A badge is not a deduction

Noir adventures, mystery role-playing games, and licensed crime stories are happy to call you a detective. The costume is cheap. What matters is whether you do any reasoning the game did not already complete. One type lets you wear the coat while the plot advances. Another lets you pick a suspect from a list the game assembled. The third withholds the list and waits for you to name the person.

The third is rarer because it can be failed, and a failed accusation is a support problem as well as a design. Players who were never given the means to know will be angry, and they will be right. A deduction game owes you the facts in a form you can recombine, and then it owes you the silence in which to recombine them. A badge and a highlighted clue is the first type wearing the third type hat.

Plenty of games hand you a badge without ever asking you to think. Noir adventures, mystery RPGs, and licensed crime stories all cast the player as a detective, but the costume is not the design. What matters is whether the mechanics put the player through anything resembling real detective work.

Look at what recent detective games actually ask for and the answers separate cleanly. One group hands you scattered evidence and asks who somebody is and what happened to them. Another hands you testimony and asks where it fails against the facts. A third buries the answer in an enormous pile of information and asks where to look next.

Those three demands make three categories, and the rest of this article uses them: deduction games, contradiction games, and investigation games.

A deduction game gives you facts and asks a question whose answer sits between them. Return of the Obra Dinn is the cleanest version. You come aboard a 19th century merchant ship where 60 passengers and crew are dead or missing, and as an insurance inspector your job is a log book: a name and a fate for every one of them.

A magic pocket watch drops you into the frozen instant of each death. Every scene is dense with detail: last words, accents, clothing, objects, the room, and who stands next to whom. Some entries write themselves. Watch a man take a bullet and the cause of death is not in question.

Names are the hard half. The game rarely states one outright, so you assemble it. The captain confesses to shooting Abigail's brother. Abigail's maiden name is Hoscut. A William Hoscut appears on the crew roll. Nobody told you that William was the man who was shot, and yet you know.

The same engine drives several other games. The Case of the Golden Idol has you picking through dioramas frozen at the moment of death, reading notes, rifling pockets, and moving between rooms before filling in a scroll describing what happened. Scene Investigators puts the crime scene in 3D and asks pointed questions about the case. Riley & Rochelle swaps murder for an archive of documents, tapes, letters, and records, and runs the same logic over it.

Strip the fiction away and these are logic puzzles. The information is all there, but no single clue answers the question. You get there by cross-referencing, by elimination, or by inferring one solid fact from two partial ones.

That is where the pleasure lives. The gap between the clues is closed by your reasoning and nothing else.

A contradiction game gives you statements instead of facts and asks you to find where they conflict with what you know. Lucifer Within Us is a good specimen. You investigate murders committed by people manipulated by demons, and each scene comes with witnesses and suspects whose accounts are laid out on a timeline across the bottom of the screen.

None of it can be taken at face value. Witnesses lie, forget, and leave things out. Abraham claims he spent the whole morning on the bushes. Nerissa says she spoke to him first thing. Play Nerissa's statement back at Abraham and the story collapses, he revises it, and the revision opens the case up further.

The pattern shows up all over the genre. Ace Attorney builds its courtroom scenes around cross-examination and the piece of evidence that a witness statement cannot survive. L.A. Noire asks you to rule on whether an answer is honest, sometimes from a face and sometimes from something in the case file.

What is being tested is narrower than deduction. You are matching a claim against a fact and looking for the mismatch, closer to Papers, Please and its stack of documents that do not quite agree with one another.

So the demand on the player is memory more than inference. An alibi has to be held up against the CCTV footage. A statement has to be corrected because some other piece of evidence proves one detail wrong. The fantasy on offer is the interrogation itself: listen closely, keep everything in your head, and pounce on the sentence that breaks.

An investigation game makes the search itself the puzzle by hiding the answer in more information than anyone could read. Shadows of Doubt is the clearest case, dropping you into a simulated city of hundreds of citizens, each with an address, a job, a schedule, fingerprints, and a web of relationships you can look up.

A single fingerprint at a murder scene would be trivial in a smaller game. Find the matching person and you are done. Here that route is useless, because the match is buried among too many people and too many dead ends. What you are really playing is the question of how to move through the mess.

A case opens with a body, an object left behind, and a print with no name on it. From there you might pull hospital surveillance around the time of death and spot somebody who does not belong. Ask his co-workers whether they know the face. Trace sightings to an apartment block, break into its security room, search the resident database, match the face. Lift prints off a door handle, and only then do you have enough to arrest.

None of that is one clever leap. It is procedure: work the leads, shrink the search space, and keep picking a next step that is worth taking. That costs far more content than an authored puzzle, because the true answer only stays hidden if there are red herrings and irrelevant records around it.

Shadows of Doubt buys that content with procedural generation, spinning up enough citizens and detail to lose a killer in. Other games build their haystacks differently. Hypnospace Outlaw uses a sprawling fictional web portal. Sherlock Holmes: Consulting Detective tucks the addresses that matter among many that do not. Her Story puts every fact on video and then withholds order, since the archive only answers search terms. Outer Wilds belongs here too, once you notice that knowledge, place, and timing are the whole progression system.

The skill these games train is reading the structure of the information. You chase a word from one clip into the next. You learn which archive holds which records, or stumble onto a store of files nobody pointed you at. Or you work a schedule and a time loop until you are standing in the right room at the right minute.

Whatever the category, the fantasy usually ends the same way: you have to say who did it. Lucifer Within Us runs on contradictions but still wants means, motive, and opportunity from you. Shadows of Doubt is a city-wide search that closes on a suspect and a folder of evidence. Scene Investigators counts naming the right culprit as one of its deductions.

A few games put the accusation at the center. Paradise Killer scatters clues and suspects across an open island, but the island is the setup and the trial is the payoff. Whenever you decide you have enough, you walk into court and argue each crime on the docket. You get one attempt, and the game will not simply confirm whether you were correct.

Whispers in the West keeps investigation light and puts the pressure at the end, with a clock running and a table of players comparing notes. The result is a social version of the same fantasy: collect what you can, put your theory against everyone else's, and commit.

Evidence gathering, interrogation, chasing leads, deduction, and accusation are all furniture from detective fiction. The design thread underneath them is the same one every time, which is that each asks the player to handle information the way a detective would.

Every detective game needs a way for the player to state a conclusion and be told whether it holds, and that answer tester is the most fragile part of the design. Written carelessly, it solves the puzzle on the player's behalf.

Picture a version of Obra Dinn that floats thought bubbles at you and asks you to link them: the captain shot Abigail's brother, Abigail was born a Hoscut, William's surname is Hoscut. Connecting those resembles thinking, but the game has already made the link you were meant to spot.

Multiple choice does the same damage from another angle. Ask who Abigail's brother is over four names and the question has announced the chain of reasoning. Four options are also few enough to guess, whittle down, and grind through.

Obra Dinn sidesteps all of that. The question is identical for every corpse, which is who they are and how they died. There is no board of pre-drawn clue links. Names come from the full list of 60, so guessing is a waste of a session. And it only confirms in batches of three correct fates, which makes brute force genuinely expensive.

Others find their own version of that discipline. Riley & Rochelle wants full dates, which is thousands of possible answers. Shadows of Doubt lets the size of the city do the work, since nobody stumbles into the right citizen. The tester stays quiet and the conclusion has to be earned.

Contradiction games tend to be the weakest here. Where the evidence list is short, a player can throw every item at every statement until something sticks, and detective work quietly turns into working through permutations.

A tester strict enough to be worth beating will strand players, so good detective games offer help that protects the thinking instead of replacing it. Six moves are available.

The first is asking precisely without asking leadingly. The Case of the Golden Idol improved when it stopped demanding whole explanatory sentences and started posing sharper, smaller prompts.

The second is a ramp. Riley & Rochelle opens on a handful of notes and one-step logic before it expects you to hold several deductions in your head at once.

The third is more than one road to the same truth. Obra Dinn frequently allows two separate ways to pin down a name. Shadows of Doubt will let you reach a killer through footprints, prints, camera footage, call records, or something else entirely. Hypnospace Outlaw sometimes rewards a shortcut and otherwise accepts the slow route.

The fourth is partial confirmation. Golden Idol validates a panel at a time rather than making you submit a case in one piece. The fifth is note-keeping: paper may be enough, but pinboards, tags, bookmarks, and free-text notes turn a mess into something you can reason over. The pinboard in Shadows of Doubt works because any person, place, or object can be pulled onto it.

The sixth is a hint that keeps things moving. Obra Dinn tells you when a fate has become solvable. Shadows of Doubt can simply produce another murder and with it another set of leads. Golden Idol keeps written hints behind an unlock step that is boring on purpose, so help exists but nobody reaches for it by reflex.

The three categories are a lens, not a taxonomy, and several good games sit outside them. Overboard flips the role and makes you the murderer, so the play is falsifying evidence, building an alibi, and steering suspicion elsewhere. Silicon Dreams turns the interview into a reading of emotional state and a lever for manipulation. Among Us counts too, because the culprit is another player and the evidence is behavior, alibis, timing, and lies.

The principle underneath is wider than any single category. Give the player clues. Ask for real thinking, whether that means a deduction, a contradiction, a lead, a connection, or an accusation. Then check the answer in a way that does not telegraph it, does not shrink the field to something guessable, and does not reward grinding.

Whether your own tester does that is not something a design document can settle. You find out by getting the thing running and trying to cheat it yourself, and by watching whether a player who does not know the answer starts guessing. Get it right and the payoff is the good one. The player is not watching a detective be clever. They are the one who saw how the facts fit together.

In the Flockbay app, write three facts and a culprit you must name by typing, not by selecting. Play it without having written the answer in the brief the player sees. If you cannot reach the name from the facts, the facts are private. Add one, do not add a menu.

A detective mystery template and the AI point-and-click maker are the tools. Three facts, one name.