The Face Memory Test is a five-round remember faces quiz in which you study one original fictional profile, then identify the same illustration from three choices before a second distinct miss. Each target appears during a two-tick Reveal. You get two visible one-second ticks to study it, or you can choose Recall now to move directly into recall. Once the target unmounts, three option buttons mount. Pick the matching profile by clicking its large illustrated button or pressing 1, 2, or 3. A correct choice adds 200 points and starts the next round. The fixed run contains five rounds, so the calculation is 5 correct matches × 200 points = 1,000 points. Missing once does not end the game: the first distinct incorrect profile is recoverable. Selecting that same wrong profile again during the same round is deduplicated and leaves the error count unchanged; a second different profile ends the run and disables further answers. Unlike a celebrity or photo-based quiz, this game uses fifteen code-drawn SVG profiles identified by fictional fixture names from Alder through Opal. They are not portraits, generated people, identity records, biometric templates, demographic symbols, or emotion labels. The game does not recognize faces, compare uploads, infer sensitive attributes, diagnose a condition, or measure memory ability. It asks one bounded question: which original illustration was visible a moment earlier. Everything runs locally, with no camera request, file upload, account, API, machine-learning model, cookie-based identity, or network call. The only product data that may remain is the shared best completed score in this browser.
The task is a five-round visual recall
One round has two clearly separated states. During Reveal, the game mounts one target illustration and no answer buttons. You can study its head outline, hair, brows, eyes, nose, mouth, and any cheek marks or freckles until the two visible ticks finish. Choosing Recall now ends this state early. The target then unmounts, and the Recall state mounts exactly three distinct illustrated options.
This state separation keeps the task auditable. The hidden target is not left behind in the active Reveal document through opacity, clipping, alternative text, a data attribute, or a concealed answer list. After the transition, the three options in the current tray are the only available choices. Unknown identifiers, options belonging to another round, cursor movement outside the three-button tray, and actions attempted during the wrong phase are atomic no-ops.
You do not need to identify a real person, recall a name, recognize a celebrity, or interpret an expression. The illustrated target is displayed for only a moment, and the alternatives are fictional geometric drawings. The objective is exact visual recall within a short, deterministic route.
How the fictional profiles are constructed
The game contains fifteen immutable fictional profile specifications made from authored SVG puzzle features. No photograph, portrait file, stock image, external asset, or generated-person record is used. Each profile combines geometric elements from a small vocabulary:
| Profile element | Authored variations |
|---|---|
| Head outline | Round, oval, or square |
| Hair | Five shapes |
| Brows | Three shapes |
| Eyes | Three shapes |
| Nose | Three shapes |
| Mouth | Four shapes |
| Optional marks | Cheek marks or freckles |
Because the faces are combinations of code-drawn primitives, the useful approach is to notice a small cluster of memorable features rather than search for a real-world identity. The names Alder through Opal label fictional fixtures. They do not describe the drawings as people, assign demographic meanings, identify an emotion, or create a biometric template. Every profile is therefore a visual puzzle component with no identity claim attached to it.
Play the remember faces quiz
- Open the Face Memory Test. Visit the Face Memory Test game and wait for the first illustrated target to enter the Reveal state. The target contains one original fictional profile.
- Study the target. Examine its defining feature combination during the two visible one-second ticks. You can also choose Recall now if you feel ready before both ticks finish. Enter or Space performs the same early-hide action during Reveal. Answer buttons are not available in this state.
- Wait for the option tray. At the end of the two ticks, or immediately after Recall now, the target unmounts. Three large illustrated option buttons then mount in a single horizontal row. The target does not remain concealed in the active Reveal document.
- Select the matching profile. Click the button showing the same drawing, press 1 through 3 for direct selection, or use Left and Right to move the visible focus before confirming with Enter or Space. Up and Down deliberately do nothing because the answers form one row.
- Continue after the result. A correct match adds exactly 200 points and opens the next target. A first distinct wrong profile adds one recoverable error, but selecting that same wrong profile again during the same round is deduplicated. A second different wrong profile closes the run.
- Complete five clean matches or restart. Five correct matches produce the fixed 1,000-point finish. If the run ends before that point, the shared Restart control returns the game to the first target and restores its two-tick countdown.
Read the fixed route and exact score
The result sequence is deterministic. There is no shuffled answer position, random portrait, timer bonus, adaptive model, or hidden score adjustment. The correct option is fixed for every round:
| Round | Correct labeled option |
|---|---|
| 1 | 2 |
| 2 | 1 |
| 3 | 3 |
| 4 | 2 |
| 5 | 1 |
These positions are the labeled buttons shown to the player, not concealed metadata. Every clean match is worth 200 points. The complete-score calculation is 5 × 200 points = 1,000 points. Because neither the portraits nor their answer positions change, the published route is repeatable and can be audited against the game’s fixed fixtures.
Use the pointer or keyboard controls
The pointer path centers on three large illustrated buttons with textual option labels and visible keyboard focus. Cursor movement outside the three-button tray has no effect. During Reveal, the Recall now control ends the countdown early. Pressing Enter or Space also hides the target.
After the options appear, keys 1, 2, and 3 select an answer directly. Left and Right move the visibly selected option within the single-row tray, and Enter or Space confirms it. Up and Down are intentionally inactive because the game has no vertical option navigation. These controls provide the same choice mechanism as the pointer path rather than creating a separate scoring system.
Status text reports the current round, match total, and error count without relying only on color. The SVG drawings use stable view boxes that reserve their dimensions and avoid layout shifts, while semantic theme variables keep the illustrations visible across the site’s supported themes. Those presentation details do not change the five-round route or the fixed score.
Understand the two-miss limit
The error count is based on distinct wrong profile IDs within a round, not the number of clicks. If you select profile A and make a mistake, then select profile A again, the second action is treated as the same miss and does not increase the count. If you then select a different profile, B, that second distinct miss ends the run. A repeated selection after the terminal state is ignored.
This rule allows recovery from a single wrong profile while keeping the challenge short. It also prevents a rapid sequence of repeated clicks from artificially adding errors. Actions made outside the current option tray or while the game is transitioning are no-ops, so stale choices cannot influence the active round.
Know what the score does and does not mean
The 1,000-point result records completion of this exact illustrated recall game. It does not recognize a face, estimate anyone’s identity, infer sensitive attributes, or claim to measure general memory ability. It also cannot diagnose prosopagnosia, label a demographic group, or classify a real person from the fictional drawings. The score is a game result, not a clinical or biometric assessment.
Privacy follows the same local design. No camera, image file, account, API, machine-learning model, cookie-based identity, or network request is involved. The SVG illustrations are generated by the game rather than downloaded, and only the shared best completed score can remain in the browser.
Try another fixed visual challenge
For a related task with a different kind of visual information, Visual Memory Test asks you to recall five fixed grid patterns, while Pattern Memory Game presents five original 3×3 patterns with four visual options. These are separate games with their own fixed routes, controls, and scoring; the Face Memory Test itself always uses five illustrated profiles, three choices per round, and a 1,000-point clean finish.