To choose your next step in Spatial Memory Game, select the option card that shows where the marked object's remembered 3D world point projects after the game's yaw rotation, then pitch rotation, and perspective step. In the Spatial Memory Game, the old screen location is not the answer; the task is to place that same world point in a new view.
That choice has two linked parts: decide which control comes next, then decide which of four cards contains the correct projection. The supported workflow is Reveal scene, Hide, and answer. Understanding what changes during that workflow is the key to making each selection accurately.
What Your Next Choice Represents
Each round contains three original objects placed at different horizontal, vertical, and depth positions. One object is marked and named as the target. During Reveal, the marked object's exact X, Y, and Z world coordinates are exposed. X represents left and right, Y represents height, and Z represents depth. The perspective floor grid and labeled axis rails make those relationships visible, but the object's shape is only an identifier; the important information is where it sits.
| Component | Function | Selection cue |
|---|---|---|
| X | Horizontal world coordinate | Track how far the target lies to the left or right of the origin. |
| Y | Vertical world coordinate | Track the target's height above or below the origin. |
| Z | Depth coordinate | Remember whether the target is nearer to or farther from the viewing position. |
| Yaw | Turns the scene around its vertical Y axis | Apply the horizontal turn before considering pitch. |
| Pitch | tilts the view around the X axis | Apply it after the yaw has established the new viewing direction. |
| Perspective | Converts rotated 3D coordinates into screen coordinates | Use rotated Z as depth, because changing depth can change the apparent position and scale. |
The target remains the same world point throughout the round. Yaw and pitch are rigid rotations, so they preserve its distance from the origin. What changes is the point's orientation relative to the camera and therefore its projection on the screen.
Make Your Move Through the Five-Round Sequence
Follow the same order in every round. There is no countdown or speed bonus, so use the Reveal phase to become confident rather than rushing to hide the scene.
- Press Reveal scene before trying to answer. A covered scene does not accept an answer, and Hide cannot run before Reveal. Once revealed, look for the dashed marker and the object's name so you know which of the three objects to remember.
- Store one complete 3D location. Use the marked object's exposed X, Y, and Z coordinates and relate it to the origin, the other two objects, and the depth grid. Do not rely only on where it appeared on the screen, because the next view is deliberately rotated and the final question concerns its new projection.
- Press Hide only when the position is clear. The objects disappear before the recall choices appear, and the round's published yaw and pitch are applied. Once hidden, you cannot inspect the original scene again, so check the horizontal, vertical, and depth relationships before leaving Reveal.
- Rotate the remembered point in the correct order. First apply yaw around the vertical Y axis, then pitch around the X axis. Keep following the same point rather than imagining that the camera has revealed a different object.
- Find the depth-sensitive projection. The four option cards show the same rotated coordinate frame with a crosshair at different projected screen positions. Choose the card where the remembered world point lands after both rotations and perspective projection. The visible candidates expose their exact screen x value out of 320 and screen y value out of 220, while the target's Reveal description provides its exact world coordinates.
- Submit, check the result, and continue. Each correct projection awards exactly 200 points. A first distinct wrong option can be repaired in the same round, and repeating that unchanged wrong option does not add another error. A second different wrong option closes the run. After a correct non-final answer, the interface returns to Reveal for the next round.
Derive the New Screen Location
The implementation uses a right-handed active convention. It rotates the literal target point with a Y-axis yaw matrix, then an X-axis pitch matrix, and finally performs perspective projection. Its authored screen mapping is:
- screenX = 160 + rotatedX × 240 / (320 − rotatedZ)
- screenY = 110 − rotatedY × 240 / (320 − rotatedZ)
The screen is 320 by 220, the focal length is 240, and the camera distance is 320. Rotated X and rotated Y determine the projected horizontal and vertical offsets, while rotated Z determines perspective depth. This is why two points with the same X and Y can appear at different screen positions: their depth values differ.
That depth dependence is also explained by Microsoft Learn's projection transform overview and independently by MDN's perspective() reference. Those sources support the general projection behavior; the game's focal length, camera distance, screen size, angles, and coordinates are its own authored display parameters.
You do not have to calculate the formula during every round. Its practical value is that it identifies the right reasoning order: preserve the world point, rotate it with yaw and then pitch, account for the resulting depth, and compare the projected location with the candidate crosshairs.
Use the Phase and Error Rules
The interface rules determine which action is available and what a mistake means. Check the current phase before submitting, because the same visible button can have different consequences depending on when it is used.
| Phase or result | What happens | Your next action |
|---|---|---|
| Covered scene | No answer can be submitted. | Press Reveal scene. |
| Reveal scene | The marked object's exact X, Y, and Z description is available. | Study the location, then press Hide. |
| Recall | Objects are hidden, yaw and pitch are applied, and four choices appear. | Select the single geometric match. |
| Correct answer | Exactly 200 points are awarded. | Continue to the next round unless this was the fifth. |
| First distinct miss | The round remains open for repair. | Reconsider the projection and try another option. |
| Repeated unchanged miss | The same round and display position do not count as another error. | Do not submit that option again; change your choice. |
| Second distinct miss | The run ends in a deadlock. | Restart when you want a fresh run. |
| Completion | Five correct rounds produce the fixed maximum of 1,000 points. | Restart to return to the first covered scene with zero score and no stored failures. |
After completion or deadlock, Reveal, Hide, cursor movement, and answer submission are frozen. Restart remains the way to begin again, and it restores the first covered scene rather than generating a new random layout.
Choose by Pointer, Touch, or Keyboard
Pointer and touch users can press any option card directly. Keyboard users can Tab into the option group, use the Arrow keys to move across its two-by-two layout, and press Enter or Space to select the focused card. The roving focus keeps the actual focused button, visible focus ring, and logical cursor aligned.
Focus also follows the round's sequence: Reveal moves focus to Hide, Hide moves it to Option A, and a correct non-final answer returns it to Reveal. Completion or deadlock moves focus to the readable terminal status. Before selection, no hidden answer label identifies the correct card, so use the coordinates, axes, and projection rather than searching for a visual correctness cue.
Keep the Result in Perspective
A correct choice shows that you matched one authored point in this game. It does not establish a general level of spatial memory, intelligence, attention, vision, or cognitive ability. The five scenes are fixed and deterministic rather than random, and they are not normed, calibrated, adaptive, or comparable between people. Display size, input method, familiarity with perspective drawings, and ordinary distraction can affect a play session without supporting a conclusion about health or ability.
The game is entertainment, not a measurement or diagnosis. It does not assess memory loss, neurological conditions, dementia, learning needs, injury, school readiness, or occupational fitness. It also runs locally in the browser with React and SVG: it requires no account, API call, WebGL, camera, device-orientation access, or motion hardware. If local storage is blocked, the shared shell cannot retain your best completed score, but the geometry and available choices remain unchanged.
The practical rule for every round is therefore simple: remember the target's X, Y, and Z coordinates, apply yaw and then pitch, use the resulting Z value for perspective, and choose the candidate whose screen x and y position matches that same world point.
Related reading: Visual Memory Test No Download: 5-Round Guide.
Related reading: How Do I Choose My Next Step in Battleships Puzzle.