How To Host

A murder mystery at home game lives or dies on one question, and you can ask it an hour in: does the room still disagree about who did it? If everyone has quietly settled on the same person and is just waiting for the reveal, the game is already over — you're just doing the admin.

Hosting a mystery night at home has never been easier to start and never been easier to get wrong. You can download something in ninety seconds. You can also spend an evening watching six adults read stage directions to each other with increasing embarrassment. The difference between those two evenings has almost nothing to do with price and almost everything to do with structure — which is the useful news, because structure is something you can check before you commit.

This is the practical version: what the phrase actually covers now, which format suits your group, what you genuinely need on the night, and the specific ways these evenings fall apart.

"Murder mystery at home game" covers three very different evenings

The term has stretched to cover formats that share a genre and nothing else. Buying the wrong one for your group is the single most common way the night goes flat, so it's worth being precise.

Format	What players actually do	Suits
Character role-play (the classic murder mystery party game)	Each guest is assigned a character with secrets, and reads or improvises through timed rounds. One person is secretly the culprit — sometimes even they don't know at the start.	Confident, extroverted groups of 6–12 who want a dinner-party performance more than a puzzle.
Case file / evidence pack	Everyone gets the same body of evidence — statements, records, photographs, financials — and works out the answer by reading and cross-referencing.	Groups of 1–6 who want to actually deduce something. Works solo. No acting required.
Hybrid (evidence plus digital gates)	A printed evidence pack whose clues unlock real login screens in a browser — an inbox, a bank portal, an archive — holding the next layer of the case.	People who want the physical satisfaction of paper and the reward loop of actually getting into something. This is what we build.

The mismatch to avoid: buying a role-play kit for a group of four introverts, or an evidence pack for ten people who came to be entertained. Both are good products in the wrong room.

The one-hour test

Here is the test that separates a mystery from a costume party with a plot attached. An hour in, with roughly half the material read, can more than one person in the room make a serious case for a different suspect?

If yes, the game is working. Everyone is holding partial information, everyone has a theory, and the remaining evidence has real work to do. If no — if the group converged on one name early and the rest of the evening is confirmation — then the game handed the answer over and dressed the rest up as investigation.

This is not a subtle failure. It is the most common one. It happens because writing three genuinely convincing suspects is several times harder than writing one culprit and two names, and it is invisible to the person who wrote the case, who already knows the answer and cannot un-know it. We wrote about that specific problem separately, because it is the thing our own design rules exist to prevent.

The practical version of the rule we build to: could someone read any single document in the pack and confidently name the culprit? If yes, the case fails, no matter how good the writing is.

What you actually need on the night

Less than most guides suggest. A murder mystery at home game is one of the cheapest evenings you can host, because the production budget went into the writing rather than into components.

• A printer, or a print shop. Home laser or inkjet on ordinary A4 is fine for almost everything. See our guide to printing a mystery at home for what actually matters and what doesn't.
• A table big enough to spread out. This matters more than people expect. Evidence games are spatial — people sort documents into piles, and piles need room. A coffee table for four is tight; a dining table is right.
• One phone, tablet or laptop if the case has digital elements. One between the group is enough, and passing it around is part of the fun.
• Two to three hours, uninterrupted. Not a hard rule, but mysteries do not survive being paused for forty minutes while food is served. Eat first, or eat while you read.
• Somebody willing to write things down. Not a host, not a game master — just a person with a pen who tracks who claimed to be where. Half the pleasure of these evenings is watching that page fill up.

What you do not need: costumes, a script, an app, an account, a subscription, or a designated narrator. Some formats use them. None of them are required for a good evening.

The four ways these evenings actually fail

1. The reveal card

In some role-play kits, one guest opens an envelope at the start telling them they are the murderer. That player now has nothing to discover — they spend the evening managing everyone else's discovery instead of having an evening. If you're using a role-play format, look for one where the culprit is determined at the end, or where even the guilty player has something left to find out.

2. The one obvious suspect

Covered above, and worth repeating because it is so common: if one character's motive is loud and everyone else's is decorative, your guests will spot it in the first twenty minutes and be polite about it for two hours.

3. The unfair clue

The opposite failure. A solution that depends on noticing a detail in the corner of a photograph, or on knowing something the pack never tells you, produces a room full of people who feel cheated rather than outwitted. The distinction that matters: a good mystery should be hard to assemble, not hard to notice. Our own working rule is that at most one link in any chain of reasoning may depend on spotting something visual — the rest has to be plain readable text a careful person will certainly find.

4. Nobody has to change their mind

The best moment in one of these evenings is when someone who was certain has to publicly abandon a theory. If the case never forces that — if the evidence only ever adds to one picture rather than demolishing another — the night is pleasant but not memorable. Misdirection is not a decoration on a mystery. It is the mystery.

Solo, couple, dinner party or team

Evidence-led cases scale down in a way that role-play cases can't. One person with a case file and a pot of coffee is a genuinely good evening — arguably the purest version of it, since nothing gets solved by committee. Two people is our most common play pattern. Four to six around a table is the sweet spot for argument.

Above about eight, an evidence pack starts to strain: there is only so much paper to go around and quieter people stop reading and start listening. That is the point where a role-play dinner party kit genuinely is the better tool, and we'd rather say so than sell you the wrong thing.

If your group can't get to the same table at all, that's a different problem with a real answer — see running a murder mystery at home online.

Before buying anything, it's worth knowing what the free end of this genre actually offers — some of it is genuinely excellent. We laid it out, with links and no email gate, in the guide to free murder mystery games.

What ours look like

Every Exhibit A case is a printable evidence packet — typically in the region of twenty to thirty pages — plus a set of working login screens in your browser. Clues in the paper open the screens; what's behind the screens sends you back to the paper. No app, no account, no download beyond the files themselves.

The design rules are public because they're the actual product: at least three suspects who stay genuinely plausible until the whole picture is assembled, plus further minor leads for texture; no single document that names the culprit; nothing important stated as a conclusion when it could be something you derive; and one full solo testsolve, run by someone who did not help design the case, before it ships.

Step Into the Story

She waited at a table set for two. He never came. Three days later, it's not a bad date anyone's talking about any more — it's a missing person.

Sophie Callahan had been talking to a match called "Daniel M." for four months on an app called Matchlight — long enough to feel like she actually knew him, long enough to finally agree to meet in person. She picked the venue, Amara Wine Bar, showed up on time, ordered for two, and waited. Her messages stopped getting replies mid-sentence. Daniel never appeared.

Three days later, her sister reports her missing. Her phone goes straight to voicemail. Her car is still sitting in the wine bar's car park, exactly where she left it.

Asked to look quietly, before anyone panics officially

The wine bar's owner knows you're good at puzzles, and she'd rather have someone she trusts poking around than a police report she can't take back. So she hands you what she's got — Sophie's message history, a few things her sister's shared, and a growing list of people who might know more than they're saying — and asks you to find out what actually happened before this becomes a story nobody can control.

Here's what makes this one different from the mystery you think you're solving: unmasking "Daniel" isn't the ending. It's the halfway point. Who he really is explains why he never showed up — it doesn't explain who Sophie actually ran from.

More than one person had a reason to be there that night

There's the ex-boyfriend who's been a little too present since the break-up. There's a coworker whose attention had already started to feel like a problem before any of this happened. There's a regular at the wine bar who asked Sophie out once and didn't take the no especially well. And there's whoever "Daniel M." really is — a person who, it turns out, was never being entirely honest with Sophie about a great many things.

Every one of them has a real motive. Only one of them was actually standing outside that wine bar when Sophie ran.

What you'll actually be doing

You're working through a genuine message export, a set of screenshots, receipts, and a couple of accounts that don't unlock until you've earned the right password from something else in the file entirely. This one's built to be worked through as a pair — trade theories, argue about who's lying, and don't assume the obvious answer just because it's the one every dating-show plot would give you.

By the end, you'll need to say — clearly, and with evidence — who actually confronted Sophie outside that wine bar, what happened in the moments after, and why. Genre instinct will point you at the obvious name. The file won't let you stop there.

If you are working out which kind of case suits your group, start with how to run a murder mystery at home.

Step Into the Story

Opening night of the most anticipated production in years. A director's long-awaited comeback. And a counterweight that fails during intermission, on a rigging line that had been inspected and cleared that very same morning.

"The Nightjar" was supposed to be Gabriel Thorne's return — his first major work since a public flop nearly ended his career six years earlier, and the most talked-about opening in the city all season. He was found backstage shortly after intermission began, struck by a falling counterweight from the fly system above the stage. A tragic accident, on paper. Except the venue's own safety officer has already flagged something that changes everything: that exact line was checked and cleared that morning. Rigging doesn't fail on its own, hours after someone confirmed it was fine.

This time, it's not a favour — it's a case number

You're not being asked quietly on the side by someone who'd rather keep this out of the papers. You're detectives with a real provincial Serious and Violent Crimes Unit, assigned a genuine case docket the moment that inspection log came to light. Search warrants, subpoenaed records, full authority — and a file that's going to the National Prosecuting Authority if what you build actually holds up.

Thorne was controlling, meticulous, and difficult in exactly the way that made him dangerous to somebody. Something he found in the days before opening night got him killed. Working out what that was is most of the job.

Three people, three logins, and a killer who isn't the obvious name

There's the production's producer, who has a great deal riding on this show and a lot to lose if certain numbers ever get properly audited. There's the lead actress — Thorne's former protégée, publicly and bitterly split from him weeks before opening — who every instinct in the room will point at first. And there's an uncredited co-writer with the oldest, most sympathetic grievance in the entire case. Add a rigging technician, a critic, and a distant relative, and you've got a full backstage's worth of people who all had a reason to want Gabriel Thorne gone.

This is deliberately our most demanding case yet — three fully-developed suspects, three separate accounts to get into, and a killer we wrote specifically to be the least satisfying answer in the room. If you're chasing the first name that feels right, you're chasing the wrong one.

What you'll actually be doing

You'll be working warrant access into two accounts and public access into a third, checking a call platform's own session log against an alibi that looks airtight until you check it minute by minute, and cross-referencing old production credits to spot a pattern nobody was supposed to notice. Every login is earned from something you found on paper first — nothing is handed to you.

By the end, you'll need to state — clearly, with a docket that would actually hold up — who is responsible, what happened during intermission, why they needed Thorne gone by that specific night, how they had the knowledge to do it, and why the two people you'll suspect first are not, in fact, guilty.

Curious how three suspects are built to stay standing this long? We wrote about it in why every suspect has to earn your suspicion.

Step Into the StoryComing Soon

The story ran on the front page. The correction ran on page two, three months later. In between, a company died.

Kesterley District Council put its waste and recycling contract out to tender for the first time in fourteen years. The long-standing incumbent lost. The contract went to a smaller family firm, run by a woman named Frances Aubrey, on a bid nine per cent cheaper — the kind of result that should have been a good news story.

Six weeks before the new contract was due to start, The Kesterley Clarion ran a front-page exposé: leaked council correspondence and a payment record that, the paper said, proved Frances Aubrey had bought the contract. The council opened an investigation. The new firm's lender pulled its funding within nine days. A firm mid-way through fitting out vehicles and a depot, suddenly with no financing and no ability to earn, doesn't survive that. Ninety-one jobs went. The council quietly reverted to the old contractor — which is the detail almost nobody notices, and the one that matters most.

Three months later, the Clarion printed a correction: the leaked material couldn't be authenticated, and the paper accepted the allegations were unfounded. Frances Aubrey sued for defamation. And that's where you come in.

You're working for the paper that might have ruined her

You're forensic investigators, instructed by the solicitors defending the Clarion, which is fighting this on a public-interest defence. Your job is to reconstruct exactly what happened to those leaked documents — because the defence, and the newspaper's own insurance cover, both turn on one question: was the newsroom fooled, or was it complicit?

That's not a comfortable position to investigate from. You're working for a client whose survival depends on what you find, and there's a real chance what you find sinks them anyway. Nothing here is a crime scene. It's disclosure — committee minutes, tender sheets, weighbridge tickets, newsroom copy, an insurer's own claim file — produced by people who all had reasons to shape what they handed over.

The forged documents are sitting right in front of you

Here's the part that makes this case genuinely different: you don't read a description of the forgery. You get the actual leaked package, exactly as the newsroom received it — and half of it is completely real. A genuine payment, to a genuine firm, for genuine work, made years before any of this started. The other half was built to sit next to it and borrow its credibility. Working out which is which, and who could possibly have built it, is the entire investigation.

Along the way you'll meet a reporter who's been suspended over a story that wasn't entirely hers to get wrong, an editor who cut a verification step that exists for exactly this situation, a councillor with a real and properly declared financial connection to one of the leaked documents, and a council compliance officer twenty-two years into a job nobody ever looks at twice. Every one of them can explain a piece of what happened. Only one of them built the forgery.

What you'll actually be doing

This one doesn't ask you to read someone's private inbox — it asks you to search institutional records the way a real forensic accountant would. You'll be comparing a document-reference format against a system migration date to catch an anachronism nobody meant to leave behind, cross-checking a company register against a client's own ledger, and doing arithmetic on a weighbridge ticket book that looks perfectly ordinary until you sort it by load type.

By the end, you'll need to say — clearly, and with the paperwork to back it up — who built the leak, how it survived a legal read on deadline, and why. Get the who right and the why wrong, and the case still doesn't hold.

Step Into the StoryComing Soon

Your mother said she didn't do it. She was right, and she knew who did.

Seventeen years ago, the sculptor Anton Vael died at Marrow Point — the family's studio and smallholding on South Africa's West Coast — poisoned during a weekend gathering with the people closest to him. His wife, Rosalind, was convicted eleven months later, on evidence that was largely circumstantial but for one damning detail: she was seen washing out the decanter believed to have carried the poison. She died in custody, fourteen months into her sentence.

You were ten years old. You're turning twenty-eight now — the exact age your mother was written into her own trust, the age she was when she married your father. That birthday unseals a personal archive she left behind, and the family solicitor hands it to you along with the rest of the estate. Inside is a letter, written in the months before she died, stating plainly that she didn't do it. And that she knew who did.

This time, nobody hired you

Every previous Exhibit A case has put you on the outside of the story — a consultant, an investigator, someone brought in to look at other people's lives. Not this one. This is your father, your mother, and your own name on the birth certificate.

The family's solicitor is honest with you about what you're actually holding: a route, not a guarantee. A petition under section 327 of the Criminal Procedure Act, addressed to the Minister of Justice, presenting new evidence that materially affects a conviction — a mechanism written for a living petitioner, with essentially no precedent for a case like yours. Your job is to assemble the evidence that petition needs. The report you eventually file isn't a police report or an insurer's finding. It's the substance of your own affidavit.

Five people were at Marrow Point that weekend. All five are still alive.

There's your father's business partner, who'd discovered a discrepancy in a foundry ledger days before your father died and had a great deal to lose if it ever surfaced. There's your mother's half-sister, mid-argument with your father that same weekend over a family cottage neither of them would back down on. There's the estate's longtime manager, who handled the household accounts and never quite explained a discrepancy of her own. And there's a woman who modelled for your father for three years — present at Marrow Point, and, as far as the original investigation ever established, entirely unremarkable.

Nobody in this file is a stranger. That's what makes reading their statements seventeen years later so much harder than it should be.

What you'll actually be doing

You're working through your mother's own unsealed email account, your father's studio business archive, and the original defence file itself — three separate archives, each released to you specifically because of who you are, not because of a badge or a warrant. Somewhere in a foundry pour log, an edition number that was supposed to mean something stops adding up. Somewhere in a caterer's schedule and a family photographer's contact sheet, a twenty-minute window narrows to exactly one person. And somewhere in your mother's own words, never sent to anyone, is the thing the original trial never had a way to hear.

By the end, you'll need to say — clearly, and with enough to actually put in front of the family's solicitor — who poisoned your father, what happened, and why. Your mother already told you she knew the answer. She just didn't get to say it in a way anyone believed.

Step Into the StoryComing Soon

Six people. One road. It went at midnight.

Ardenfell Observatory sits on an upland plateau, ninety kilometres from the nearest town, reached by a single road that crosses a shallow river at a low-water ford. Six people were there for a three-night observing run. On the second night, the rain didn't stop, the ford went under, and the only way in or out of Ardenfell disappeared with it.

Somewhere around 02:00, Dr. Miriam Hallam — a revered astronomer on what everyone knew was her last observing run before her health ended her career — fell roughly seven metres from the exterior catwalk of the main dome onto the concrete below. It should have been a tragic accident on a dark, badly-lit walkway. Then the site manager's own report mentioned a catwalk gate that shouldn't have been open, and a witness account that didn't quite hold together.

It was forty hours before the road was passable again and anyone official could reach the site. Forty hours, six people, one building, and — for at least one of them — the growing certainty that somebody in that room had probably killed her.

Forty hours is a long time to get your story straight

You're detectives with the Hartsmere Constabulary Major Crime Unit, brought in the moment a probable accident became a suspected homicide. But this isn't a scene anyone preserved for you. The first accounts were taken over a satellite phone by a non-police first responder, and by the time you actually arrive, six people have had forty hours together to talk, to tidy, and — whether they meant to or not — to align.

What you get instead is something better than an untouched crime scene: a set of very human, very self-interested statements, and the one category of witness at Ardenfell that had no motive and no forty hours to reconsider anything — the observatory's own machines. Dome rotation logs. Instrument exposure logs. A power event log. A long-exposure camera pointed at the sky the whole time. None of them can lie, and none of them are on anyone's side.

Everyone at Ardenfell had a reason to be up there

There's a visiting instrumentation scientist commissioning new equipment, with a professional history nobody on site actually knows. There's a postdoctoral researcher whose entire career is riding on this run's data, and who's worked out something about Miriam he's not saying out loud. There's the site manager — twenty years at Ardenfell, the only person who holds a key to that catwalk, and the only one who knew exactly which section of handrail was unsafe. And there's a technician who lied about being asleep, and a science journalist Miriam had threatened to freeze out of a story that mattered to her.

Opportunity doesn't help you here — everyone on site had it. This one gets solved on means, motive, and the moment a person's own account stops matching what a machine wrote down.

What you'll actually be doing

You'll build a reconstruction grid — six names down the side, the hour that matters across the top — and fill it in twice: once from what each person says, and once from what the site's own systems recorded. Somewhere in that grid is exactly one cell where those two versions cannot both be true. You'll need warrant access into the site's own mail system, a production order for an off-site university record no one at Ardenfell could have touched, and a password assembled from two documents that, on their own, have nothing to do with this observatory at all.

By the end, you'll need to state — clearly, and against a docket that either holds or doesn't — who is responsible, what actually happened on that catwalk, and why. Rule out the obvious answer first. It's there for a reason.

Step Into the Story

Anton Draganov died quietly, of a heart attack, in his own home. Within a week, his private study had been searched and the one thing he never let anyone touch — a lifetime's worth of secrets, kept in a cipher only he was supposed to understand — was gone.

For decades, Anton was one of the region's most careful black-market antiquities dealers: never flashy, never caught, and — it turns out — never entirely honest even with the people who trusted him most. His personal ledger wasn't just old business. It was proof of exactly what he took, and from whom, across a lifetime nobody else ever fully saw.

Not police work — a recovery, and a race

You're investigating for The Provenance Trust, a private cultural-heritage recovery organisation that had been quietly building a case against Anton for years, waiting for the right moment to either prosecute him or negotiate the return of what he'd taken. His death didn't close that file. It may have just made things harder — because whoever has that ledger now has decades of his secrets, and at least one shipment of genuinely stolen antiquities he never actually delivered.

Nobody's died as a result of the theft. Nobody's called the police, and nobody's going to. What's actually at stake is whether a hidden cache of looted cultural property gets returned to where it belongs — or vanishes into the market for good, in the hands of whoever gets there first. That might not be you.

A house full of mourners, and only one who went into the study

The memorial gathering after Anton's funeral brought together everyone who ever had a reason to care what happened to that ledger: a relative back in the country for the first time in years, a business partner of two decades who gave a short, genuine remembrance to the room, and the caretaker who'd managed Anton's house — and, it turns out, spent a great deal of time near that study — for his final eighteen months.

During that fifteen-minute remembrance, with the whole gathering's attention on one room, the rest of the house sat essentially unwatched. Someone used that window. Working out who requires more than a guess — you'll need an alibi, a motive, and an opportunity that actually survives being checked against the minute.

What you'll actually be doing

This is one of our more demanding cases: a genuine substitution cipher to work out for yourself, using a photograph, a shipping record, and one deliberately misleading example designed to teach you the mechanic without handing you the answer. Once you've got it, you'll be applying it to a real coded ledger entry — and reading three separate accounts, one of which was never meant to be found at all.

By the end, you'll need to say — clearly, and with evidence — who actually took the ledger during the gathering, what drove them to it, how they had the opportunity in that one specific window, and why the other two names on your list aren't, in fact, responsible.

If you want to judge this one against anything else on the market first, the nine checks in our buyer's guide are the same ones we hold our own cases to.

Tips & Guides

Most people who get stuck on a murder mystery game are not missing a clue. They are reading the evidence in the wrong order, and arguing before anyone has finished reading.

A good case file is designed so that no single document gives the answer away. That is the whole craft of it — and it means the skill you need is not cleverness so much as method. Here is the method, in the order that actually works.

1. Read everything once before you theorise

The single most common mistake. Someone finds a suggestive detail on page four, says it out loud, and the whole table starts building a case around it before anyone has seen page twenty. From that point on, every new document gets read as evidence for or against one theory rather than on its own terms.

Read the whole pack first. No theories, no accusations. It feels slow and it saves you an hour.

2. Separate what someone claims from what a record shows

This is the distinction that solves cases. A witness statement is not evidence of what happened — it is evidence of what a person chose to say. A bank statement, a call log, an access record, a timestamp: those are different in kind, because nobody wrote them to be read by you.

Keep two lists as you go. What people say. What the paperwork records. The answer usually lives in the first place those two disagree.

3. Build a timeline before you build a suspect list

Take a sheet of paper, put the names down the side and the hours across the top, and fill it in twice — once from the accounts people give, once from what the documents actually establish. Somewhere in that grid there is a cell where both versions cannot be true.

Suspect lists invite you to pick a favourite. Timelines do not care who you like.

4. Chase the boring documents

Everyone reads the dramatic letter. Fewer people read the ledger rows either side of the one that matters, or the third page of a maintenance schedule. Well-built cases hide the load-bearing fact in ordinary paperwork precisely because that is where a real investigator would have to look.

If a document seems too dull to be relevant, that is often the point.

5. Work out why each suspect is not guilty

Flip the question. Instead of hunting for the culprit, try to clear each person properly — and insist on a checkable fact for each one, not a feeling. A suspect you can only clear with "they seemed sincere" is a suspect you have not cleared.

You will usually find there is exactly one person you cannot rule out. That is the answer, and now you can say why.

6. Say the whole answer out loud before you commit

Who, what, and why. If you can name a person but cannot say what they did or what drove them to it, you have guessed. A well-built case will not let a guess score the same as a proof, and being able to state all three is the difference between finishing a case and finishing it honestly.

Where the digital gates fit

If your case has login screens as well as paper — ours do — the method does not change, but the order matters. A password you have not earned yet is not a puzzle to brute-force; it is a signal that you have not found the document that gives it to you. When you are stuck at a gate, go back to the paper rather than guessing at the keyboard.

And once you are in, read the system properly rather than skimming for the one item you came for. An inbox or an archive is written to be searched, and the thing you were not looking for is often the thing that closes the case.

Practising on something free

The instinct this all trains — treat every alibi as a claim, not a fact, until the paper trail backs it up — is exactly what Golden Age detective fiction was built to exercise. If you want to sharpen it before spending anything, our guide to free murder mystery games lists genuinely free, legal options with no email gate.

And if you are still deciding what kind of evening to run, start with how to run a murder mystery at home. For the design side of the same coin, why every suspect has to earn your suspicion explains what a case is doing to you while you solve it.