Skip to content
Lizely

Find the Mistake Game

Inspect five original visual-rule boards, find one unambiguous anomaly on each, and finish the deterministic route on exactly 1,000 points.

Privacy: your files never leave your device. All processing happens locally in your browser.

How to use

  1. 1.Read the rule above the board, then compare all twelve visible details for direction, count, or stripe orientation.
  2. 2.Tap a tile, or focus the board and use the Arrow keys to scan before selecting with Enter or Space.
  3. 3.Find the one rule-breaking detail on each of five fixed boards to earn 200 points per board and exactly 1,000 total.

About Find the Mistake Game

Find the Mistake Game is a five-round visual anomaly hunt built from original geometric fixtures. Each board presents twelve code-drawn details in a four-column by three-row grid and states one plain rule. Eleven details follow that rule, while exactly one breaks it. Your task is to compare the visible details, select the single exception, and clear all five boards for a fixed maximum score of 1,000.

Every round changes both the visual object and the property that matters. Signal Yard asks you to compare arrow direction: every signal should point northeast. Dot Registry uses circular badges and asks you to count three dots. Gate Ledger asks you to compare the number of horizontal bars. Wheel Workshop changes the count of spokes. Flag Archive asks whether a diagonal stripe rises from left to right. The anomaly is always visible through the named property, so no board depends on color alone, external knowledge, a trick question, hidden animation, or an unstated convention.

The five fixtures are deterministic. There is no shuffled layout, random anomaly, daily challenge, network response, or adaptive difficulty. Starting the game again gives you the same five auditable boards and the same clear rules. This makes the result reproducible across pointer, touch, and keyboard play. It also lets independent tests define their own literal twelve-cell geometry and verify that every production board has exactly one rule violation.

Pointer and touch users select directly on the canvas. The interaction converts the visible position to a logical grid cell, including deterministic behavior at the outer edge and exact row or column boundaries. Keyboard users focus the board, move a dashed scanner with the Arrow keys, and select the current tile with Enter or Space. The scanner's accessible label describes the current tile's public visual property, such as its direction or count. Both input paths finish by calling the same pure cell-selection reducer and therefore create the same success result or wrong-cell signature.

The artwork is rendered into one responsive canvas instead of twelve answer-bearing hotspot elements. The page does not create hidden target buttons, correct-cell attributes, answer classes, coordinate attributes, or per-tile test IDs. Only the currently scanned cell is described for keyboard access. The canvas backing store follows the rendered CSS box, caps device-pixel scaling for predictable memory use, and redraws through ResizeObserver when its size changes. At a narrow mobile width, the four equal columns remain inside the container and each visible cell remains comfortably larger than a 44-pixel touch target.

A correct selection awards exactly 200 points and immediately advances to the next authored board. Five correct selections produce exactly 1,000 points. A normal tile is a wrong selection, but the first distinct wrong cell is repairable. Selecting that same cell again does not consume another error because the signature combines the current round and logical cell. Selecting a second different normal cell closes the run in deadlock. Finding the anomaly after one miss advances normally and clears that repairable round error.

Completion and deadlock are terminal. Later pointer selections, keyboard selections, cursor movements, or invalid coordinates cannot change the score, round, feedback, or error history. Restart returns the entire state to the exact initial value: first board, first scanner cell, zero score, no errors, and no terminal flag. The shared GameShell also retains its standard local best-score display, Restart control, and double-Escape Boss Key behavior.

The interface uses readable text in addition to geometry. Each board prints its rule above the canvas, the status reports board number and distinct misses, and an aria-live message explains a correct selection, a repairable miss, completion, or deadlock. Focus is not moved on initial load. When completion or deadlock first occurs, focus moves to the terminal status so keyboard users receive the result. System-modified shortcuts are ignored, repeated keydown events do not submit multiple selections, and all controls use visible focus styling.

Independent tests keep a separate literal oracle for the five twelve-cell boards. They verify the sole violating cell for every rule, cell centers, exact row and column boundaries, the inclusive far canvas edge, invalid coordinates, cursor boundaries, pointer and keyboard equivalence, repeated-miss deduplication, second-miss deadlock, repair after one miss, exact 200-point progression, exact 1,000-point completion, Restart equality, and frozen terminal states. A source scan also rejects answer-bearing data attributes and requires the single hidden canvas plus shared pointer and keyboard entry points.

Everything runs locally in the browser. The game sends no selections to a server, uploads no image, requests no hardware permission, calls no paid service, and adds no package dependency. The shapes, rules, layouts, scoring, and error policy are Lizely-authored entertainment fixtures without external reference meaning.

Find the Mistake Game is recreational entertainment. It does not measure visual acuity, attention, intelligence, reaction time, cognition, educational achievement, employability, neurological function, or any physical or mental health condition. A score only records completion of these five fixed boards under their printed game rules.

Methodology & sources

Define five original twelve-cell boards and one public normal-value rule per visual kind, without storing an answer or correct-cell field. Determine a violation by comparing the selected cell's public property with that kind's normal value. Convert pointer coordinates to the same logical four-by-three cell used by keyboard selection, with deterministic exact-boundary and inclusive far-edge mapping. Award exactly 200 points for the sole violating cell, clear the repairable error on advance, and cap five rounds at 1,000. Sign wrong choices by round and cell so the same selection deduplicates while a second distinct normal cell deadlocks. Freeze every selection and cursor reducer after completion or deadlock, and make Restart equal the initial state. Render the public fixtures into one responsive canvas with no answer-bearing hotspot DOM. Verify all rules and geometries against independent literal oracles plus source-side-channel scans.

Frequently asked questions

Does every board have exactly one mistake?
Yes. Each fixed twelve-cell board contains eleven details that follow its printed rule and one visible detail that violates it.
How do keyboard controls work?
Focus the board, move the scanner with the Arrow keys, and use Enter or Space to select the current cell. Pointer, touch, and keyboard use the same selection reducer.
Does the score measure vision, attention, or intelligence?
No. The score only records completion of five original entertainment boards. It is not a visual, cognitive, educational, workplace, medical, or psychological assessment.

Related tools

Mini Games guides

View all