What this test measures
Visual working memory is the short-lived store that lets you keep something you have just seen available for a few seconds after it is gone — the layout of a page you glanced at, where a button was, which squares were filled in. It is small, it fades fast, and it is not the same thing as remembering facts or events.
This test loads it in two directions at once. More lit tiles increases how much you have to hold. A bigger board spreads the same load over more space, so you also have to scan and locate. Both go up as you progress, which is why one level number can stand in for two dimensions of difficulty.
Pattern span, not the Corsi test
If you have taken a grid memory test before, it was probably a Corsi block-tapping task: blocks light up one at a time and you reproduce the order. That version leans on the spatial-sequential side of memory — a route through space rather than a picture.
This test does something deliberately different. The whole pattern appears at once, and the order you tap in is ignored. That makes it a pattern span task, the design Della Sala and colleagues introduced precisely in order to pull the visual component of memory apart from the spatial-sequential one. The practical consequence: a score here and a Corsi score are not interchangeable, and neither is a substitute for the other.
The level ladder
Density is held at roughly a third of the board, so a larger grid is harder because it is wider to scan — not because it is suddenly crowded.
| Level | Board | Lit tiles | Density |
|---|---|---|---|
| 1–2 | 3×3 | 3, then 4 | 33–44% |
| 3–5 | 4×4 | 5, 6, 7 | 31–44% |
| 6–8 | 5×5 | 8, 9, 10 | 32–40% |
| 9–12 | 6×6 | 11 to 14 | 31–39% |
| 13 and up | 6×6 | one more each level, to 18 | up to 50% |
How to read your result
Your score is the highest level you cleared. Because the board changes underneath you, the interesting question is usually not the number itself but where you fell off. Stalling at the jump from 4×4 to 5×5 is a different pattern from stalling at eleven tiles on a board you already know.
A first attempt is almost always an underestimate. You are learning the rhythm of the test as much as you are being measured by it, and most people gain a level or two in their first few sessions before settling.
Strategies — and why they inflate the score
There are things that reliably help:
- Chunking into shapes. Seeing “an L in the top-left and a pair on the right” is three things to remember instead of five.
- Verbal labels. Silently saying “three across the middle row” converts part of the picture into words.
- Splitting the board. Treating a 6×6 as four quadrants and counting within each.
These are not cheating — but they do change what is being measured. Once you are labelling patterns in words you are partly testing verbal memory and strategy rather than raw visual span. This is one of the well-known reasons span tasks are noisy, and it is the usual explanation when someone's score jumps several levels between two sessions.
What else moves your result
Screen size matters more here than in most tests: a 6×6 board on a small phone is genuinely harder to take in at a glance than the same board on a laptop. Viewing distance, glare, interruptions and how tired you are all count. So does whether you look at the whole board or let your eyes hop from tile to tile — the pattern is meant to be taken in as a picture.
Where this task comes from
Grid-based span tasks sit inside Baddeley's working memory model, in the part that handles visual and spatial material rather than sound and language. The Corsi block-tapping task has been in use since the early 1970s, and standardised versions with published normative data exist — Kessels and colleagues administered one to 70 healthy controls alongside 70 patients with brain lesions, and reported percentiles and cut-off scores, along with the finding that right-hemisphere patients performed worse than left-hemisphere ones.
That work is the reason to be careful with a web version, not a reason to trust one. Those numbers were collected with a physical board, a fixed procedure and a trained administrator. A page you opened on your phone while the kettle boiled is not that, and should not be read against those percentiles.
Questions people ask
What is a good score on a visual memory test?
Most people clear somewhere in the middle of the ladder, but a browser test on an unknown screen at an unknown viewing distance has no norms worth quoting. Published normative data exist for clinical versions of these tasks under controlled conditions, and that is exactly why they should not be applied to a web page.
Use your own first attempt as the baseline.
Is this the same as the Corsi block-tapping test?
No. The Corsi task lights blocks one at a time and asks you to reproduce the order, which leans on spatial-sequential memory. This test shows the whole pattern at once and ignores order, which makes it a pattern span task — a design introduced specifically to separate the visual component of memory from the spatial-sequential one.
Scores from the two are not interchangeable.
Does the order I tap the tiles in matter?
No. Tap them in whatever order you like. Only the set of tiles matters, which is what keeps this a test of the visual pattern rather than of a remembered sequence.
Why does the board get bigger instead of just adding more tiles?
Because the two make different demands. More lit tiles on the same board increases how much you have to hold; a bigger board spreads the same load over more space, so you also have to scan and locate.
Growing both keeps the density roughly constant at about a third of the board, so later levels are harder without simply becoming crowded.
Do you store my results?
Your best and last score are saved in your own browser using local storage. Nothing is sent to a server, there is no account, and clearing your browsing data removes them.
References
- Della Sala, S., Gray, C., Baddeley, A., Allamano, N., & Wilson, L. (1999). Pattern span: a tool for unwelding visuo-spatial memory. Neuropsychologia, 37(10), 1189–1199. PubMed
- Kessels, R. P. C., van Zandvoort, M. J. E., Postma, A., Kappelle, L. J., & de Haan, E. H. F. (2000). The Corsi Block-Tapping Task: standardization and normative data. Applied Neuropsychology, 7(4), 252–258. PubMed
- Corsi, P. M. (1972). Human memory and the medial temporal region of the brain (doctoral dissertation). McGill University.