The Multiple Object Tracking Test is a deterministic five-round browser game that hides two target discs behind six identical ones and scores exactly 200 points for every correct unordered pair of final positions. Each round opens with a Preview phase that highlights two discs with accent rings, then unmounts those rings as soon as motion starts so every visible disc shares the same shape, color, size, and accessible description. Tracking advances frame by frame on a fixed 650-millisecond timer, and each frame is a complete permutation of disc identities 0 through 5, which guarantees one disc per location and prevents overlaps or missing objects. Finish motion now jumps to the same final frame without skipping validation, so both the automatic timer route and the manual jump route end at the same declared positions. The Answer phase replaces the moving SVG with six numbered position buttons; the player toggles exactly two and submits them as an unordered set. Five correct rounds therefore produce 1,000 points, with no speed bonus, random path, adaptive motion, click penalty, or hidden multiplier. Every path, target pair, frame interval, and error rule is an original product fixture, and the only thing stored across sessions is the shared best completed score in the local browser.
The Three Phases of Each Round
Every round of the Multiple Object Tracking Test cycles through Preview, Tracking, and Answer, and each phase has its own input surface and visual style. Preview shows the six-disc field at rest with two ringed targets highlighted for memorization. Tracking unmounts the rings, animates the discs between fixed coordinates, and lets the player call Finish motion now to land on the same final frame that the 650-millisecond timer would have reached. Answer replaces the animated SVG with six numbered position buttons and a single submit control that stays disabled until exactly two positions are selected.
The three phases are deliberately non-overlapping. You cannot toggle position buttons during Tracking, and you cannot start motion from Answer. Because each frame is a complete permutation of six identities, the same set of positions always appears in Answer; only the requested target pair changes between rounds. This means the inputs you give in Answer can be reasoned about independently of how you got there, and the validator only checks the unordered pair against the targets' last-permutation coordinates.
| Phase | What the player sees | What the player can do |
|---|---|---|
| Preview | Six discs at rest, two with accent rings | Memorize targets, press Space, or choose Start motion |
| Tracking | Six identical discs moving across permutation frames | Watch each 650 ms tick or press Space / Finish motion now |
| Answer | Six numbered position buttons and a submit control | Toggle exactly two positions, then submit the unordered pair |
Preview: Reading the Two Ringed Discs
Preview is the only phase that gives you a visual handle on the targets. Two of the six discs wear accent rings, and everything else about them stays consistent with the rest of the field, so the rings are the sole cue. Memorize the ringed pair's coordinates rather than the disc identity, because as soon as Tracking starts the rings are unmounted and every disc shares the same fill, stroke, size, and accessibility description.
If you lose track of which two positions you picked, you cannot recover them later. The submit step at the end of the round does not reveal your Preview choices; it only tells you whether the unordered pair you toggled in Answer matches the two positions those original discs occupy in the final frame. Because Preview and Tracking are non-overlapping, the only safe place to anchor on the targets is the brief moment before you press Space.
Tracking: Following Six Identical Discs
Tracking is the most mechanically constrained phase. Discs move between fixed coordinates on a 650-millisecond timer with a visible CSS transition, and Finish motion now advances to the same final frame the timer would have reached without skipping any intermediate validator state. Because every frame is a permutation of disc identities 0 through 5, the field never has two discs in the same cell and never leaves a location empty. This keeps tracking decisions about identity rather than about ambiguity resolution.
There is no random path or adaptive motion: the disc trajectories are fixed fixtures, so the same five rounds always produce the same routes. Reduced-motion styling removes the CSS transition while preserving each permutation state, which means the discs still arrive at the correct final coordinates for that frame even when the browser suppresses animation. The independent test restates every target identity, every frame permutation, and every expected final target position as literal fixtures, and the field reserves a stable five-by-three aspect ratio so the discs stay on screen as the layout reflows around them.
Answer: Selecting Two Final Positions
Answer is the only phase where you produce inputs the game records. The moving SVG is replaced by six numbered position buttons labeled 1 through 6. Toggling an active position deselects it, toggling an inactive position adds it until two are selected, and attempting a third selection does nothing. The submit control stays disabled until exactly two positions are held, which means you cannot submit one or three positions by accident.
Because validation is unordered, positions {2, 5} and {5, 2} describe the same pair and produce the same outcome. Selections themselves never count as errors, so you can reconsider freely before pressing Submit. The wrong-signature dedupe is round-aware and order-independent: submitting the identical unordered pair twice does not consume another error. To record a second distinct miss and close the run, you must change at least one position and submit another wrong pair.
How to Play the Multiple Object Tracking Test
The whole game follows the same three-input loop, so a single set of steps covers the full five-round route. The contract below is the verified operating sequence; every round and every keyboard shortcut aligns with it.
- Memorize the two ringed discs during Preview, then press Space or choose Start motion. Preview shows the field at rest with two accent rings; once Tracking begins, the rings are unmounted and you cannot recover the target identities.
- Follow both identical discs through the fixed frames; wait for motion to finish or choose Finish motion now. Each frame is a six-item permutation that advances every 650 ms, and Finish motion now jumps to the same final frame the timer would have reached.
- Toggle exactly two final positions and submit; solve all five fields for exactly 1,000 points. A correct unordered pair awards exactly 200 points and returns the field to Preview for the next target set.
Pointer and touch users get the same start, finish, position, submit, and restart controls with large targets and visible selected states. Keyboard players use Space to start motion from Preview and to finish it during Tracking, the digit keys 1 through 6 to toggle final positions, and Enter to submit a complete pair. The ledger reports frame count, selected count, score, and misses, so a player who finishes the run can read exactly how each round resolved.
Scoring, Errors, and the 1,000-Point Run
The scoring rule is fixed. Each correct unordered pair returns 200 points and resets the field to a fresh Preview with two new target rings. Five correct rounds therefore produce 1,000 points, with no speed bonus, random path, adaptive motion, click penalty, or hidden multiplier. There is no score compared against population norms, and the run is not framed as a measurement of any cognitive trait.
Error handling is equally deterministic. A wrong pair does not consume an error until you press Submit, and submitting the identical unordered pair again is recognized as a repeat and does not consume another miss. To register the second distinct miss and close the run, you must change at least one position and submit another wrong pair. Two distinct wrong submissions therefore deadlock the round and end the run.
| Outcome | Result | Next state |
|---|---|---|
| Correct unordered pair | +200 points | Return to Preview with two new targets |
| Wrong unordered pair | First miss recorded | Stay on the same round until you change at least one position |
| Repeat the same wrong pair | No new miss | Stay on the same round until a distinct wrong pair is submitted |
| Second distinct wrong pair | Run closed | Ledger reports the final score and miss count |
Restart restores the first target preview, frame zero, zero score, and no selected positions or errors. The independent test proves each frame contains all six identities exactly once, advances timer boundaries, checks Finish motion now, selects correct positions in reverse order to prove unordered validation, executes the full 200-through-1,000 route, repeats an identical wrong pair, changes one position to trigger deadlock, rejects invalid locations, and verifies phase and terminal atomicity. Only the shared best completed score is kept in this browser, so a fresh tab always starts from zero.
Keyboard, Pointer, and Reduced-Motion Play
Keyboard play covers the complete game. Space starts motion from Preview and finishes it during Tracking, so the same key handles both transitions in their respective phases. Keys 1 through 6 toggle the six numbered position buttons in Answer, and Enter submits a complete unordered pair. Pointer and touch users receive the same start, finish, position, submit, and restart controls, with large targets and visible selected states that announce the current selection count.
The game requires no precision drag, pointer-only gesture, sound, upload, or hardware sensor, and everything runs locally with no camera, eye tracker, file input, account, API, model, external asset, random service, or personal data. Answer buttons announce only final position numbers and their selected state, which means the assistive surface matches the validator's unordered-pair model. Reduced-motion styling removes the CSS transition while preserving each permutation state, so players who suppress animation still see every disc arrive at the correct cell. Double-press Escape opens the simulated spreadsheet boss screen and double-press Escape again resumes the unchanged field.
Other Fixed Visual Memory Formats
The Multiple Object Tracking Test is part of a family of fixed-fixture visual memory games that share the same five-round, 1,000-point skeleton but use different task surfaces. For static layout recall, Position Memory Game shows three shapes on a 3 by 3 board and asks you to restore each hidden position across five fixed files. Visual Memory Test extends the same skeleton to grid patterns with explicit reveal and hide controls, while Pattern Memory Game presents five original 3 by 3 patterns and asks for each exact hidden record from four options. Face Memory Test keeps the same deterministic contract but swaps the stimuli for five illustrated profiles and three visual alternatives each.
For readers who want to compare the phase breakdown above with the two existing walkthroughs, the Moving Object Memory Game: Track Two Discs to 1,000 guide treats the same five-round loop from a memorization angle, while the Attention Tracking Game: Score 1,000 in 5 Rounds guide frames the same rules from a scoring perspective. The phase-by-phase view on this page focuses on what each phase requires of the player and how the validator treats selections, repeats, and deadlocks.