To match the synonyms in a verified way, Synonym Match pairs each prompt word with one candidate that shares the same part of speech and the same disclosed sense, then checks every winning pair against Princeton WordNet 3.1 and the Cambridge Thesaurus. The game presents five fixed English rounds, each with one prompt word, its grammatical category, a short sense note, and four candidate words. Players choose the single candidate that can express the same meaning in the same grammatical role. A correct match awards exactly 200 points, so five correct rounds reach exactly 1,000. The interface runs entirely in the browser, accepts touch input and full keyboard control, and freezes at completion. Because every prompt, candidate, and verified answer is sourced rather than guessed, the game is recreational vocabulary practice that gives a reproducible answer for each round rather than an adaptive assessment.
How to Play Synonym Match
Every round of Synonym Match shows the same three-block layout: a prompt word, a part-of-speech label, a short sense note, and four candidate words. The task is to choose the one candidate that can replace the prompt without changing the disclosed meaning or the grammatical category. The following steps walk through a single round and a complete run.
- Read the prompt word, its part of speech, and the short sense note so you match the intended meaning rather than a broad association.
- Look at the four candidate words and identify the one that can express the same disclosed meaning in the same grammatical role.
- Tap the candidate on a touch device, or move the keyboard cursor with Left or Up to go to the previous candidate and Right or Down to go to the next. Selection wraps at either end.
- Press Enter on the keyboard to submit the highlighted candidate. On touch, the tap is itself the submission.
- Match all five fixed rounds for exactly 200 points each. Five correct rounds produce exactly 1,000 points.
The first distinct wrong choice keeps the round available so the player can reconsider. Selecting the same wrong candidate again does not consume another mistake because the mistake is recorded as a deterministic signature made from the round and candidate index. A second different wrong choice deadlocks the run until Restart, which restores round one, candidate one, zero score, and no wrong signatures.
How Synonym Match Verifies That Two Words Match
Synonym Match does not rely on a generated similarity score or an uncited list assembled from intuition. Every winning pair is checked in two independent reference sources. The primary check uses Princeton WordNet 3.1, whose dictionary data groups words into synsets that represent a shared lexical concept and records the part of speech for each entry. For a candidate to win, the prompt and the selected answer must occur in the same relevant adjective or verb synset. This is stronger evidence than a numerical similarity because it requires both words to be listed under the same lexical concept rather than merely being close in vector space.
The Cambridge Thesaurus provides a second cross-check. Each of the five prompt entries was fetched successfully during implementation, and the matching candidate was found in its page content. The comparison is performed for the intended sense shown in the interface. This matters for polysemous words, where the same spelling carries different meanings in different contexts. The disclosed sense pins each round to a single meaning so that a different sense of the same prompt cannot be matched accidentally.
Synonym questions become unreliable when a game treats every related word as interchangeable. A word about speed is not automatically synonymous with every word about motion. A word associated with service is not automatically equivalent to every action performed near another person. For that reason, the game rejects candidates that are merely topical, opposite, or grammatically mismatched. Other related word-matching games apply the same pinned-meaning approach to a different lexical property, such as the rhyming approach described in Match Rhyming Words by Their Stressed Sound Tail, which judges words by stressed sound tails rather than spelling.
Touch and Keyboard Controls
Pointer and touch play are direct. Every candidate is a native button at least 44 pixels high, uses wrapping text, and sits in a one-column mobile grid that expands to two columns on wider screens. Tap a candidate to submit it. The game container uses bounded width, minimum-width protection, and overflow containment so long words cannot force horizontal page scrolling on narrow phones.
Keyboard play covers the full task. Left or Up moves to the previous candidate, while Right or Down moves to the next. Selection wraps at either end of the candidate list. Press Enter to submit the highlighted candidate. The highlighted state represents only the player's keyboard cursor and does not reveal which candidate is correct.
The interface never exposes the answer through rendered attributes. All four buttons use the same structure, and correctness is not placed in visible text, screen-reader-only text, aria labels, data attributes, CSS class names, or ordering metadata. Candidate order is fixed and varied between rounds, so a repeated position pattern does not disclose the answer to anyone watching the screen.
The shared GameShell provides Restart and describes the controls alongside the score. Its double-Escape boss key switches to the spreadsheet-style cover and back without altering the round, the score, or the error history. The boss key exists so the player can step away from the screen quickly when needed and return to the same point in the run.
Scoring, Mistakes, and Restart
Each correct match awards exactly 200 points. Five correct rounds produce exactly 1,000 points. Completion freezes later input at exactly 1,000 points so accidental keystrokes cannot push the score beyond the verified total. The frozen state also blocks any invalid index, fractional index, or unknown cursor direction as an atomic no-op that changes nothing.
Wrong choices use an exact deterministic signature made from the round and candidate index. Selecting the same wrong candidate again does not consume another mistake. The first distinct wrong choice leaves the round available so the player can reconsider. A second different wrong choice deadlocks the run until Restart. This rule avoids penalizing an accidental duplicate event, such as a double tap or a repeat key press, while keeping guessing finite.
Restart restores round one, candidate one, zero score, and no wrong signatures. Best score is stored locally when browser storage is available, and the game sends no selected words to a server, requires no account, calls no paid API, and adds no package dependency. The following table summarizes how each round outcome affects points, the round state, and the run state.
| Round Outcome | Points Change | Round State | Run State |
|---|---|---|---|
| Correct selection | +200 | Advance to next round | Continues |
| First distinct wrong choice | 0 | Stays on the same round | Continues |
| Same wrong candidate repeated | 0 | Stays on the same round | Continues, no new mistake |
| Second distinct wrong choice | 0 | Deadlock | Requires Restart |
| Fifth correct round completed | Final total 1,000 | Frozen | Frozen, input ignored |
What Is Fixed and What Is Not
The five rounds are fixed rather than adaptive. Early and later rounds vary in word length, grammatical category, and how plausible the distractors may look, but the game does not claim to infer a player's vocabulary level or change difficulty from performance. A fixed fixture keeps the source evidence, score path, regression tests, and production behavior reproducible across every play.
Every round contains four unique candidate strings and exactly one verified answer. Distractors are ordinary English adjectives or verbs selected to be clearly non-equivalent in the disclosed sense. Some share a part of speech with the prompt, ensuring that players still need semantic judgment rather than dismissing options on grammar alone. Others demonstrate boundaries such as opposition or unrelated meaning. The game never accepts a broad association as a correct match, no matter how natural the connection feels.
Synonym Match is recreational vocabulary practice. It is not an adaptive assessment, a language certification, a reading-level diagnosis, or a measure of intelligence. Focused tests keep the reference and product layers separate. An independent literal fixture contains externally checked true pairs plus antonym and unrelated-word boundaries. Production rounds are compared with those expected words, parts of speech, and senses. Additional tests cover four unique candidates per round, exact 200-point progression, duplicate-error suppression, two-error deadlock, cursor wrapping, invalid commands, terminal freeze, and Restart. The cited sources support the lexical relationships only; candidate order, scoring, interface, and mistake policy are Lizely product decisions.
Where the Game Sources Its Word Pairs
The lexical evidence behind Synonym Match comes from two public references. Princeton WordNet 3.1 supplies the synset groupings and the part-of-speech tags. Each winning pair in the game has been checked against the same relevant adjective or verb synset in that data. The Cambridge Thesaurus supplies the independent cross-check for the disclosed sense. Sample pages such as assist and begin confirm the same-meaning relationship at the sense shown in the interface.
The sense pinning is what separates this game from a topical word-matching game. Two words may be loosely associated without being interchangeable. Synonym Match treats that distinction as the core of the task. The player who learns to read the part-of-speech label and the sense note will answer correctly more often than the player who only spots a thematic connection.