and they do not share a series.

A level eight on pattern is not a level eight on sequence. Holding positions is a different load from holding relations, and reversing those relations is a third load. One chart for all three would hide the cost of order inside a lump.

The board, the flash timing and the start click stay the same. What changes is only the walk you have to replay, so a gap between the three scores is about that walk and not about a different grid.

This is a measurement toy, not a medical or psychological test. No score here says anything about your health, attention or cognition, and it is not a screening tool of any kind.

Questions

What is a good score on the visual memory test?

Most people finish somewhere around level nine to eleven, and getting past the middle teens is genuinely strong. The chart on this page shows the full spread, which is a better answer than a single threshold.

Why does the grid get bigger?

Because adding squares alone would eventually fill the grid and make the task trivial by elimination. Growing the grid keeps each new level a real increase in difficulty.

Can I improve at this?

Your score usually improves over the first few attempts as you start remembering shapes rather than individual squares. That is a strategy improving rather than your memory capacity growing, and the gains flatten off quickly.

Where did the sequence and reverse tests go?

They are the other two buttons on this page. Each still writes its own chart, because a pattern level and an order level are not the same achievement.