The standard "which word is the odd one out" question asks you to pick the single word whose meaning does not belong with the others, and the Odd One Out Word Game turns that question into an eight-round pinned-sense route with a 12-second timer, fully disclosed WordNet senses, and a transparent quick bonus. Each round declares a semantic category in plain English, lists four numbered words, and asks you to select the one word whose Princeton WordNet 3.1 noun sense falls outside that category before the visible timer expires. After you answer or the time runs out, the page reveals short sense labels for every option so you can see exactly which sense was scored, then advances to the next round. The full run covers eight rounds, three supported members and one pinned distractor per round, and 32 pinned noun senses in total. The result is a clearly framed entertainment exercise that treats ambiguity as a visible limitation rather than a hidden grading trick, so the score you earn reflects only the pinned senses and the categories shown on the page.
What the Game Actually Asks You
The category appears in the prompt before you answer. The eight categories are fixed: birds, trees, edible fruit, musical instruments, table-cutlery items, weekdays, flat geometric figures, and emotions. Each round offers four numbered buttons with one short headword, and one of the four has a pinned WordNet sense that does not belong to the displayed category. Three of the four options have a pinned sense that does belong, so the correct answer is the sense that is outside the category, not the spelling that is outside an intuitive list.
This is a deliberate shift away from guessing. The headword "piano," for example, can also name a musical direction, and "eagle" can describe a golf score, but those senses are not the ones being scored here. The prompt names the category in plain English, the buttons pin to one sense per option, and the feedback then names the pinned sense so you can confirm what was actually being asked. The category table is treated as reference data, not generated from spelling or intuition at runtime, and independent tests freeze all 32 sense-key and offset pairs, require four unique option IDs, require three members and one distractor, and verify that the answer position varies across the route.
Try a pinned-sense round of the Odd One Out Word Game and compare your reading against the labels the page shows after each answer.
How to Play in Three Steps
- Start the eight-round route, read the complete displayed category, and inspect all four numbered words.
- Choose the one pinned sense outside that category before 12 visible seconds expire; number keys 1 through 4 match the buttons.
- Read all four disclosed sense labels, continue to the next round, and interpret the final score only within this repeated entertainment route.
You can click a button or press the matching number key. Native buttons remain available to keyboard, touch, pointer, zoom, and assistive technology, so the route is reachable without a mouse. A backward timestamp, a stale timeout, a repeated pause, a double answer, a held number key, an invalid option, an old feedback button, an old feedback button press before the next prompt opens, a pre-start action, and a terminal action are no-ops, so a stray input cannot change the recorded answer after the round has settled.
The Eight Visible Categories
The eight rounds cover a different semantic field so the route stays varied. The categories are birds, trees, edible fruit, musical instruments, table-cutlery items, weekdays, flat geometric figures, and emotions. Each round shows exactly one of those categories, and each option pins a single WordNet noun sense that has been verified as a member of the category or, in one case, as the distractor outside it.
| Round category | What it groups | Disambiguator used in the prompt |
|---|---|---|
| Birds | Flying and perching bird species | Figurative senses such as a golf score are excluded |
| Trees | Tree species with a noun sense | Wood and figure senses are excluded |
| Edible fruit | Culinary fruits eaten as food | Edible fruit rather than botanical fruit |
| Musical instruments | Instruments that produce music by design | Music-direction senses such as piano are excluded |
| Table cutlery | Eating utensils used at a dining table | Table knife rather than bare knife |
| Weekdays | Named days of the working week | Calendar weekday names without regional variants |
| Flat geometric figures | Plane shapes with no thickness | Flat geometric figure rather than the broader word shape |
| Emotions | Named feelings or affective states | Figurative desire senses such as hunger are excluded |
The wording matters. "Edible fruit" rather than "fruit" excludes a botanical reading. "Table knife" rather than "knife" excludes a blade weapon. "Flat geometric figure" rather than "shape" excludes three-dimensional solids. Capitalized March is preserved as the calendar month rather than the verb to march. Those small disambiguators are part of the prompt, so a technically tidy round never quietly uses a different sense.
How Scoring Works
The game begins with 800 points, which is 100 points per round across the eight rounds. A correct answer retains the 100-point round value and adds five points for each whole visible second remaining, so a fast answer scores more than a slow one. A wrong answer or timeout removes the 100 points for that round. The total never goes below zero, so a missed round drops the total to 700 rather than into negative numbers.
The timer is based on monotonic visible elapsed time rather than callback count or device speed. A delayed browser timer recomputes the actual elapsed interval before it can settle a round, so a slow device or a long-running callback cannot silently rob you of seconds. If you hide the tab or open the spreadsheet-style boss screen, the active round pauses at the document boundary. When you return, the round remains paused until you explicitly resume with the preserved remaining time. Hidden or covered time is subtracted once, the old phase token becomes stale, and returning does not silently restart the clock.
| Event during a round | Round value | Speed bonus |
|---|---|---|
| Correct answer before timeout | Retained (100) | +5 per whole visible second remaining |
| Wrong answer before timeout | Removed | Not added |
| Timeout | Removed | Not added |
| Pause via hidden tab or boss screen | Preserved while paused | Remaining time preserved while paused |
| Resume after explicit resume | Unchanged | Unchanged |
The status line reports the round, errors, final quick bonus, and current score in words, so you always know what the page has recorded. No response is uploaded, and the shared game shell may store a best point total locally on your device, but there is no account, leaderboard, vocabulary profile, adaptive category generator, or remote model involved in the score you earn.
Why Polysemy Is Honest, Not Hidden
Most odd-one-out puzzles rely on the reader guessing which reading the author had in mind, but this game fixes that guesswork. The prompt names the category in plain English, the buttons pin to one sense per option, and the feedback names the pinned sense after you answer. Polysemy is a visible limitation rather than a hidden grading trick: eagle can describe a golf score, piano can be a musical direction, square can name a city place, and hunger can be a figurative desire, and those alternatives are not the senses being scored here. The page only claims that one displayed pinned sense is outside one displayed category, and never claims that the same spelling cannot relate differently in another literal, metaphorical, regional, or technical context.
This matters for the score you earn. A reader who knows that eagle also scores a golf term would otherwise waste seconds wondering which sense to compare. Here, the prompt explicitly names the category "birds," the pinned sense is a noun sense of eagle as a bird, and the distractor in a round is chosen so its everyday reading would not reasonably fit the category. The product rejects any distractor whose common headword sense would fit the displayed category, so you never see a round where the distractor would pass for a member under the everyday reading.
The same honesty applies to disambiguation. Capitalized March is preserved for the calendar month, table knife stands in for bare knife, edible fruit stands in for botanical fruit, and flat geometric figure stands in for the broader word shape. Each preserved disambiguator narrows the category to the sense being scored, which keeps the score you earn tied to the pinned sense rather than to whichever reading happened to cross your mind first. A second review against the Cambridge Dictionary entry for apple confirms the everyday edible-fruit reading that the edible-fruit round is scored against.
What the Feedback Labels Show
After your answer or timeout, the page reveals short sense labels for all four options before the next round opens. These labels come from the same pinned WordNet sense that the round was scored against, so you can compare the member senses against the distractor sense and confirm what the round was actually asking. If a sense label does not match the reading you had in mind, that is a visible limit of the category rather than a grading error, and the page itself answers that the decision applies only to the pinned sense and visible category.
The route is a repeated entertainment route, not a clinical or cognitive assessment. The result does not measure lexical retrieval, semantic fluency, vocabulary size, language proficiency, intelligence, cognition, cognitive health, diagnosis, school placement, or improvement. Repeating the eight known rounds naturally changes short response time, so the score should be read as a transparent result from this fixed game rather than as a stable personal metric. Best point totals may be stored locally on your device, but no account, leaderboard, vocabulary profile, adaptive category generator, or remote model is involved in the score you earn.
Semantic categories are derived from Princeton WordNet 3.1, whose complete Copyright 2011 notice is published with the site. The linked Cambridge Dictionary headword pages act as independent ordinary-sense checks only, and the Cambridge text and examples are not copied. No response is uploaded, and the shared game shell keeps the rest of the bookkeeping on the page itself. The status line reports the round, errors, and final quick bonus in words, so you always know what the page has recorded before you close the tab.
For a deeper look, see Phrasal Verbs vs Single Word Verbs: An Audited Quiz.
For a deeper look, see Simple vs Complex Sentence Order in a Word Game.