Skip to content
Lizely

Vocabulary Quiz

Match five words to four visible meanings while correct and wrong choices adapt the next candidate difficulty.

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

How to use

  1. 1.Read the visible word and all four candidate meanings, then choose the definition that matches the word most closely.
  2. 2.Use Tab plus Enter or Space, or tap a large definition button; a correct choice raises difficulty and a wrong choice lowers it.
  3. 3.Complete five correct rounds for exactly 1,000 points, while avoiding a second distinct wrong definition because it locks the session until Restart.

About Vocabulary Quiz

Vocabulary Quiz is a deterministic five-round definition-matching game with a small adaptive difficulty loop. Each round displays one English word and four visible candidate meanings. Choose the closest meaning to earn 200 points and advance. A completely correct five-round path therefore produces an exact score of 1,000. The game is designed as a short recreational challenge, not as a measurement instrument.

Adaptation changes the question experience rather than merely changing a label. The session starts in the middle of three bands. A correct selection raises the next band, up to the challenge ceiling. A wrong selection lowers the current band, down to the foundation floor, and the four visible candidate meanings change immediately. Easier sets use more clearly separated distractors, while harder sets place more closely related adjective senses beside the target meaning. The underlying prompt remains stable during a wrong-answer adjustment so the player can learn from the easier contrast.

Candidate positions are intentionally varied. The correct definition does not always occupy the first, last, or any single repeated location. Across the deterministic winning route, correct choices appear in four different button positions. This guards against position guessing and makes the player read the text. Every candidate is rendered as ordinary visible content; color alone is not used to communicate its identity.

Scoring is deliberately simple. Every correct round is worth exactly 200 points regardless of difficulty. Difficulty controls which candidates appear, not the reward. A correct choice immediately advances the round and resets that round's miss signatures. A player may complete a round after one wrong selection and still receive 200, but a second distinct wrong definition locks the session. This keeps the maximum easy to audit while preserving meaningful adaptive behavior.

Wrong submissions use normalized signatures. Candidate text is trimmed, converted to lowercase, and internal whitespace is collapsed before comparison. Submitting the same wrong definition again with different case or spacing is deduplicated and does not consume the second miss. A second different visible wrong definition sets deadlocked state. Once locked, all four candidate buttons are disabled and subsequent submissions are atomic no-ops. Restart restores the exact starting round, middle difficulty, empty attempt set, zero score, and unchanged deterministic candidate order.

Inputs outside the four currently visible definitions are ignored atomically. Empty values, non-string values, and stale definitions from a different difficulty set cannot alter the board, score, difficulty, message, or attempts. Completed state is protected in the same way. These boundaries are tested directly at the pure-logic layer, separate from the React interface.

The interface exposes only a public question object containing the visible prompt, current difficulty, and four candidate texts. It contains no correct marker, correct index, answer-key field, hidden solution, correctness class, or data attribute before grading. Buttons use neutral test identifiers based only on their visible position. Aria labels repeat the visible option number and definition text without adding correctness metadata. After a correct choice, the game moves directly to the next round instead of inserting an answer-key element into the page.

The eight spellings and displayed adjective senses are audited reference data. Princeton WordNet 3.1 is the primary lexicon. Its official HTTPS archive returned HTTP 200, and all entries were read from dict/index.adj. Independent Cambridge Dictionary pages returned HTTP 200 for every spelling and exact ordinary sense. Answer-specific URLs remain in the maintainer evidence manifest, while the public database uses only generic WordNet and Cambridge sources so citation links do not expose the question key before grading.

The automated test owns a private literal copy of all eight definitions, every candidate set across five rounds and three difficulty bands, the prompt sequence, and the expected winning indices. It verifies four unique visible options at every level, the presence of exactly one matching meaning, correct-up and wrong-down transitions, an actual candidate-set change after a wrong answer, varied winning positions, normalized duplicate-error handling, second-distinct-error deadlock, exact 200-point increments, exact 1,000 completion, terminal immutability, and deterministic Restart. Eight external golden cases connect each stored meaning to both dictionary sources without deriving expected text from product functions.

The final label is carefully scoped as a recreational session band. It is explicitly not a diagnostic, certification, standardized assessment, proficiency decision, educational placement tool, or real vocabulary-size estimate. Five fixed items cannot support those claims. The band only summarizes what happened in this brief game session and should not be used to make decisions about a person.

Everything runs locally in the browser as a client-side L1 game. Choices, difficulty changes, misses, and scores are not uploaded for grading, and no account is required. Native buttons support Tab, Enter, and Space for keyboard operation, while every definition option has at least a 44-pixel interaction height for touch. Minimum-zero grid tracks, wrapping definition text, breaking feedback, and an overflow guard keep the two-column desktop design from causing structural horizontal overflow on a 390-pixel phone. The shared GameShell supplies score, local best-score behavior, Restart, and Boss Key.

Methodology & sources

Eight adjective spellings and senses are checked in Princeton WordNet 3.1 dict/index.adj and independently on exact Cambridge pages retained in the evidence manifest. Public metadata uses generic source URLs to avoid revealing answer mappings. Questions have four visible candidates; correct raises difficulty by one, wrong lowers it by one, and each correct round awards exactly 200. Tests use private literal definitions and candidate matrices.

Frequently asked questions

Related tools

Mini Games guides

View all