12 Questions People Get Wrong [Trivia Traps]

The questions people get wrong most are rarely the obscure ones. They are the ones where a famous shortcut fights the real answer and wins.

This list comes from a scan of LearnClash production answer data on 2026-07-27. The scan read 39,113 question documents, found 24,065 with first-encounter answers, and kept the 267 questions that had at least 30 first-time answers and a clean version history. The rows with a clear wrong-answer pull then went through a fact-check, and the 12 below are the most-missed of the ones that came out clean in that July check, with wrong rates from 64 percent up to 90. One of them, the clams row, was demoted to contested by the August 2026 re-check of this page; it stays in the table flagged as such rather than being swapped out quietly, and the row’s entry explains why.

A correction first. An earlier version of this page ranked questions from a May 2026 export with a 10-answer minimum. That floor was too low: a 100 percent miss rate at 12 answers is an anecdote, and nine of the old twelve sat at samples between 11 and 15. Only three of the original questions reached the new 30-answer floor, and one of those, the Santayana quote, was pulled for a wording problem I explain below, which leaves the dung beetle and pure ethanol as the two survivors in the ranking. The old number one, the Bulgarian head-nod, is still missed by every player who has seen it, all 13 of them as of July, and 13 is not a ranking. Everything else here is new, and every number is from the July scan.

How LearnClash Ranked Its Most Missed Questions

Method, in one line: a bounded read-only scan of LearnClash production answer data on 2026-07-27, counting each player’s first answer to each question in duels (no timeouts, no joker-assisted answers) and ranking by the answer-weighted share that was wrong, with a 30-answer minimum per question.

Because only first encounters count, each player appears once per question. Nobody’s tenth attempt at a fact they already learned dilutes the rate. The n in each row is the number of first-time answers, and the “most-picked trap” is the single wrong option chosen most often, with its share of all answers.

The hardest-questions data page exposes the broader difficulty ranking. The misconception dataset answers a different question: which specific wrong answer attracted the most choices, with the published CSV and its limits attached.

The 12 Most Missed Questions on LearnClash

These are the 12 questions with the highest first-try wrong rates among the fact-checked misconception rows of the 2026-07-27 LearnClash scan, ordered from most missed to least. Row 2 shows the app’s answer key, which the August re-check rates as contested (its entry explains why it is listed with that flag and not as a nutrition fact). It is not the 12 highest wrong rates in the whole data set; the rows that out-missed these but were held back are named in the section on numbers I would not headline. If you want the longer version of this ranking, the 50 hardest trivia questions page runs the same miss-rate method across a full four-round set.

Source: read-only scan of LearnClash production answer data, 2026-07-27. Wrong rate is the share of incorrect first-encounter answers; n is the number of first-time answers.

1. Why stars twinkle (90% wrong, n=31)

This row is the reason I trust the whole data set. The correct answer, that Earth’s atmosphere bends starlight on the way down, polled 9.7 percent. Four options are shown per question, so blind guessing alone would score 25 percent.

Players did not merely fail to know the answer. They were pulled to a wrong one: 29 percent chose “Stars’ atmospheres bend their light rays.” Stars do have atmospheres, but the shimmer is made in ours, in pockets of air at different temperatures and densities. The star’s own light is steady; as Caltech’s Cool Cosmos puts it, stars do not really twinkle.

2. Clams beat beef liver for iron, on paper (89% wrong, n=37)

Beef liver is the canonical iron food, so a shellfish outdoing it feels wrong, and most players who guessed a shellfish picked oysters. The answer key comes from an older generation of USDA data. The 2005 Dietary Guidelines for Americans ranked canned drained clams first among iron sources, at 23.8 milligrams per 3-ounce serving, with organ meats such as liver at 5.2 to 9.9.

That ranking has not aged well, something the August 2026 re-check of this page turned up. A 2009 study in the Journal of Food Composition and Analysis bought canned clams off Honolulu supermarket shelves and measured 1.4 milligrams of iron per 100 grams on average for chopped and minced products against 30.5 for whole baby clams, a spread that no label predicted. USDA’s current Standard Reference entry for canned drained clams (FoodData Central 171976) sits at 2.68 milligrams per 100 grams, and Healthline’s iron ranking now puts a 100-gram serving of clams at up to 3 milligrams, under liver’s 6.5.

So the row is contested, and I would rather say so than quietly drop it. Whole baby clams beat beef liver on total iron by a wide margin, chopped and minced clams do not, and clam iron is mostly non-heme, which the body absorbs less readily than the heme iron in liver. The 89 percent miss rate is real, but part of it belongs to the question, which needs a wording review before anyone cites it as a nutrition fact. That is why the table flags this row: it measures the trap players fell into, not a settled nutrition ranking, and the question is queued for a rewrite in the app.

3. Ethanol sits between carbs and fat (88% wrong, n=41)

Alcohol never registers as an energy source, so the most-picked answer was dietary fiber, at 34 percent. Only 12 percent of players knew that pure ethanol delivers about 7 calories per gram, between carbohydrate at 4 and fat at 9, as Medical News Today lays out. This question survived from the May version of the list, and with a bigger sample its rate barely moved.

4. The shortest war lasted about 40 minutes (82% wrong, n=33)

The Six-Day War drew 42 percent of answers, more than double the 18 percent who got it right, because six days already sounds impressively short for a war. The Anglo-Zanzibar War of 27 August 1896 was over the same morning it started. The British ultimatum expired at 9:00 am, the ships opened fire at 9:02, the sultan’s palace stopped resisting around 40 minutes later, and histories quote figures between 38 and 45 minutes.

5. The dung beetle out-pulls the ant (80% wrong, n=35)

Ants are the cultural shorthand for tiny-but-mighty, and the leafcutter ant took 43 percent of answers against 20 percent for the dung beetle. The horned dung beetle can pull over 1,000 times its own body weight, per Science. This is the other survivor from the May list, and it doubles as a lesson in sample sizes: at 12 answers it looked like a 92 percent miss, at 35 answers it settled to 80.

6. Early samurai were archers first (78% wrong, n=32)

No other question with 30 or more answers tilts this hard toward one wrong answer. 56 percent of players chose “curved longsword” and 22 percent chose the bow; the scan’s own binomial check puts the odds of that split arising from random guessing at well under one in a million. The yumi bow was the mainstay of the samurai arsenal long before the sword, and their martial way was known as kyuba no michi, the way of horse and bow, as TheCollector explains. Pop culture rewrote the job description.

7. Dogs see yellow and blue (73% wrong, n=41)

Red and green pulled about 41 percent of answers, and those are precisely the colors dogs cannot tell apart. Dogs are dichromats with two cone types. They see the world in yellows, blues, and grays, roughly like a person with red-green color blindness, as the American Kennel Club describes.

8. Rickets, not osteoporosis (73% wrong, n=37)

Osteoporosis is the well-known bone disease, so it soaked up 32 percent of answers, but it is a disease of aging bone. The childhood condition caused by vitamin D deficiency is rickets, which leaves growing bones weak and soft and typically shows as bowed legs, per the NHS.

9. A Venusian day outlasts its year (70% wrong, n=37)

Players reached for the giant planet, Jupiter, at 32 percent. Venus is the odd one: it takes about 243 Earth days to rotate once but only about 225 to orbit the Sun, and it spins backwards, per NASA.

10. Cleopatra’s native tongue was Greek (66% wrong, n=103)

This question has 103 first-encounter answers, more than any other question in this top 12, so its numbers are on firmer ground than anything else on the page. A quarter of players answered Egyptian. Cleopatra was a Ptolemy, a Macedonian Greek dynasty that ruled Egypt for nearly three centuries without learning the local language; the World History Encyclopedia has the dynasty background, and Plutarch’s Life of Antony records her as the first of her line who learned it.

11. “Play it again, Sam” is never said (66% wrong, n=32)

Nobody in Casablanca says “Play it again, Sam.” Ilsa says “Play it, Sam,” and the misquote grew from there, as Phrase Finder documents. The most-picked wrong answer was “Here’s looking at you, kid,” which Bogart genuinely does say, so 28 percent of players picked the line that is famous for being real over the line that is famous for being fake.

12. Sloths climb down to defecate (64% wrong, n=33)

Once a week a sloth leaves its tree, descends to the forest floor, and relieves itself, despite the predator risk. 39 percent of players guessed it was to drink fresh water. The leading scientific explanation involves a mutualism between sloths, the moths in their fur, and the algae they graze from it, as Scientific American reports.

Why Smart Players Miss These Trivia Questions

LearnClash players miss these 12 questions because each collides with a stronger story, and the July data quantifies the collision: in 21 of the 267 LearnClash questions with 30 or more first-encounter answers, one specific wrong answer out-polled the correct answer. The published dataset keeps 10 of those 21 after the fact-check; the other 11 sat inside sampling error at 30 to 40 answers. That is not ignorance. That is confident, shared, wrong knowledge.

The misses cluster into three patterns:

A miss like this is worth more than an easy correct answer. Getting it wrong, seeing why, and meeting the question again later is retrieval practice with a built-in emotional spike, the same mechanism covered in our testing-effect guide. In LearnClash a missed question does not vanish; it comes back through a three-stage spaced repetition system after you have had time to forget it.

The Numbers I Would Not Put in a Headline

The stat that would have led this page is the Santayana quote question. 48.4 percent of players credited “those who forget the past are condemned to repeat it” to Winston Churchill and only 3.2 percent to George Santayana, at n=31, the widest gap of any question with 30 or more answers. I wanted it at the top. The fact-check killed it: our question text quotes a paraphrase that matches neither Santayana’s exact 1905 wording nor Churchill’s later version, which makes “Churchill” a defensible answer as asked. The question is queued for a rewrite, and until that ships the stat stays out of the ranking.

The Roman urine mouthwash question (88 percent wrong) went the same way, because historians file the practice under myth. Six more rows sit in the same holding pen for wording fixes, from a hens’ teeth question that only holds for mutant embryos to a Great Red Spot question whose two-to-three-Earths answer is only true historically, since the storm has been shrinking.

A second kind of row was left out on purpose. Questions like “What was the name of the grueling Spartan training system?” run wrong rates near 76 percent, but the wrong answers scatter across every option with no dominant pull. That measures a knowledge gap, not a trap, and this page is about traps.

Two fact-checked rows also sat just above the sloths in wrong rate: Thucydides explaining events by human rather than divine causes (65 percent wrong, n=43) and Persephone’s pomegranate seeds (64.5 percent, n=31). Neither made a publish tier in the July summary, so I left them out rather than promote them after the fact.

Two rules I now follow: never publish a rate without its n, and treat margins under about 15 percentage points at samples of 30 to 50 as noise. That second rule is why entries 8 through 12 are presented as most-missed questions and not as “the wrong answer beat the right one” claims, even where the trap edged ahead in the raw counts.

So this page shrank in confidence while growing in evidence. The May version had flashier numbers on thinner ice. For the wide-angle picture of what the player base does know, the LearnClash statistics page has the corpus numbers, and the history trivia questions and space trivia questions lists give you more rounds in two of the categories that show up above.

Or meet these 12 the way our players did, mid-duel with no answer column to peek at. Get LearnClash and see which traps get you.