Mystery adventures are particularly difficult for lots of reasons. D&D is a game primarily about combat, and mysteries put a lot of burden on the DM to construct a gripping narrative. Drama often emerges from the dice rolling and mechanics in a standard fight-this-monster scenario, but in a mystery the drama is not as emergent from the game itself. In the ruleset itself, there are limited mechanics for running a mystery, like investigation and deduction.
The Problem With Investigation
One can simulate investigation through a series of rolls, but for a dedicated mystery adventure this falls short. A mystery requires a DM or adventure author to come up with a compelling mystery, such as a crime, and a method of figuring out the solution that is neither obvious nor impossible. And, it has to account for game breaking spells like the dreaded Speak With Dead.
The main thing you’d hope player characters are doing is investigating, and that means looking for clues. The nature of TTRPGs as a format however, make this difficult.
When running an investigation, you have to decide, do you;
Allow the players to lead the investigation, proactively asking to investigate things as the DM narrates the result? This risks the players missing all the clues entirely, simply because they didn’t know to look for them.
Or, does the DM telegraph the presence of clues? This risks revealing the presence of clues too obviously.
Clues must be telegraphed lest they get lost entirely, but telegraphing clues without giving them away is very difficult. And then you have the problem of players reading in things that are not clues (see: Critical Role chair incident).
The Problem With Deduction
The next stage, once a detective has collected evidence or clues, is deduction; interpreting what a clue means.
Imagine how this works in detective fiction. The detective finds a clue, and you read their deduction of this clue and accept it at face value. But is this the player’s experience? The problem with the player's interpretation of clues is the problem of validation. How does a player that makes a deduction validate that they are correct?
Does the DM tell them “that’s correct”?
Or does the DM remain silent while the player wonders if their interpretation is valid?
Let's keep these problems in mind for later, when we'll try to tackle them.
Where Can You Look To Improve Your Mysteries?
In fiction, writing mysteries that can be solved by the reader is more difficult than writing mysteries that the author solves for the reader.
Imagine the latest piece of detective fiction you experienced. Was the crime something that you could solve before the end, and know you were correct? Probably not. The climax of a detective story is not the reader or viewer solving the mystery, it’s the detective solving the mystery. It’s the reveal! The reveal to the audience is supposed to be unexpected, you’re not supposed to figure it out yourself.
But in RPGs the player is the detective, and so they not only need to solve the mystery, but also know that their solution is correct.
So if we can’t really look to detective fiction to help us run a good piece of interactive mystery fiction, then where can we look?
Video games.
The 3 Type of Detective Video Game
Game Maker’s Toolkit (in this video) identifies 3 types of detective video game;
1. Deduction
In deduction games, you align information to come to a conclusion.
They can be as simple as logic puzzles e.g. Jack murdered Anne. Anne murdered George. Jack is married, George is not, and we don’t know if Anne is married. Did a married person murder an unmarried person?
Essentially, it’s where you can gather information, and when you have enough, the conclusion can be inferred. Sometimes, the web of clues you need to piece together is like sudoku where you can fill in the clues yourself once you have enough contextual information.
2. Contradiction
Contradiction games involve players collecting witness testimonies and finding which ones don’t line up, to deduce who is lying.
There are similar logical puzzles to deduction games, but in this case it’s not about what information aligns, but instead, what information does not align.
3. Investigation
The final type of game is an investigation game, where the conclusion is out there in a sea of evidence. It’s not so much that you have to make logical leaps, but instead, create a system of sorting through the sea of evidence to narrow down to a conclusion.
In these games, we might imagine clues show Anne was murdered by someone with red hair who cast a fireball. So a redheaded pyromancer. But in this scenario, instead of just George and Jack as suspects, the DM has a list of a million residents of Waterdeep. So how are you going to approach it to narrow it down? Go the pyromancy college and find a list of alumni, ask about people with red hair etc.
In these games, the method is the puzzle.
Let’s imagine these in a TTRPG. Firstly, Investigation games, with their reliance on an enormous sea of evidence would be difficult to write and run. There’s surely a way of doing it, but it would be difficult!
Deduction and Contradiction games however, are a lot easier. They don’t rely on quantity of evidence, instead they require a limited set of information that poses a puzzle. Contradiction puzzles are social in nature, relying on witness testimonies, and so they lend themselves to roleplay. And the great advantage for TTRPGs that Deduction and Contradiction mysteries have, is that they can be self-validating.
When you play sudoku you don’t need the creator of the puzzle to tell you that you got it right. The correct answer tells you it is correct.
If we use similar logic puzzles, we can help solve the problem of deduction. Self-validating clues where the player's interpretations are inherently correct or incorrect.
Next we need to solve the problem of…
Telegraphing Clues
These mysteries hinge on the players discovering clues, so we need to make sure that they can be discovered. Sure, you can roll Investigation to skip that part, but that doesn’t feel especially satisfying, and leaves room for failed rolls that stop the session progressing. So we need some robust methods of encouraging players to search for clues in the places that clues can actually be found.
Suspect Testimonies Don’t Need Telegraphing
One clue can be the witness and suspect testimonies themselves. Generally, if you’re running a closed mystery where there’s a group of suspects, you won’t need to telegraph to players that they should talk to the suspects themselves. We can rely on the player's understanding of the tropes of detective fiction to interrogate suspects.
Prescribed Set of Clues
One thing that you can do, is prescribe the players an existing template that clues can fit into. For example, in Cluedo you know that the clues are; Weapon, Murderer and Place. In an open, sandbox game, it’s not as simple, but you can still create a set of clue categories that give the player’s direction.
Some kind of UI can do this, where a set of potential clues are laid out. By giving the players a template to fill in, it prescribes the types of clues that they are looking for, without removing the open-ended problem solving that makes TTRPGs special.
Other trackers, templates and methods of presenting information can help players give structure to their investigation. Give them a map of the location, and they’ll start searching and making notes in each room. Give them a pinboard with some string and portraits of NPCs and they’ll start piecing together the relationships between the different people, without you ever having to tell them that’s how they should approach the mystery.
Some kind of interface can encourage the players to look for clues in the right places and investigate with a certain method in a way that feels organic and immersive, even if it's totally prescribed and artificial.
Missing Clues and Story Progression
Unfortunately, no matter how many trackers and handouts you provide, you have to accept that players will miss clues, and plan accordingly. These clues, if they are essential to the progression of the adventure, become chokepoints. That’s why The Alexandrian proposes the Three Clue Rule. You can read more about it on his blog, but essentially it’s this: provide 3 clues for any conclusion otherwise the players will miss them.
In general, this applies to all puzzles in D&D - you should make sure any problem has at least 3 solutions. Often though, the best puzzles are not the ones with prescribed solutions. This is the same for clues.
Sandbox puzzles like crossing a ravine are the best, because the DM doesn’t need to plan a solution. It’s open-ended enough that the players will come up with something. Clues are the same.
For example, if a clue is the secret known by an NPC, there are many social (and anti-social) approaches to finding it out. It’s essentially a puzzle without a prescribed solution.
Anywhere that the players have to overcome a challenge to progress is a chokepoint. However, in mystery adventures, chokepoints are overcome with logic and deduction, rather than combat. This is the main reason they are so difficult to run. Where you can, make sure it is signposted to players that they are progressing; small victories are important in mysteries, because we are building to a single, final reveal.
Plans Do Not Survive Contact With The Players
Always account for the fact that the players can break the game. Congratulate them if they figure out the mystery in 30 seconds, and be ready to make the final encounter or aftermath a bigger part of the story. But remember, you can’t predict everything.
There are some things we can predict thought. Duncan from Hipsters & Dragons has a great comprehensive article that goes through how to deal with the different mystery-game-breaking spells.
Here though, we'll at least tackle the worst offender; Speak With Dead.
Speak With Dead doesn’t need to be a problem, it can actually be an engaging part of the mystery. In response to Speak With Dead in a murder mystery, you can;
1. Make the conditions of the murder so that dead body can’t reveal anything crucial; they didn’t see the murderer, or didn’t recognise them.
2. Even better, they did see the murderer and that’s a clue! Spoilers for our mystery The Hrossroad Inn: the murderer is a shapeshifter, so the appearance of the murderer is an intentional hint. The player in this case, are meant to find out what the dead body’s got to say.
3. You could just nullify the spell; no head on the body, anti-magic etc. This is a little boring, but it could hint that the murderer was prepared for this spell and planned accordingly.
There are many various directions you won’t expect that the players might take the story. A closed circle mystery where the scenario is limited to a smaller location and set of suspects is a great way to reduce the risk, as well as create detailed scenario with depth.
That’s all for now! If you want to see some of these principles in action you can download our murder mystery: The Hrossroad Inn from our Discord!
A group of travelers arrive at the inn. A murder. Acid rain. And a dark secret.