Show How Questions Will Be Written for Your Quiz Format
Show How Questions Will Be Written for Your Quiz Format
A quiz pitch usually arrives with five good questions. They land, the room laughs in the right places, and nobody asks where the sixth one comes from.
But five good questions prove taste. They do not prove a supply. Your format is promising something that runs past the pilot: every episode, a set of questions contestants can actually answer, that the show can defend when someone objects, that don't repeat last week's trick, and that don't quietly expire. That promise needs a method, and the method is usually the missing page.
Here is what to put on it.
Answering is a task before it is a topic
A category name is not a task. Neither is the adjective "accessible," which has never once told a question writer what to do.
Write down what the contestant has to perform. Recalling a fact is not the same operation as recognising a face in a photograph, which is not the same as working out which of four things happened earlier, which is not the same as joining two facts that never appear in the same sentence. Each draws on different knowledge, fails in different ways, and rewards a different kind of preparation.
"We ask about history" describes a subject area. "We name a city from three clues read at five-second intervals, with the third clue narrowing the field" describes a task. The second one tells a writer what to supply, tells you how long the material needs to be, and tells a viewer what they are watching someone do.
Then describe what the contestant can see and hear, and when. If the four options appear on screen while the question is read, the writer is writing for the eye and the ear at once. If they appear only after the clock starts, the reading is a beat the round has to budget for. Those are different writing jobs. Say which one you're doing.
Build a miniature category, not four favourites
Four questions chosen because they're good is a tasting menu. Four questions chosen to show the range of a round is a sample.
The example below is invented — a fictional show, a fictional city, a fictional metro network — so the mechanics stay visible without borrowing anyone's actual question bank.
Say the show is Waypoint, a proposed half-hour quiz. Round three is "In Order." Four cards appear. The prompt names a property. The team has thirty seconds to arrange the four cards according to it. Then the reveal shows the correct sequence and the fact behind each card.
A sample category, "Green Line openings":
1. Chronology. Cards: Cinder Row, Dunmore, Fenwick Cross, Gorse Lane. Order by the year each entered public passenger service.
2. Measurement. Cards: Bramble Hill, Ashgate, Cinder Row, Fenwick Cross. Order by depth below street level. Nothing here rewards date recall; the player is reading a figure out of a shape.
3. Recognition. Four photographs of platform tiling. Order by the decade the station was built. The evidence is visual, and the writing job is choosing photographs that are legible and distinguishable at broadcast size.
4. Connection. Four river crossings. Order by the year a metro train first used each. No single card can be answered from one fact; each requires joining a bridge's completion to a line's opening.
Same round, four different things asked of a player, and no chance to learn one trick and repeat it. That is what a sample range is for. Notice that nobody has called anything easy, medium or hard — those words are a claim about an audience, and you have not tested one yet.
Keep the presentation constraints in view while you do this. A genuinely good standalone question can still be wrong for the round. "Which Vellick Metro station has the longest platform?" is fine trivia, but the round requires an ordering of four cards, and this produces a single name, so it cannot sit here. A forty-word prompt is another failure: the round allows roughly eight seconds of reading before the clock starts, and a prompt that long eats the thinking time it was supposed to protect.
A supported answer is not the same as an unambiguous question
This is the distinction most pitch decks miss entirely.
You can prove your intended answer and still have a second defensible answer. Support for one reading is not exclusivity.
Suppose the first draft of the chronology question read: "Which of these stations opened first?" It's short, it sounds like a quiz question, and it has a problem. "Opened" covers at least two different events. Cinder Row entered public passenger service in 1912. Dunmore's site carried engineering trains from 1910 and opened to passengers in 1913. Under the passenger reading, Cinder Row is first. Under the first-train reading, Dunmore is. The order flips depending on a word nobody defined.
Narrowed:
Prompt: Put these four stations in order of the year each entered public passenger service. Intended answer: Cinder Row (1912), Dunmore (1913), Fenwick Cross (1930), Gorse Lane (1955). Boundary stated: "public passenger service," not first use of the site. Alternative reading checked: Dunmore's engineering traffic from 1910; Fenwick Cross's temporary works platform from 1928. Both are excluded by the wording, and the reveal says so, so a contestant who ordered by first trains can see exactly where their reading diverged.
Both versions of the question are legitimate. An "earliest rail service of any kind" round is perfectly good television. But it's a different question, and the pitch has to pick one and say which.
So every question carries an answer note. At minimum: the intended answer, the boundary the prompt actually states, the alternative readings someone might reasonably offer, and whether those alternatives change the answer or merely need acknowledging. When the answer note is written next to the question, a later editor who shortens the prompt for timing can see what the shortening costs. When the note lives in a different document, the boundary disappears from the question and nobody notices until a contestant appeals.
Specialist or disputed material should go to someone who knows the field, and the question should be rewritten if the boundary turns out not to be stated. That is a revision, not a failure. The failure is discovering it on air.
Name who writes, who checks, who decides
A pitch should be able to say, in plain terms, who proposes material, who researches it, who reviews it, who resolves an objection, and who owns the current version of a question.
Those functions can be one person on a small production or five on a large one. What the pitch cannot leave blank is whether they exist. A format that assumes questions appear has already assigned the work; it has just assigned it to whoever is unlucky.
For a useful point of comparison, the Jeopardy! site published a 2018 account of its research role that treats clue writing and clue research as separate jobs and names the problem of questions with more than one possible correct response. That is one program's historical description of its own production, not a staffing model you're obliged to copy. Take from it the shape of the division, not the headcount.
Version control matters more than it sounds. Any question that has passed review has a current wording, and the wording is the thing that makes the answer defensible. Keep the answer note attached to it. A category file that lists answers without their boundaries is a pile of future arguments.
Walk one question all the way through
Then rehearse it. Take the narrowed Cinder Row question and follow it through delivery: the cards appear, the prompt is read, the clock runs, the team arranges, the reveal shows the sequence and each date.
Watch for the dependencies a written description hides. Can the team read card four while the prompt is still being read? Does the reveal hold long enough to show both the order and the boundary note about Dunmore's engineering trains? Does the host's phrasing of "public passenger service" land as a constraint or as clutter?
What a run-through gives you is clarity evidence. It shows whether the question is legible, whether the reveal explains itself, whether the timing works. What it does not give you is difficulty evidence. A room of production staff and friends is not an audience sample, and thirty seconds of one team's hesitation is not a fairness measurement. If your pitch wants to claim the round is calibrated for a general audience, that claim needs testing you haven't done yet — so leave it out, and let the clarity evidence stand on its own.
Renewal is the last piece, and it's the one that separates a format from a pilot. Where does fresh material come from — staff writers, freelancers, a research contract, contributed submissions? What happens to a question once the episode airs, given that a streamed show publishes its own archive? How does the show review material tied to a moving fact, like the longest platform on an expanding network?
An honest pitch says: questions are dated from the moment they're written, they get reviewed before reuse, and here is the review. It doesn't promise an inexhaustible bank. Nobody has one.
Where to leave it
Close with a single question in full — prompt, intended answer, stated boundary, alternative readings, source and check date, reveal, and the initials of whoever cleared it — and then the sequence that produces the next one.
That pairing does the work. One traceable sample shows the room you can actually make a question. The sequence behind it shows the room there will be more. Keep three things separate while you present it: what has been researched and is correct, what has been rehearsed and is clear, and what has been measured and is difficult. The first two you can demonstrate now. The third is a claim you'll earn later, with an audience, and the pitch is stronger for saying so.
Frequently asked questions
Why isn't a set of five strong questions enough to pitch a quiz format?
Five good questions prove taste, not supply. The format promises something that runs past the pilot: every episode, questions contestants can answer, that the show can defend when someone objects, that don't repeat last week's trick and that don't quietly expire. That promise needs a method, and the method is usually the missing page.
What's the difference between describing a topic and describing a task?
"We ask about history" names a subject area. A task says what the contestant performs — recalling a fact, recognizing a face, working out which of four things happened earlier, joining two facts — plus what they can see and hear, and when. Whether the four options appear on screen while the prompt is read, or only after the clock starts, changes the writing job, so the pitch should say which one it's doing.
How can a question have a supported answer and still be ambiguous?
Support for one reading isn't exclusivity. "Which of these stations opened first?" flips depending on whether "opened" means public passenger service or first use of the site: Cinder Row entered passenger service in 1912, while Dunmore's site carried engineering trains from 1910. Every question should carry an answer note — the intended answer, the boundary the prompt actually states, alternative readings, and whether those readings change the answer or merely need acknowledging.
What does a rehearsal actually prove?
Clarity evidence: whether the question is legible, whether the reveal explains itself, whether the timing works. It doesn't prove difficulty. A room of production staff and friends is not an audience sample, and thirty seconds of one team's hesitation is not a fairness measurement, so a claim that the round is calibrated for a general audience is a claim still waiting on testing.
Who writes, checks and decides on questions, and where does fresh material come from?
Those functions can be one person on a small production or five on a large one, but the pitch has to say they exist — who proposes material, who researches it, who reviews it, who resolves an objection, and who owns the current version. Renewal matters just as much: staff writers, freelancers, a research contract or contributed submissions, plus a rule for what happens to a question after the episode airs. Questions are dated from the moment they're written and get reviewed before reuse.