How to Build a Jeopardy Game for Class: Categories, Clues and Scoring
If you want to know how to make a jeopardy game for class, the useful answer is not a template. A board is four moving parts: categories, values, clue format, and scoring. Each one has a failure mode that shows up in front of thirty students, and each one is cheap to fix before the first bell. This guide covers the standard 100–500 ladder, when a second round doubles it, how to write clues that carry exactly one fact, and the hosting rules you must announce before the first clue so nobody argues about points in the middle of a review lesson. A generated board is a first draft. The teacher still owns the facts.
The four parts of a working board
A board is a grid, and every cell in that grid is a promise that something answerable is behind it. Before you write a single clue, decide these four things.
Categories. Column headers. A traditional game uses six. Six columns of five gives you thirty cells, which is roughly one review block. Fewer than four categories and the board feels thin; more than six and the text shrinks below readable size on a projector at the back of the room.
Values. The point ladder down each column. The standard is 100, 200, 300, 400, 500. This is not decoration: the ladder is a difficulty signal. Students learn within two clues that the bottom row is where the hard material lives, and they bid their confidence accordingly.
Clue format. How the prompt is worded and how the answer is revealed. Pick one convention and hold it for the whole game. Mixing formats mid-board is the fastest way to lose a class.
Scoring. Who gets points, when, and what happens on a wrong answer. This is the part teachers improvise, and it is the part that causes the argument in round two.
Get all four settled in a planning doc before you open any tool. The build takes twenty minutes once the content exists.
Designing categories that support five clues each
The most common structural mistake is a category that is really a single fact wearing five costumes. "The French Revolution" sounds like a category. In practice you can write two good clues for it and then you start padding.
A category is usable if it supports a genuine difficulty gradient. That means five distinct facts of rising obscurity, each independently answerable.
The gradient test
Write your five answers first, as a plain list, without clues. Then read them top to bottom and ask whether a strong student gets the 100 instantly, gets the 500 only with effort, and a weaker student gets the 100 and misses three of the rest. If the list is flat — all five equally hard, or all five equally easy — the category is broken, not the clues.
Strong category shapes:
Category shape | Why it works | Example framing |
|---|---|---|
Named set with internal ordering | Five members exist, ordered by fame | Major rivers by length, elements by atomic number |
Process with stages | Each stage is a separate fact | Steps of mitosis, stages of a bill becoming law |
Vocabulary cluster | Terms defined in the same unit | Literary devices, lab equipment |
People in one field | Five names at different fame levels | Reformers, composers, mathematicians |
Cause and effect chain | Each link is testable separately | Events leading to a specific outcome |
Weak category shapes: "Chapter 7", "Miscellaneous", "Potpourri", "Things we covered". These give you no gradient because they give you no organizing principle.
Plan six categories and write five answers for each, then cut the weakest category entirely. You will end up with five strong columns rather than six uneven ones, and a five-column board reads fine on a projector.
The 100–500 ladder, and when a second round doubles it
The single-round ladder is 100 through 500, five tiers. The doubling rule is simple: a second round (often called Double Jeopardy in the format most people recognize) uses the same five-tier structure but with values doubled — 200, 400, 600, 800, 1000 — and usually introduces fresh categories rather than reusing round one.
You should run a second round when one of these is true:
- You have more than thirty facts to review and the class period allows it.
- The material splits naturally into two halves (two units, two chapters, two skill sets).
- You want a distinct late-game swing where a trailing team can close a gap.
You should not run a second round when the period is under forty minutes, or when round one already ran long because of disputes. Doubling the values does not double the pace; it usually slows it, because students take the higher-value clues more seriously and argue about them longer.
Scoring arithmetic to keep in mind: a full single-round sweep is 100+200+300+400+500 per column, which is 1500 per column, 9000 across six columns. A second round at doubled values is 18,000. If you are tracking scores on a whiteboard, write the maximum at the top so students can see how much board is left, not just how much they have.
Put the point values in the cell corners, not in the clue text. When the value is embedded in the sentence, students read the number aloud while you are still reading the clue, and you lose the room.
Writing a clue so exactly one fact comes through
A clue has one job: make exactly one fact retrievable. Everything else in the sentence is scaffolding.
Three rules that hold up under pressure.
One interrogative per clue. If your clue contains "and", check whether it is joining two questions. "Who wrote this, and in what year?" is two facts. Split it.
No accidental hints. Students are excellent at pattern-matching surface features. If every answer in a column starts with the same letter, they will notice by the third clue. If your clue for a 300 uses the same phrasing as your clue for the 100, they will notice that too.
Answer in a consistent grammatical form. If your answers are names, all five should be names. Mixed answer types within a column make the reveal moment feel arbitrary.
The 500-first method
Write the 500 first. Then the 400. Then down to the 100.
The reason is calibration. The hardest clue is the one you actually have to construct — it needs a specific, discriminating detail that a prepared student knows and an unprepared student does not. Once that ceiling is set, each lower tier is a matter of removing one constraint. The 100 clue is the fact stated nearly plainly. The 300 adds a layer of indirection. The 500 requires synthesis.
If you write bottom-up (100 first), you naturally anchor on the easy version and then struggle to make the top of the column meaningfully harder without it becoming unfair or obscure. You end up with a column that is really four 100s and one 300.
Read every clue aloud once at speaking pace. If you stumble, the class will stumble. Clues that read fine silently often have three subordinate clauses that nobody can hold in working memory while also recalling a fact.
Hosting: who clicks, how teams answer, what happens on a wrong answer
The board is the easy part. The room is the hard part.
Who clicks
Default: you, the teacher, at the keyboard, with the board on the projector. This keeps pacing under your control and prevents the cursor from becoming a toy. If you have a reliable student helper, they can click, but give them one instruction: do not open a cell until you say the category and value out loud.
If students click, you will lose between ten and forty seconds per clue to navigation mistakes. Across thirty clues that is a meaningful fraction of a period.
How teams answer
Buzzer systems are nice but optional. The reliable low-tech method: each team has a designated responder, and answers are given verbally after you call on the team that signaled first. Hand raised, or a physical object tapped on the desk, both work.
The important structural choice is whether all teams can attempt every clue or only the first to signal. First-to-signal is faster but shuts out quieter students. All-teams-write is slower but gives every student retrieval practice. For a review game, all-teams-write on the 400 and 500 rows only is a reasonable hybrid: fast at the bottom of the board, deliberate at the top.
Wrong answer: deduct or pass?
Both are defensible. Pick one and state it before the first clue.
Rule | Effect | Best for |
|---|---|---|
Wrong answer deducts the value | Discourages guessing, slows the game | Competitive classes, older students |
Wrong answer passes to next team, no deduction | Keeps energy up, more attempts per clue | Mixed-ability classes, review focus |
Wrong answer costs nothing, no pass | Fastest, least punitive | Short periods, first exposure to material |
The pass rule has one failure mode: on a 500, you can cycle through every team and burn three minutes on one clue. Cap the passes at two.
The one rule to announce before the first clue
State, out loud, before the board opens: the answer must be given in the form the clue asks for, and the decision on whether it counts is made once, immediately, and is not revisited.
That single sentence prevents the mid-game scoring argument. The dispute is almost never about whether a student knew the fact. It is about whether "Ulysses" counts when the answer is "Odysseus", or whether a partial answer earns partial credit. Decide in the moment, move on, and if a team feels wronged, tell them you will review the ruling after the game — then actually review it, after.
Announce the scoring rule and the answer format in the same breath. Most scoring disputes are actually format disputes wearing a costume.
The teacher still owns the facts
A generated board is a first draft. This is the part that matters most and gets skipped most often.
Whatever tool produces your skeleton — categories, values, and placeholder clues — it is producing text that looks like a finished board. It is not. Language models produce plausible clues. Plausible is the dangerous word. A clue that reads confidently and is subtly wrong will teach thirty students the wrong fact, and they will remember it, because it arrived in a game.
Fact-checking protocol that fits in a planning period:
- Every clue gets its answer verified against your actual course material or a primary reference. Not against another generated sentence.
- Every answer gets checked for the alternate-spelling problem. Decide now whether you accept both forms.
- Every 400 and 500 gets a second read for ambiguity. Hard clues are where two answers are both technically correct.
- Delete any clue you cannot verify in under a minute. Replace it with something you know cold.
Budget your time accordingly. Writing thirty clues from scratch takes a long while. Verifying thirty generated clues against a textbook takes considerably less, but it is not zero, and it is the part that determines whether the game actually taught anything.
The fast path
If you want to make a jeopardy game for class without spending an evening on layout and score tracking, the sequence that works is:
- Decide your categories and write the five answers per category as a plain list. This is the intellectual work and nothing automates it.
- Generate the board skeleton — grid, values, clue cells, reveal behavior, scoring. Tools like the 3D game builder in Neta Studio will produce a playable board, and if you want a review game that is also a 3D space students can walk through, Agent Sprite Forge and Mini World let you dress the room without hand-building geometry.
- Verify every fact. This is where your remaining time goes.
- Rehearse the hosting rules out loud once, alone, before class.
The order matters. Generating first and thinking about categories second produces a board that looks complete and plays badly. Writing the answers first means the generation step is mechanical and the verification step is targeted.
The craft is in the category gradient, the single-fact clue, and the 500-first writing order. A generator handles the grid. It does not handle the pedagogy, and it does not know your students or your unit. Generate the skeleton, spend your time on fact-checking rather than typing, and announce the scoring rule before the first clue. Everything after that is just running the room.
Keep going
Build the world it walks in
A character, a tile or a model is a material with nowhere to stand. Every world on the 3D game builder was made from one written description and published as a page you can play — including Mini World and The Cyclops' Island.