A good mystery can be one of the most satisfying adventures to run in a tabletop RPG. There's something wonderful about watching players argue over evidence, compare their notes, question witnesses, and finally make a connection you've been waiting two sessions for them to notice.
There's also something deeply uncomfortable about realizing that they've spent forty minutes discussing a clue you mentioned purely as background decoration.
The Game Master has an unusual advantage during an investigation: they already know the truth. They know who committed the crime, what the culprit wanted, which witnesses are reliable, and where the evidence can be found. From that position, the solution often seems perfectly obvious.
For the players, however, the mystery consists of incomplete information delivered through descriptions, conversations, dice rolls, and whatever details they happened to write down.
That difference creates one of the central challenges of running an investigation. You need to provide enough information for the party to reason about what happened, while resisting the temptation to guide them toward the answer whenever their theories begin heading in the wrong direction.
The goal isn't to make the mystery impossible to solve. It's to make solving it feel like something the players actually accomplished.
Start with what happened, not what the players are supposed to discover
Let's build a small investigation.
In the riverside town of Blackwater, a merchant named Elias Voss has been found dead inside his locked counting house. His business partner claims he was working late. The town watch believes the murder was committed by a thief who entered through a window overlooking the canal.
The truth is less straightforward.
Elias discovered that his business partner, Mara Bell, had been falsifying shipping records to conceal an illegal trade. During an argument, Mara struck him with a heavy brass seal. She then staged the room to suggest a burglary and arranged for a warehouse clerk to remove the incriminating ledgers.
That's what actually happened.
Notice that we haven't decided how the players will solve the case. We haven't written a sequence requiring them to speak with the watch captain, inspect the counting house, visit the warehouse, and finally confront Mara.
Instead, we know the events that occurred before the characters arrived.
That gives us something far more useful than a predetermined investigation route: a consistent version of reality.
If the party questions Mara, we know what she's hiding. If they inspect the warehouse, we know what happened there. If they interview a boatman who works near the counting house, we can decide what he might reasonably have seen.
The investigation becomes a process of discovering an existing truth rather than following a route through prepared scenes.
Evidence should exist independently of the party's decisions
When planning a mystery, I find it helpful to separate evidence from the order in which players might encounter it.
In Blackwater, the murder leaves several traces. The brass seal is missing from Elias's desk, a small amount of blood remains beneath the cabinet where Mara tried to clean it, and the warehouse clerk has altered a shipping record. A canal worker also remembers seeing Mara leave the building later than she claims.
Each detail exists because of something that happened during the crime or its cover-up.
The players may discover these clues in almost any order. They might inspect the counting house immediately or spend their first evening talking to the people who worked with Elias. They may suspect Mara early, or they may pursue the burglary theory suggested by the watch.
That's fine. Their approach changes how the investigation unfolds, but it doesn't change what happened.
The important thing is that the evidence remains consistent.
If the players ignore the counting house and investigate the warehouse, the altered records are still there unless someone has had a believable opportunity to remove them. If they question Mara before finding the ledgers, her responses should reflect what she knows and what she believes the party has discovered.
Once you've established the truth, improvising around unexpected player decisions becomes much easier.
A false lead should have an explanation
Red herrings are common in mystery adventures, but they're also easy to mishandle.
Suppose the players discover that Elias had recently argued with a local moneylender. The moneylender had threatened to ruin him financially, and someone saw the two men fighting in public several days before the murder.
This is suspicious, and the party has every reason to investigate it.
However, the argument doesn't need to be a deliberate trick placed by the GM. It can simply be another conflict in Elias's life.
The moneylender may have wanted repayment, while Mara wanted to prevent Elias from exposing the illegal shipping operation. Both had reasons to be angry with him, but only one committed the murder.
A good false lead is supported by believable circumstances. It becomes problematic when it's designed exclusively to waste the players' time or when the GM deliberately withholds information that would allow them to evaluate it.
If the party spends an hour investigating the moneylender, that investigation should produce something useful, even if it doesn't identify the murderer.
Perhaps the moneylender has a genuine alibi. Maybe the players discover that Elias was unexpectedly wealthy despite his financial troubles, which raises questions about the business records. They might also learn something about his relationships that changes their understanding of the victim.
The investigation hasn't failed simply because the party pursued the wrong suspect. They have eliminated one possibility and learned more about the case.
That's detective work.
Don't change the murderer to match the players' favorite theory
This is a tempting improvisational technique, especially when the players develop a particularly clever explanation.
Imagine that the party becomes convinced the warehouse clerk committed the murder. They've noticed that he altered the shipping records, had access to the counting house, and behaved nervously during questioning.
Their theory is coherent. In fact, it's almost better than the explanation you prepared.
You could decide that they're right and change the culprit behind the scenes.
There are styles of collaborative storytelling where that approach works beautifully. But if you're running a mystery intended to be solved through evidence, changing the truth can undermine the entire exercise.
The clues no longer necessarily point toward an established solution. Instead, the solution changes according to whichever theory seems most entertaining.
Players may never notice, but the investigation loses something important: the possibility of being wrong.
Being wrong is part of what makes solving a mystery satisfying.
The warehouse clerk can be guilty of helping conceal evidence without being responsible for the murder. The players may correctly identify his involvement while misunderstanding his role, and further investigation can reveal the distinction.
That creates a more interesting outcome than simply confirming their first convincing theory.
Let the suspects behave like people
One of the easiest ways to make an investigation feel artificial is to treat suspects as containers of information.
The players ask the right question, the NPC gives the relevant clue, and the investigation advances. Ask the wrong question, and the NPC has nothing useful to say.
Real conversations don't work that way, and RPG investigations become much more engaging when witnesses and suspects have their own concerns.
Mara Bell doesn't merely exist to conceal the identity of the murderer. She's trying to protect her business, maintain her reputation, and avoid being arrested. She may be frightened, defensive, or unusually cooperative if she believes she can direct suspicion elsewhere.
The warehouse clerk has a different problem. He helped alter the records because Mara threatened to expose his own petty theft from the company. He's afraid of her, but he's also afraid of the town watch discovering what he's done.
Those motivations affect how each character responds to questioning.
If the party approaches the clerk aggressively, he may refuse to cooperate. If they offer protection, he might reveal what he knows. If they accuse him of the murder, he could panic and attempt to flee.
You don't need to prepare every possible line of dialogue. You need to understand what each person wants, what they know, and what they're willing to risk.
The players can handle the rest.
When players misunderstand a clue, resist the urge to correct them
Let's say the party discovers a scrap of blue cloth caught on a nail near the counting house window.
You intended it as evidence that someone staged the burglary. The cloth came from a warehouse covering used to protect cargo from the rain.
The players, however, remember that the town watch wears blue uniforms. They become convinced that a guard was involved.
Your first instinct may be to clarify that the cloth is a different shade of blue or provide additional details that steer them toward the warehouse.
Sometimes clarification is appropriate, particularly if the misunderstanding comes from an ambiguous description. If the characters would obviously recognize the difference between uniform fabric and coarse cargo cloth, explain that.
But if the players are drawing a reasonable conclusion from incomplete information, let them investigate it.
They might question the watch, inspect uniforms, or compare the cloth with fabric found elsewhere. The resulting discoveries can help them eliminate their theory or uncover something genuinely unexpected.
The important distinction is between correcting a communication problem and correcting a player's deduction.
The first is part of being a fair GM. The second can quietly turn an investigation into a guided tour of your preferred solution.
Use a clue board to understand the investigation
As a mystery becomes more complicated, keeping track of relationships can become difficult.
In Blackwater, we have a victim, several suspects, a counting house, a warehouse, altered ledgers, a missing brass seal, and witnesses whose statements contradict one another.
You can document all of this in a conventional scenario, but a visual representation can make the connections much easier to understand.
A clue board allows you to arrange the important people, places, and pieces of evidence and see how they relate to one another. You might connect Mara to the shipping records, the warehouse clerk to the altered ledger, and the counting house to the missing seal.
That approach is useful during preparation, particularly when you're checking whether the investigation contains enough information to support its conclusions.
It also helps during the session. If the party unexpectedly questions the canal worker before visiting the warehouse, you can quickly see what that character might reasonably know and which parts of the case their testimony could affect.
Talehaven's Board provides a visual workspace for organizing campaign elements and their relationships. Combined with the Adventure module, you can keep investigation scenes, NPCs, locations, documents, and other materials available as the party explores the case.
The application doesn't solve the mystery for you. It simply gives you another way to understand and manage the information you've prepared.
For an investigation involving several suspects and competing explanations, that can be particularly useful.
Interactive demo
The Blackwater Case
A merchant has been murdered. Examine the evidence, connect the clues, and decide who did it.
Elias Voss was found dead inside his locked counting house. The town watch believes a thief came in through the canal window.
Evidence
- The Altered LedgerTwo pages of the shipping ledger have been rewritten in a careful hand. The entries cover the weeks Elias had been questioning, and the corrections all run the same way: cargo that never officially arrived.
- The Missing Brass SealThe desk set is complete except for the seal, which was heavy enough to be used as a weapon. Nothing else of value was taken from the room, which sits badly with the idea of a thief.
- The Canal WitnessA canal worker loading a barge remembers the counting house door opening well after dark. He recognised Mara Bell. Her own account has her leaving before sunset.
- The Warehouse Clerk’s StatementDaniel Cross admits he moved the ledgers out of the warehouse office that night. He insists he never went near the counting house, and he is noticeably more frightened of Mara than of the watch.
- The Moneylender’s ArgumentHalf the street saw Victor Hale threaten to ruin Elias over an unpaid debt. It was loud, it was public, and it was three days before the murder. Several people can say where Victor was that night.
- The Blue ClothCoarse blue cloth, the kind used to cover cargo against rain, snagged on a nail by the window. It suggests someone came in from the canal - or wanted it to look that way. The town watch also wears blue.
Suspects
- Mara BellBusiness partner. Says she left before sunset. Handles the shipping records.
- Daniel CrossWarehouse clerk. Moved the ledgers. Afraid of somebody, and it is not the watch.
- Victor HaleMoneylender. Threatened Elias in public. Was elsewhere that night.
Reveal the solution
The records, the timeline and the missing seal all point the same way.
Elias found the falsified shipping records, and Mara is the one who kept them. The canal witness puts her at the counting house after dark, which contradicts her own account. The only thing taken from the room was the heavy object that could have killed him. Daniel moved the ledgers for her, and the blue cloth staged a burglary that never happened.
A fair mystery doesn’t require players to follow your preferred route. It gives them enough consistent information to reach the truth themselves.
What if the players accuse the wrong person?
Eventually, the party may decide it has enough evidence to act.
Unfortunately, the person they accuse might be innocent.
This isn't necessarily a failure of the adventure. It's an outcome of the players making a decision based on their understanding of the evidence.
If they publicly accuse the moneylender without convincing proof, he may demand an apology, threaten legal action, or use his influence to make their investigation more difficult.
If they persuade the town watch to arrest the warehouse clerk, Mara gains time to destroy evidence or leave Blackwater.
The consequences should follow from what the characters have done, rather than existing simply to punish them for choosing incorrectly.
Players should also have reasonable opportunities to reconsider their conclusions. A new witness might come forward, a suspect's alibi could be verified, or the behavior of the actual culprit may create additional evidence.
A good mystery doesn't require the party to be right immediately. It allows them to revise their understanding as they learn more.
The best revelation is one the players reach themselves
There's a special kind of satisfaction in watching a group connect several pieces of information without assistance.
Someone remembers that Mara claimed to have left the counting house before sunset. Another player points out that the canal worker saw her there after dark. A third realizes that the altered shipping records explain why she needed Elias's ledger removed.
The party begins comparing its evidence, and you can see the moment when the explanation finally makes sense.
It's tempting to confirm the theory immediately, particularly when you've been waiting for them to reach the correct conclusion.
But you don't need to.
Let the players decide what to do with their discovery. They may confront Mara, search for the missing brass seal, persuade the warehouse clerk to testify, or attempt to trap her into contradicting her earlier statement.
Their theory has given them a direction, but they still have to act on it.
That's what makes the investigation feel like part of the game rather than a puzzle whose solution automatically ends the adventure.
A practical checklist for your next RPG mystery
Before running an investigation, check a few things.
- Do you know exactly what happened? Establish the truth before deciding how the players might uncover it. A consistent sequence of events makes improvisation easier.
- Can the important conclusions be reached through more than one source? If the entire investigation depends on finding a single note in a hidden drawer, you're creating a fragile adventure. Give players several reasonable opportunities to discover essential information.
- Do your suspects have believable motivations? They should behave according to their own interests, not simply dispense clues when questioned.
- Can the players reasonably form incorrect theories? A mystery where every clue points obviously toward the same person may not offer much investigation. Competing explanations are interesting when the evidence allows players to evaluate them.
- Does the case continue when the players make mistakes? Consider what happens if the wrong suspect is accused, an important witness is ignored, or the culprit realizes someone is investigating.
- Will the players have to do something with their conclusion? Discovering the truth is satisfying, but deciding how to prove it, whom to confront, and what consequences to accept can make the final part of the adventure even better.
You don't need an enormous conspiracy or twenty suspects to create a satisfying investigation. A small case with believable characters, consistent evidence, and meaningful decisions can occupy an entire session.
The mystery belongs to the players once the session begins
Preparing an RPG investigation requires a different mindset from writing a detective story.
In a novel, the author controls when the detective finds a clue, which questions are asked, and when the final revelation occurs. At a tabletop RPG table, the players make those decisions, sometimes in an order you never anticipated.
Your job is to create a situation that can withstand that freedom.
If the evidence is consistent, the suspects have understandable motivations, and the consequences follow logically from what happens, you don't need to steer the party through every step.
They may solve the case faster than expected. They may spend an entire evening investigating an innocent person. They might even reach the correct conclusion for completely unexpected reasons.
That's part of the appeal.
When the players finally put the pieces together, the satisfaction comes from knowing they did the work themselves.
And as the Game Master, you get to sit there pretending you weren't worried about the whole thing falling apart twenty minutes earlier.
