Skip to content
Testpartout

Human performance · Memory

Visual Memory Test

This visual memory test measures how many tile positions you can hold in mind at once. A pattern of highlighted tiles appears on a fixed 6 × 6 grid for one and a half seconds, then disappears, and you select the tiles you remember.

Every level you complete adds one more tile to the next pattern, starting at three. Your result is your visual span: the largest number of tiles you recalled correctly in a single completed round. It runs entirely in your browser, with no permissions and no sign-in.

  • Fixed 6 × 6 grid
  • Starts at 3 tiles
  • 3 lives
  • No permissions required

Not started

Level 1 · 3 tiles · Lives remaining 3 of 3

Press Start test. Level 1 shows 3 tiles for 1.5 seconds, then you select them. Tiles are only selectable during the recall phase; presses while the pattern is shown do nothing. Keyboard users can move between tiles with Tab and select the focused tile with Enter or Space.

Session in progress

Not started
Visual span
—

Shown when the session ends

Highest level completed
—
Successful rounds
0
Lives used
0 of 3
Personal best
—

Saved in this browser after your first completed session

Only your best visual span, the level it was reached at and a timestamp are stored in this browser. No pattern, tile press or session history is kept, and nothing is uploaded.

Your result

What your result means

Visual span is the largest pattern you reproduced without a single wrong tile. It is a count of spatial positions, not a score, a percentage or a rating. A span of six means one completed round contained six target tiles and you selected all six and nothing else.

The number describes one short session on one screen, under whatever conditions you were in. Attention, screen size and distraction can affect it from session to session, so the useful comparison is with your own earlier attempts on the same device rather than with anyone else.

  • Only completed rounds count. If a session ends before you finish a single round, the visual span is 0 — starting a level is not the same as completing it.
  • The grid never changes size, so a level is harder only because it holds more tiles, not because the geometry moved.
  • A failed round costs one life and gives you a brand-new pattern at the same level, so you can never memorise the board you just missed and retry it.
  • Testpartout shows no percentile, no global average and no performance label, because no real aggregated data exists yet. When it does, it will come from measured results — never from estimates.

Procedure

How the test works

  1. 01

    Start when you are ready

    Nothing flashes until you press Start test. A memory pattern never appears while you might be reading the page.

  2. 02

    Memorise the pattern

    Level 1 highlights three tiles on the 6 × 6 grid for about one and a half seconds. Highlighted tiles carry a filled marker as well as a colour.

  3. 03

    The pattern hides

    The highlights disappear together and the grid becomes selectable. Nothing you press during the presentation is counted.

  4. 04

    Select what you remember

    Tap or click the tiles that were highlighted, in any order. Each correct tile is marked with a tick, and selecting an already-recalled tile again does nothing.

  5. 05

    One wrong tile ends the round

    Selecting a tile that was not part of the pattern fails the round immediately, costs one life and reveals the tiles you missed.

  6. 06

    Progress or retry

    A completed round moves you up a level and adds a tile. A failed round keeps the same level with a new pattern. After three failed rounds the session ends and reports your visual span.

Methodology

How Testpartout measures this

The method below is exactly what the code on this page does — nothing is described that is not implemented.

What the implementation does

  • The board is a fixed 6 × 6 grid of 36 cells for every level. Grid geometry never changes between levels in this version.
  • Level 1 uses 3 target tiles. Each completed level increases the target count by exactly one, so level 2 uses 4 tiles, level 3 uses 5, and so on.
  • Each round's pattern is generated fresh in your browser, using crypto.getRandomValues() where available, and always contains the required number of unique cells.
  • The pattern is shown for a nominal 1500 ms, timed by the browser, and then hidden in a single step.
  • Tiles are only selectable during the recall phase. Pointer presses and keyboard activation during the presentation cannot affect the round.
  • A round succeeds when every target tile has been selected and no other tile was selected. Selecting an already-recalled target again is a no-op, not a second success.
  • A single non-target selection fails the round immediately. A resolved round cannot fail twice, so rapid extra presses never cost more than one life.
  • A session has 3 lives. A failed round costs one life and repeats the same level with a newly generated pattern — never the pattern you just failed.
  • After the third failed round the session completes. Visual span = the largest target count of any successfully completed round; a session with no completed round reports 0.
  • 36 tiles is the maximum: the grid holds 36 cells, so completing a 36-tile round ends the session with a visual span of 36 and no further level is created.
  • Every round begins from an explicit action: Start test, Start next level or Retry level. That control is outside the grid and is never part of the measured input.
  • If the page becomes hidden while a pattern is being shown or recalled, the whole session is cancelled and produces no result.

Limits of a browser measurement

  • The 1500 ms exposure is browser-timed. Background load, battery-saver modes, display refresh rate and compositing all mean the visible duration is approximate, not laboratory-controlled.
  • Screen size, pixel density, brightness and viewing distance change how easily a pattern can be taken in, and none of those can be measured from a web page.
  • This is a browser interaction test, not an assessment instrument. It is not standardised, not validated and not a clinical measure of memory.
  • A single session is noisy. Attention lapses, interruptions and unfamiliarity with the interaction can move the result more than small differences in ability.
  • Nothing prevents pausing to write the pattern down or photograph the screen. A local result that was not produced from memory is meaningless, but it is only your own browser storage.

Privacy

What happens to your data

This section describes the behaviour of the code on this page, not an intention.

  • The test runs in this page. No pattern, tile press or session history is sent anywhere.
  • No account, sign-in or device permission is required to run the test.
  • Only your best result is written to this browser's local storage: the visual span, the level it was reached at and a timestamp. No board, no press log and no identifier are stored.
  • Random patterns exist only in memory for the round being played and are never persisted, so a restart cannot replay an earlier board.
  • Clearing your browser storage removes the personal best. There is no copy of it anywhere else.
  • Results are not uploaded, aggregated or ranked. If Testpartout ever offers aggregated comparisons, it will be explicit and based on real measured data.

Troubleshooting

If the test does not behave as expected

Nothing happens when I press a tile

Tiles are only selectable during the recall phase. While the pattern is being shown, and after a round has resolved, the grid is disabled on purpose so a stray press cannot score.

The pattern disappeared before I was ready

Exposure is a fixed one and a half seconds and is not adjustable in this version. Press Start level only once you are looking at the grid — nothing runs until you do.

The session was cancelled

Switching tabs, minimising the window or locking the screen hides the page, and a hidden page cannot present or measure a pattern honestly. The session is discarded and your personal best is untouched.

I lost a life for a tile I meant to press

Any tile outside the pattern ends the round, wherever you meant to press. On a small screen, tap slightly towards the centre of a tile rather than the gap between two.

My result did not save

Only a completed session that beats your stored best is saved, and a cancelled session never saves. Private browsing and storage-blocking settings can also prevent writing to local storage.

I want to use the keyboard

During recall the tiles are ordinary buttons: Tab moves between them and Enter or Space selects the focused tile. The start and retry controls are separate, so activating one never selects a tile.

Factors

Factors that affect a visual memory result

The same person can produce different spans on different days and different screens. These factors can move the number without your ability changing.

Screen size and viewing distance

A larger grid at a comfortable distance can be taken in with fewer eye movements than a small phone screen held close or far away.

Display visibility and contrast

Brightness, glare, a dimmed battery-saver screen and low-contrast conditions can all make the highlighted tiles harder to register in one and a half seconds.

Attention and distraction

Notifications, background conversation and doing something else at the same time can affect a short memory task and can increase variability between attempts.

Familiarity with the interaction

Familiarity with the timing of the presentation can influence repeated attempts, so early and later attempts may differ.

Chunking strategy

Grouping tiles into shapes, rows or pairs is a common strategy and can change results. It is not cheating, but it means two people are not necessarily doing the same thing.

Fatigue and time of day

Tiredness can affect short-term memory tasks, so repeated sessions late in a long day may not be comparable with fresh ones.

FAQ

Frequently asked questions

What does visual span mean?
It is the largest number of tiles you recalled correctly in a single completed round. If your best completed round had six target tiles and you selected all six without a wrong tile, your visual span is 6.
How long is the pattern shown?
A nominal one and a half seconds, timed by the browser. Display refresh, background load and battery-saver modes mean the visible duration is approximate rather than laboratory-exact.
Why does one wrong tile end the round?
Because the round is testing whether you recalled the pattern, not how many guesses it takes to find it. Allowing wrong selections would let anyone reach a high span by pressing every tile.
Does the grid become larger at higher levels?
No. The board stays a 6 × 6 grid of 36 tiles at every level. Only the number of target tiles increases, by one per completed level, so the scoring stays transparent and repeatable.
What happens after I lose a life?
You keep the same level and get a completely new random pattern of the same size. The pattern you missed is never repeated, so you cannot learn a board and retry it. After three failed rounds the session ends.
Does switching tabs cancel the test?
Yes. If the page becomes hidden while a pattern is being shown or recalled, the session is cancelled, no result is produced and your personal best is left untouched.
Is my pattern or click history uploaded?
No. Nothing leaves your browser. Patterns exist only in memory for the round being played, and only your best visual span, its level and a timestamp are stored locally.
Why don't I see a percentile or an average?
Because Testpartout has no real aggregated data yet, and inventing one would be misleading. Comparisons will appear only when they can be built from real measured results.
Does this test diagnose memory problems?
No. This is a browser interaction test, not a medical or psychological assessment. It cannot diagnose or rule out any condition, and a low or high result says nothing about your health. If you are concerned about your memory, speak to a qualified clinician.