A partially hidden image game is a reveal-and-guess puzzle where each round starts with a scene covered by opaque tiles that you remove one at a time to expose small fragments, then commit to one of three titles before the possible score falls. Picture Reveal Guessing Game uses a deterministic three-by-three cover over five original code-drawn scenes — Moon Garden, Kite Harbor, Cactus Camp, Teacup Train, and Lamp Library — and follows the visible score rule max(40, 240 minus 40 times the reveal count). One revealed tile is worth 200 points, two 160, three 120, four 80, and any state from five through nine revealed tiles is locked at 40 points. A fully covered scene cannot be guessed, so the three title buttons stay disabled until you press R or click the Reveal next tile control. A correct title adds the currently displayed possible score and starts the next round fully covered, while two distinct wrong answers in the same round end the run. Everything runs in this page with no account, storage permission, camera, file picker, network service, remote random source, or personal data behind it, and only the shared best completed score can remain in localStorage.

partially hidden image game
Partially Hidden Image Game: Score Drops With Each Reveal

What a partially hidden image game actually does

A partially hidden image game sits between pure hidden-object searches and full picture quizzes: the artwork is already on the canvas, but most of it is masked by an opaque grid, and your only job is to read the clues that the visible fragments give you. Each round begins with a complete cover and zero information, and the artwork gradually appears in a fixed useful-first order that you do not control. The interesting decision is timing — guess now while the score is high, or reveal more visual evidence and accept fewer points.

Picture Reveal Guessing Game follows that contract exactly. The five scenes are SVG geometry authored inside the project with the site's color tokens, so there are no photographs, generated image files, stock assets, third-party icons, downloads, or external reference data. The names describe fictional scenes only, not real landmarks. That keeps every round visually distinct while keeping the puzzle fully transparent about what you are seeing.

The five original scenes you cycle through

Every run walks through the same five scenes in the same order, and the round counter advances one step at a time. Each scene carries its own three title buttons in fixed positions and its own nine-position reveal permutation, so the useful-first order is unique to that round even though the cover grid always looks the same. The masked canvas uses a generic accessible label and tile-position test IDs, and it does not announce the hidden scene title through alternate text while the active puzzle is covered, so a screen reader user gets the same reveal-and-name loop as a sighted player.

RoundSceneTheme
1Moon GardenNight flora and a glowing sphere
2Kite HarborTied sails and a docked boat
3Cactus CampDesert plants around a small tent
4Teacup TrainStacked cups on rolling rails
5Lamp LibraryBookshelves lit by hanging lamps

Restart returns you to a fully covered Moon Garden with zero score and no recorded errors, which makes the run easy to reset for practice between attempts.

How to play Picture Reveal Guessing Game

The mechanics are simple enough to describe in five steps, and every step has a single keyboard equivalent if you would rather not touch the mouse.

  1. Press R, or click the Reveal next tile button, to uncover the first useful fragment of the original illustration. The three title buttons stay disabled until at least one tile is visible, so a fully covered scene cannot score.
  2. Read the visible fragment and look at the three title buttons. If the fragment already names the scene, choose that title with keys 1 through 3 while the possible score still shows 200 points.
  3. If the fragment is too thin, press R again to remove the next opaque panel in the round's fixed permutation. Watch the possible score drop to 160 after two reveals, 120 after three, 80 after four, and 40 once five or more tiles are uncovered.
  4. Commit to one of the three titles. A correct choice adds the currently displayed possible score and advances to the next fully covered scene.
  5. Repeat for all five rounds. Reveal one tile per round, pick the right title every time, and you have the perfect 1,000-point route — 200 points multiplied by five rounds equals exactly 1,000.

Pointer and touch users receive the same labeled controls with practical target sizes, so the keyboard-only route and the tap-only route produce identical scores.

The score formula behind every reveal

The scoring rule is the most important thing to internalize before you start revealing tiles, because every additional click either buys evidence or costs points — never both. Each round starts fully covered with zero possible score, and the moment you uncover the first tile, the possible score jumps to 200 points. Every additional reveal subtracts 40 points from the ceiling, down to a hard floor of 40 points once five or more tiles are uncovered.

Tiles revealed in the roundPossible score for a correct guess
0 (fully covered)Locked — no guess allowed
1200 points
2160 points
3120 points
480 points
5 through 940 points

The formula max(40, 240 minus 40 times the reveal count) gives the same numbers as the table, and a correct title awards exactly the displayed possible score for that round. There is no timer bonus, no streak multiplier, no random shuffle, and no hidden modifier, so the highest reachable total is fixed: 200 points per round for five rounds equals exactly 1,000 points.

For a worked example, take the perfect route: round one reveals 1 tile for max(40, 240 − 40 × 1) = max(40, 200) = 200 points, and the same calculation repeats for rounds two through five. Adding 200 + 200 + 200 + 200 + 200 = 1,000 points. A cautious player can reveal more tiles and still complete the game with a lower, honestly calculated total, and the text beside the canvas repeats the current possible score, reveal count, round count, solved count, and errors without depending on the picture alone.

Keyboard, pointer, and touch controls compared

The control layer mirrors the deterministic design of the round. Every input does one thing and the game does not branch based on whether you used the keyboard, the mouse, or a touchscreen.

  • R — Reveal next tile. Removes one opaque panel in the round's fixed order.
  • 1, 2, 3 — Choose the matching title button. Disabled until at least one tile is uncovered.
  • Escape, Escape — Double-press to show the simulated spreadsheet boss screen, then double-press again to resume the same cover state.
  • Pointer and touch — Labeled Reveal next tile and three title buttons with practical target sizes. No drag, no long-press, no swipe.

Invalid choice indices, guesses before the first reveal, extra reveals after all nine positions, and actions after completion or deadlock preserve the state, so a mistimed key never costs you points.

Errors, deadlocks, and what restart actually does

The error system rewards reading the fragment before guessing. A single wrong title is recoverable — you pick a different title in the same round and the run continues from the same cover state. Repeating the same wrong title in the same round is deduplicated and the error count stays where it was, so spamming one button does not artificially inflate your mistakes.

A second distinct wrong title in the same round closes the run. After the second different miss, the canvas freezes at the current cover state, the title buttons are disabled, and you can only restart from a fully covered Moon Garden with zero score and no recorded errors. The independent logic test for the game proves that this terminal freeze and deterministic restart behavior is enforced in code, so a deadlocked run never advances or scores further. Only the shared best completed score can remain in localStorage; there is no account, storage permission, camera, file picker, network service, remote random source, or personal data behind the round.

The same project hosts several other original visual games that exercise different perceptual tasks rather than copying this progressive reveal-and-name loop.

  • Hidden Object Game — Search an original illustrated attic for five named objects. Useful when you want a pure visual search without a guessing decision on top.
  • Pattern Memory Game — Memorize five original 3×3 patterns and pick each hidden record from four options. Useful when you want a memorization layer instead of a recognition layer.
  • Shadow Matching Game — Compare five original specimens with shuffled silhouettes. Useful when you want to train shape abstraction rather than color and detail recognition.
  • Memory Match Game — Match six fixed symbol pairs on a deterministic twelve-card board. Useful when you want a low-pressure memory exercise without a scoring clock.