Sequence Memory Test

What it measures

Serial recall: holding items in order. That is a genuinely different task from holding a set of items, which is why most people score noticeably differently here than on the visual memory test, where order does not matter.

Most people reach a sequence of eight or nine. Getting past twelve is unusual. Because each round adds exactly one step and a single wrong square ends the run, scores bounce around more between attempts than on almost any other test here — one lapse at step ten costs you the whole round.

Why order is harder than quantity

Remembering which squares lit is a set. Remembering the order is a set plus the relationships between its members, and the second part degrades faster. The characteristic error is telling: people rarely forget a square entirely, they swap two adjacent ones.

That is a well-described pattern in serial recall — errors cluster at transpositions rather than omissions, and they concentrate in the middle of a sequence. The first couple of items and the last couple survive; positions four to seven of a nine-step sequence are where runs die.

Strategies that genuinely help

• Convert it to speech. Naming the squares — "top left, middle, bottom right" — moves the task into verbal memory, which most people rehearse more reliably than spatial patterns. It is the single biggest improvement available.
• Number the grid and read it as digits. A nine-cell grid becomes a number, and then every chunking trick from the number memory test applies.
• Chunk into shapes. Three squares forming a diagonal is one thing to remember rather than three.
• Rehearse during the flashes, not after. The gap between the last flash and your first click is where sequences are lost.

Deliberate design choices

No sound. The classic version of this game pairs each square with a tone, which makes it a mixture of spatial and auditory memory and disadvantages anyone who cannot hear the cues or is playing with the sound off. This version is visual only, so it measures one thing and works for everyone.

Never the same square twice in a row. Two identical flashes in succession are indistinguishable from one long flash, so the player loses a step through a rendering ambiguity rather than a memory failure. The generator explicitly prevents it.

Fully keyboard operable. The nine cells are real buttons and the number keys map to them, so nothing here requires a mouse or a trackpad. That is unusual for this kind of test and it costs nothing to do properly.

Where runs actually end

Not where people expect. Very few runs end because the whole sequence was forgotten — almost all of them end on a single transposition, usually somewhere in the middle third.

The first two or three steps are protected by having been rehearsed most often, and the last one or two are still fresh from having just been seen. It is the middle that has neither advantage, and it is where a nine-step sequence quietly becomes an eight-step sequence with two items swapped.

Knowing that changes how to rehearse. Most people repeat the sequence from the start each time, which reinforces the part that was already safe. Starting your rehearsal from the middle is awkward and noticeably more effective.

Why it gets harder faster than it looks

Each round adds one step, so the task appears to grow linearly. The failure rate does not. If your chance of holding any individual step is high but not perfect, the chance of holding an entire sequence falls off a cliff as it lengthens, because you need every step right rather than most of them.

That is why the difference between reaching eight and reaching twelve feels enormous despite being four steps. It is also why scores on this test bounce around more between attempts than on any other test here, and why a single run tells you very little.

A note on what it is not

This is not a clinical instrument and not a diagnostic. Sequence tasks do appear in cognitive assessment, but with standardised administration and normative data that a web page cannot supply. A poor score here means you had a poor run on a browser game.

Your results are stored in this browser only. There is no account, nothing is uploaded, and clearing your browsing data clears them.