Memory
Sequence Memory Test
Blocks light up one at a time; you tap them back in the same order. The sequence grows by one block each time you get it right. Your score is the longest sequence you reproduced correctly — a count of positions, not a percentile, and not the same measurement as a clinical block-tapping span.
Sequence memory
Blocks light up one at a time, one every two seconds. When they stop, tap them in the same order. Each time you get it right the sequence gets one block longer. Two misses at the same length ends the test. There is no undo — a tap is final.
What your equipment contributed
- Your display
- not measuredassuming a typical 60 Hz panel
- Each block lit for
- 1000 ±16.7 msquantised to whole frames
- Board rendered at
- not measuredCSS pixels, square
Timing barely reaches this score: a block's 1000 ms exposure is quantised to whole frames, but the result is a count of blocks rather than a measured duration, so a frame either way changes nothing. The board size is the limit that does matter here, and we can only half state it. The task this test derives from uses a board of fixed physical dimensions at a fixed viewing distance, so every participant sees the same spatial layout. Yours is whatever your viewport allows across — and a CSS pixel is not a fixed physical length, so we cannot convert that to centimetres, and we certainly cannot know how far your eyes are from the screen. A board spanning your whole phone at arm's length and one occupying a corner of a large monitor are different tasks. We report the number the browser gives us and decline to derive a physical size from it.
How this is measured
How the sequence is presented
Nine blocks sit in an irregular arrangement — no two share a row or a column. One block at a time turns light, on a two-second cadence: lit for 1000 ms, then 1000 ms dark before the next. The dark gap is what makes two taps read as two separate events rather than one highlight wandering across the board. During presentation the blocks are disabled, so an early tap cannot be registered as an answer at all.
Why the layout is irregular, and why blocks never repeat
A 3×3 grid would let you name the cells and rehearse the names, converting a spatial span into a verbal one — you would end up taking a worse version of our number memory test without being told. The physical board is irregular for that reason and so is ours. For the same kind of reason a sequence never revisits a block: the board-based task doesn't, and a repeated position is an easier item than a fresh one, which would make some rounds quietly easier than the length printed beside them.
How the test ends, and what the score is
The sequence starts at two blocks and grows by one every time you are right. Two misses at the same length end the test — one slip is not enough to establish that a length is beyond you, which is why the published procedures give more than one attempt at each length. Your score is the longest sequence you reproduced correctly. A tap is final: there is no undo and no confirm step, because the task this derives from has neither and a correction affordance would be measuring something else.
The ceiling is the board, not you
There are nine blocks, and a sequence cannot revisit one, so nine is the longest sequence this instrument can present. Reproduce all nine and we report ≥9, not 9 — past that point we stopped measuring you and started measuring the board. This is the same reasoning our colour test uses when a visitor sees a difference finer than the display can render: naming the ceiling is more useful than pretending the number is a result.
What we cannot tell
We cannot distinguish a mis-aimed tap from a forgotten position. We show you where each miss diverged — “correct for four, then diverged” is different information from “wrong from the first tap” — but both end the round identically and both count the same. We also cannot tell whether you traced the sequence with a finger, sketched it, or were interrupted. And because the test is purely visual and spatial, we deliberately do not announce the lit blocks to a screen reader: doing so would turn it into a verbal sequence task and produce a number that looks like this one but measures something else.
What your equipment contributes
Timing, very little — each 1000 ms exposure is quantised to whole frames by your display, but the result is a count of blocks rather than a duration. Your screen size matters much more, and we can only half measure it. The panel above reports the rendered board in CSS pixels, which is the most the browser will tell us. A CSS pixel is not a fixed physical length, and nothing in a web page can measure your viewing distance, so we state the number we have and refuse to convert it into centimetres or into a correction.
How your result compares
Spatial span has been studied on a physical board for decades, so a defensible reference figure exists. In the largest recent normative sample we found, 340 healthy adults aged 21–89 (mean age 51.6) averaged a forward span of 4.97 blocks (SD 1.03). The examiner touched the cubes at one cube every two seconds, and a length was passed on two of three trials correct at a length.
What we matched, and what we could not
Matched: the presentation rate of one position every two seconds, the principle of more than one attempt per length, and the irregular non-repeating layout.
Not matched: theirs was a physical board of fixed size, touched by an examiner sitting opposite, at a fixed distance, with standardised sequences rather than random ones and a two-of-three rule rather than our two-misses rule. Ours is a flat screen of unknown physical size at an unknown distance, tapped with a finger or clicked with a mouse, with sequences drawn at random.
And the format difference is not hypothetical. A pilot study of 40 students ran both the standard board and a tablet version and found forward reproduction was more accurate on the physical board. Its authors concluded that computerization “has serious consequences for the cognitive concepts that the Corsi Task is assumed to assess”. That is a published finding that our format scores differently from the one the 4.97-block figure came from — and it is not a constant we could subtract, because the study was a pilot on 40 young adults with a different device from yours.
Read 4.97 blocks as a bearing on what typical looks like for a related task on different apparatus — not as a line you have passed or failed.
No percentile is shown. There are three independent reasons, any one of which would be enough. The source gives a mean and an SD but not the distribution's shape, and spans are bounded integers rather than a smooth curve. The normative sample is Italian adults on a physical board, and you are not on a physical board. And the one study that compared the two formats directly found they differ — so even a perfectly characterised distribution for the board task would not locate you on the screen task. A percentile drawn from BlinkBench's own visitors would be worse than any of that: self-selected, unvetted, and taken on screens of every size.
Reference range: Facchin, A., Pegoraro, S., Rigoli, M., et al. (2024). Regression-based normative data for Corsi Span and Supraspan learning and recall among Italian adults. Neurological Sciences, 45(12), 5707–5718.
Format comparison: Claessen, M. H. G., van der Ham, I. J. M., & van Zandvoort, M. J. E. (2015). Computerization of the standard Corsi Block-Tapping Task affects its underlying cognitive concepts: a pilot study. Applied Neuropsychology: Adult, 22(3), 180–188.
Last reviewed: July 2026
Frequently asked questions
Why aren't the blocks in a neat grid?
Because a grid is a different test. Regular rows and columns let you name the cells — top-left, middle, bottom-right — and rehearse the names, which quietly converts a spatial memory task into a verbal one. The physical board this test derives from is irregular for exactly that reason, and ours is too. Our nine positions are our own scatter, not a reproduction of any published board's measurements, but no two of them share a row or a column.
Why does a block never light up twice in the same sequence?
Because the board-based task never taps the same block twice, and a repeat is an easier item than a fresh position — you already have it. Allowing repeats would make some rounds quietly easier than the length printed beside them. It also means nine is the hard ceiling: the board runs out. If you reproduce all nine we report your result as ≥9, because at that point we stopped measuring you and started measuring the instrument.
Is this the same as a Corsi block-tapping score?
No, and the difference has been measured. A pilot study of 40 students ran both the standard physical task and a tablet version, and found forward reproduction was more accurate on the physical board — the authors concluded that computerization "has serious consequences for the cognitive concepts that the Corsi Task is assumed to assess". So a screen version is not a clinical Corsi span with a different case. That published finding is why we show you a reference figure as context and no percentile at all.
Does it matter whether I'm on a phone or a monitor?
Almost certainly, and we can't correct for it. The physical task uses a board of fixed dimensions at a fixed viewing distance, so everyone sees the same layout. Our board scales to your viewport, and the browser reports its size only in CSS pixels — which are not a fixed physical length, and say nothing about how far away you are sitting. A board filling a phone held at arm's length and one in the corner of a large monitor are different tasks. We show you the pixel figure and decline to convert it into a claim.
Related tools
The longest string of digits you can hold and recall, presented one per second to match the research the score is compared against.
Open tool →The largest grid pattern you can reproduce after a brief exposure.
Five trials measuring how fast you respond to a colour change, reported as a median with the share your own display could account for.
Open tool →