A rhyming words game compares ending sounds, not ending letters, and the only way to score 1,000 points across five rounds is to match by the phonemes from the last stressed vowel onward. The Rhyme Match game runs five fixed four-choice rounds, awards 200 points per correct answer, and finishes on exactly 1,000 — a result you reach by selecting the correct candidate for light, cake, moon, train, and blue. Each candidate list contains exactly one word whose stressed vowel tail matches the target word's CMUdict tail, while the other three fail for a specific phonetic reason such as a different stressed vowel or a different final consonant. Because the correct option appears at positions 1, 3, 0, 2, and 1 across the five rounds (using zero-based indexing), you cannot clear the game by tapping the same slot every time. Spelling alone will mislead you: rough and cough share the written ending "ough" yet rhyme by no CMUdict measure, and moon and moan share the same final letter without sharing the same ending sound.
What a Rhyming Words Game Actually Tests
Most "rhyming words game" pages collapse the concept into a spelling shortcut: same ending letters, marked correct. A genuine rhyming words game tests ending sounds, and a sound-based definition has to say exactly which sounds count. Rhyme Match uses the CMU Pronouncing Dictionary as its reference. For each target word, the game locates the last vowel that carries primary stress (1) or secondary stress (2), then collects every phoneme from that vowel through the end of the word. The resulting sequence is called the stressed tail. A candidate only counts as a rhyme if its stressed tail is identical to the target's and its pre-tail onset is different, so a word paired with itself is automatically excluded.
Because the rule is explicit, it can be tested. The maintained CMU Pronouncing Dictionary repository and its raw dictionary file serve as the primary ARPAbet sources, and every one of the ten words the game plays is independently cross-checked against a Cambridge Dictionary page. The rule, the source files, and the cross-checks are all auditable, which is what lets the game claim an exact 1,000-point perfect score rather than a probabilistic one.
Why Spelling Fails and Three Boundary Pairs Show It
The most common reason people get a rhyming words game wrong is that they trust spelling. Three boundary pairs make the failure visible. Rough and cough share the conspicuous written ending "ough," yet CMUdict records their tails as AH1 F (rough, R AH1 F) and AA1 F (cough, K AA1 F), so the stressed vowels differ. Light and late both finish with the phoneme T; their tails are AY1 T and EY1 T respectively, again differing at the stressed vowel. Moon and moan both end with N, but their tails are UW1 N and OW1 N — same consonant, different vowel. Each example is exactly the kind of distractor the game uses to keep spelling from working as a rhyme oracle.
| Pair | Spelling match? | Stressed tails (CMUdict) | Why the rhyme fails |
|---|---|---|---|
| rough / cough | Both end in "ough" | R AH1 F vs K AA1 F | Different stressed vowels (AH1 vs AA1) |
| light / late | Both end in "t" | L AY1 T vs L EY1 T | Same final T but different stressed vowels (AY1 vs EY1) |
| moon / moan | Both end in "n" | M UW1 N vs M OW1 N | Same final N but different stressed vowels (UW1 vs OW1) |
Because the game uses these failure modes as built-in distractors, your only reliable strategy is to listen for or recall the actual pronunciation rather than to read the letters.
How to Play Rhyme Match in Five Rounds
The whole game is five rounds long. Each round shows one target word at the top and four neutral candidate words below. You pick the candidate that shares the target's complete ending sound from the last stressed vowel onward, then submit.
- Open the Rhyme Match page and let the first target word load; the page is self-contained and runs entirely in the browser with no download, account, or audio.
- Read the target word, then mentally or aloud isolate the vowel that carries primary or secondary stress and pronounce everything from that vowel to the end of the word — that sequence is the stressed tail you are trying to match.
- Tap, click, or focus the candidate whose pronunciation produces the same stressed tail; the page will not reveal the correct tail before you submit, so this step is purely phonetic.
- Select using a finger, mouse, pointer, keys 1 through 4, the arrow keys, or Space, then choose Submit match or press Enter; options use one column on narrow screens and two columns when space permits, and every control is at least 44 pixels high.
- Watch the page response: a correct submission reveals a short review note with the previous pair and its verified CMUdict tail; a wrong submission does not reveal the answer, and repeating the identical wrong selection is a no-op while a second distinct wrong selection closes the run until Restart.
- Clear five rounds in this manner. Each correct round adds 200 points and a perfect run ends on exactly 1,000.
The Five Rounds and Their Verified ARPAbet Tails
Every round in the game is a literal fixture. The targets, in order, are light, cake, moon, train, and blue. Their correct partners — night, lake, spoon, rain, and glue — and their ARPAbet tails are fixed, and the correct option sits at position 1, 3, 0, 2, and 1 in the five rounds respectively (zero-based). That distribution is what keeps a single repeated guess from clearing the game.
| Round | Target word | Correct candidate | Stressed tail (CMUdict) | Correct option position |
|---|---|---|---|---|
| 1 | light | night | AY1 T | 1 |
| 2 | cake | lake | EY1 K | 3 |
| 3 | moon | spoon | UW1 N | 0 |
| 4 | train | rain | EY1 N | 2 |
| 5 | blue | glue | UW1 | 1 |
Round-by-round cross-checks against Cambridge word pages confirm each pair's general English pronunciation, and the focused test contains its own literal pronunciation fixture alongside a separately written tail oracle that does not import runtime answer tables.
Controls, Keyboard Access, and Error Handling
Rhyme Match is built so that the same five rounds are reachable by touch, pointer, or keyboard. Touch and pointer users see four answer controls and a full-width submit button, all at least 44 pixels tall, laid out in one column on narrow screens and in a two-by-two grid where space allows. Keyboard users move through that grid with the arrow keys, select the highlighted choice with Space or keys 1 through 4, and submit with Enter. Text wraps inside options, and nothing forces a horizontal scroll, so the layout stays usable on small screens as well as wide ones.
Wrong attempts are handled by signature, not by count. The page hashes the current round and the chosen option into a short signature and refuses to record a duplicate. The first time you submit a wrong selection, the page simply reports it. A second, different wrong signature closes the run and the game waits for Restart. Invalid option indexes, submitting with nothing selected, moving the cursor past the option grid, and every action after completion or deadlock all preserve state without producing a third error. This means you can experiment freely without racking up phantom mistakes.
Scoring, Best Run, and Boss Key
Scoring is exact, not approximate. Each correct round adds 200 points, and the game awards exactly 1,000 only when all five rounds are answered correctly. Your best score is stored locally in the browser as an optional record and never leaves your device; if you prefer to reset, Restart wipes the current run while leaving the local best untouched. A double press of the Escape key opens a neutral boss screen overlay, which is useful when you want to step away from the keyboard without the round state being visible.
Because the page makes no network calls, loads no pronunciation audio, and adds no third-party dependencies, the entire experience fits in one tab. There is no microphone input, no account flow, and no telemetry, which keeps the rhyming words game fully private.
Honest Limits: Accents and What the Game Doesn't Claim
Rhyme Match makes a deliberately narrow claim. The five answers match under the cited CMUdict entries and the disclosed stressed-tail rule, and that rule has been independently cross-checked against Cambridge word pages for all ten played words plus the rough and cough boundary pair. The game does not claim that every English accent realizes every vowel identically, that spelling predicts speech, or that the five-round set is a universal rhyming dictionary.
Practical consequences follow from that scope. If your accent produces a different vowel for any of the ten words — for example, a merged pin/pen vowel that collapses AY1 and EY1 in some American Englishes — the stressed-tail rule may not match your intuition for that round. The same goes for blue and glue, whose tails UW1 assume the standard distinction between the two vowels. The interface states this limitation in plain language next to the rule itself, so you can adjust your expectation before you start. If you want to explore word games that rely on a different kind of match — meaning, opposites, or collocations — you can compare a rhyming words game to something like an opposite-words game that also scores 1,000, where the test is semantic rather than phonetic. For a quick reference on the pronunciation of any single word, the Cambridge page for blue or cake shows both British and American transcriptions you can compare against the in-game review note after a correct round.
For a deeper look, see Pop Bubbles Game Has No Popping Sound — Here's Why.