What We Have Learned From Building a Murder Mystery

Building a murder mystery is not the same as writing one. A novelist knows exactly which clue the reader sees and when. We hand that control to the player and hope the case still holds together.

A good case is a chain, not a pile

Our earliest cases threw evidence at the player and hoped something would stick. It did not work. A pile of clues is not a mystery, it is a filing task. What makes a case satisfying is a chain: the mortician names the weapon, the weapon points to a room, the room contradicts an alibi, and the alibi collapses under questioning.

If any link in that chain is missing, the player is guessing. If every link is present, they feel clever. The difference is not difficulty. It is whether the reasoning is available to be found.

The lie has to be catchable

Every suspect in BACKSTABBERS can be evasive, but only one of them is guilty, and the guilty one has to be catchable. A liar who never slips is not a puzzle, it is a wall. That is why you can now put a piece of evidence in front of a suspect and watch what happens: if it genuinely contradicts what they told you, they crack.

Nothing about that moment is scripted. The suspect is reasoning about the evidence you chose, against the account they already gave you.

Where the experience breaks down

Some lessons were less flattering. Clues that link to nothing teach players to distrust the whole evidence board. A timer that runs while you read the instructions feels like a cheat. A hint that points at the wrong suspect is worse than no hint at all. We have fixed all three.

What comes next

The Daily Case is where most of this thinking ends up: one mystery, twelve minutes, the same case for everyone, no account required. It is the shortest honest version of the thing we are trying to build.

Every failed case teaches us something. Usually about ourselves.