How to run a mystery without hiding the solution too well
The Talehaven team9 min read
You built a beautiful mystery. The players walked in, missed the one clue that mattered, and spent forty minutes searching a room you had already emptied of anything useful.
The problem is almost never that the players are slow. It is that the whole case was resting on a single thread, and the thread snapped.
Short answer
Use the three-clue rule: for every conclusion the players need to reach, plant at least three independent clues pointing at it. Then a missed roll, an ignored NPC or an unsearched room costs you nothing, because two other paths still lead to the truth.
An investigation that does not stall
Design backwards from the solution
Decide first what actually happened, who is behind it and where the proof sits. You build the puzzle in reverse: only once you know the truth can you work out what traces it would leave. Skip this and your clues point nowhere, because you do not know where either.
The three-clue rule
For every conclusion the players must reach, prepare at least three independent clues. Not because players are dim, but because they will miss one, misread the second, and only click with the third. A case hanging on a single clue will collapse sooner or later.
Give each clue more than one way in
The same fact should be findable several ways: in conversation, in a document, in examining the scene. The group that never talks to NPCs still trips over it in the ledger. A clue with one entrance is a clue half your tables will never see.
False leads that still teach something
A good false lead is not a dead end - it is a wrong answer that carries real information. The players suspect the wrong person, but chasing them turns up something true about the case. A lead that just burns an hour and ends in nothing is a punishment, not a puzzle.
Never gate a clue behind one roll
If finding the key piece of evidence rides on a single Perception check, you have designed a session one bad die can end. A roll can decide how fast, or at what cost, the players find something. It should never decide whether they find it at all.
When the party stalls, move the world
An investigation is not only the players hunting the culprit. If they get stuck, let a clue come to them: another victim, a witness who cracks, a note from an NPC. That is not cheating - it is the world reacting while the players hesitate.
The disappearances at the Alpendorf mine
Four miners have vanished in two months. The official line is that they left for the city. The truth: the night foreman, Greta, has been sending them into sealed shaft four to dig silver off the books, and whatever is down there does not let them climb back out.
Clue one: the seal on shaft four is broken and the pick-marks are fresh.
Clue two: a widow mentions the extra night shifts her husband was paid for in silver.
Clue three: the pay ledger records bonuses on nights the mine was officially shut.
The false lead: the village blames the hermit on the ridge. He is innocent, but he has watched lanterns climb to shaft four after midnight. Wrong suspect, right direction.
Common challenges
- One clue, one roll, one entrance - and the whole case stalls the moment anything goes wrong.
- A puzzle so clever that only you know the answer; the players flounder and you cannot help without spoiling the "surprise".
- False leads that go nowhere and do nothing but eat the session.
- Punishing players for not asking the exact question you had in your head.
Keep the clues in one place, the secrets in another
Lay the case out, hide the solution in your GM notes, and share with the players only what they have already uncovered.