The Forge Archives

Ok, so there was a thread recently that got my brain working on the idea of Mystery games.  I started shuffling through random thoughts about the nature of the mystery genre, how it's applied to RPGs, how I've run them in the past, and what I think would make a good mystery game.

My intention in putting my thoughts here at the Forge is to pick everyone else's brains on the subject; "What makes, or would make for, a good mystery rpg?"

I took a quick peek at past threads with the keyword 'mystery', but didn't see anything that tackles this topic directly.  If anyone knows of a thread or two, I'd love a link.

In all my previous games whenever a mystery of some kind came up, the way I'd always handle it was to create something like a flow-chart of clues.  I'd take a fesh page of paper and write a simple description of an introductiory scene at the top.  Then, I'd draw lines to new scenes, and each of these lines represented a clue that might be found in that introductory scene.  So on & so forth 'till the PCs followed the trail of clues to the solution of the mystery.

While I'm having trouble thinking of what would consitute a good way to run a mystery, I think I can identify a few distinctly distasteful habits from my usual method.

First, I'm in the terribly bad habit of rolling to see if the PCs find the clues.  Perception or spot rolls, investigation rolls, Libary Use; all these and more are guilty of inspiring me to bad habits.  That is to say, because these skills or abilities were included in the games I chose to run, I did what was implied by the mechanisms & provided a chance for the PCs to fail in a less than entertaining way.  I know that this topic has been covered to some extent here at the Forge, but I think it's still a good starting point when considering what would make for a good mystery rpg.  

The second bit that keeps floating around in my noggin is kind of looking at the mystery genre, trying to see what's 'really' going on.  The two investigators I keep thinking of are Poirot and Sherlock.  Each has their sidekick to underline how kewl they really are, right?  So, by looking at Watson or Capt. Hastings, we see what Sherlock and Poirot can do simply by seeing what they cannot.  And it seems to me that what the investigators do not do is -find- clues.

Lemmie see if I can explain what I mean.

The investigators don't so much see the clue sitting there, as they identify that what they see is a clue.  Hastings might see the item sitting about the crime scene, but it's Poirot that makes the connection between the item and the plot at hand.

I'm imagining that the 'secret' to a good mystery game might just lie in that idea, but I'm having a terrible time prying it loose.  It's like that last little piece of tasty walnut within the shell.  Can't shake the bugger loose.

Anywho, I don't know if any of this makes any sense, or if it's all ground that's already been covered somwhere, but I'd so much like to see a tasty mystery game spawn from it, I'd love to hear everyone's thoughts.

-Eric

The first thing that came to mind while reading your question was the relationship map from Sorcerer. If you don't know about these I'll try to do my best to explain them in a moment. But first...

Unless the PC's in your game are Police Officers or Detectives then almost all of the modern mystery fiction is useless for your purposes. The PC's aren't going to have access to the sort of resources that a Police lab has (i.e. DNA, finger-print comparison, and so on). Therefore, the PC's will instead have to start asking a lot of questions. They'll have to uncover the "dirty laundry" of those who are or might be involved, and most people aren't going to make it known that they've been screwing their business partners wife. And this is where the relationship map comes in.

What you need to do first is complete the crime. Ask yourself the W's. Who killed the victum(s), what did they use to kill them, where did the crime take place, when did it happen, why did the killer(s) do it? Once you've answered these questions you'll have a small list of names of the pricipal players.

Now, get yourself a blank sheet of paper and space these names out like they were on a family tree. Next figure out all of the people around this pricipal group who have some sort of insight into the principal groups relationships... parents, wives, husbands, kids, affairs, close friends, business partners, neighbors, etc... Each of these people will be more or less willing to divulge information about "the tragic event".

Now, put these secondary players on the map, grouping them near the pricipal player that they were closest to. Draw in the lines that represent the different relationshps involved, but only the inportant. (We don't need a line between John the banker and Billy the gas station attendant unless they were having an affair or something...) Make sure to lable each line according to the type of relationship involved (marriage, kids, affairs, etc..) Oh, don't forget to include on the map people who are already dead but may have had or even have been the reason for the crime.

Once you've got all these people on the map, you've got your mystery. You know who's who. You know why Mrs. Johnson was screwing the guy at the butcher shop, and why her husband killed her sister thirty years ago. Some of these relationships may have nothing to do with the mystery at hand at all, but that doesn't make them any less important to the NPC's invovled.

One last thing to remember. EVERYBODY LIES! I can't stress this enough. Most of the people lie for no reason at all, we can't help it.

Hope this helps.
Merrick

Over on RPG.net, at least, the stock advice on running mysteries is that you should not have a preset answer. Leave a lot of clues, wait for the players to develop theories, and let the first or second theory they don't abandon on their own turn out to be correct.

I've had more success with that approach than with the more obvious way of deciding who it is and planting a series of clues that the players are likely to completely miss, but I have not tried the relationship-map way of running a mystery and cannot compare the two.

Here is another approach:

Decide the answer. Tell the players the answer.

Then make them generate the clues and the path of reasoning that gets their characters to that realization. It's worked for me every time.

QuoteIf anyone knows of a thread or two, I'd love a link.

Eric:

The difficulties with running mystery-style games got a definite work out in a series of threads last spring about Call of Cthulhu, particularly:

Cthulhu's Clues: Why failure is not an option

Also in that series of threads:
Drifting to R'lyeh (on CoC in general)
Hot Lead and Hypocrisy (on guns in CoC)
(but the Cthulhu's Clues thread is most relevant for your purposes)

QuoteThe investigators don't so much see the clue sitting there, as they identify that what they see is a clue. Hastings might see the item sitting about the crime scene, but it's Poirot that makes the connection between the item and the plot at hand.

I think you are on to something there, and I like Shreyas' suggestion a lot too. The most drastic solution to the mystery problem I can think of is the game InSpectres, which was created in response to just these issues.

Rob

One technique I've used in the past is to allow players to collect undetermined clues as well as predetermined one.

Here's an example: characters arrive at a crime scene and start looking around.  I give a few obvious clues "The victim was stabbed very precisely with a knife.  It's the kind of thing they teach military special forces.  The knife is gone." or "The victim was killed without a sign of a struggle."  I'll usually give some predetermined clues without any rolls required (this is one of my tips in that Cthulhu's clues thread - deliver essential clues as freebies, and require rolls only for useful, but optional tips).  

Now, at this point, I'll have people start making rolls, such as to analyze the crime scene by taking fingerprints and all that stuff.  If there are any of those optional clues, I'll let them know based on the rolls ("The killer wears size 12 boots").  But I might also say "There's nothing obviously useful in the evidence you've collected, but you might spot a connection later on - keep this as a clue."  

This undetermined clue can then be used later in the investigation if the players get stuck.  For example, in the essential clues, I mentioned the killing tactic, so players have an obvious special forces lead.  But maybe they pursue that and don't talk to the right people (or perhaps even fail to recognize the importance of a clue that's given).  At that point, they can cash in the undetermined clue to get themselves unstuck and I'll give them information that will get them on the right track.  For example "You remember that you found traces of mud at the crime scene.  While on the base, you saw the same kind of mud near Barracks 12."

I think this does a good job of genre emulation here, too.  A lot of Sherlock's clues aren't necessarily noticed or understood right away.  At some point in the story, someone's missed appointment or a mis-placed item suddenly makes sense and nails down the killer.

It also does an excellent job of summarizing modern crime data.  A good forensics team can collect reams of evidence, much of which isn't useful until more evidence is collected.

The only other tip I have: don't make criminals too intelligent.  Serial killers often want to be caught - they won't just turn themselves in, but they will get intentionally sloppy.  Arsons (and many others) desperately want recognition for their work, so they almost always tell someone.  Many criminals (serial killers included) often show up at their own crime scenes and offer to help the police, or they'll collect news clippings or talk about the crimes non-stop with friends or family.  And let's not forget the plain stupid: the guy who robbed a bank with a note written on his own ATM receipt, or the federal reserve robbers who spent $4 million in the month following the theft (and who then refused to give $50,000 to an accomplice, who went to the police).  

But many GMs envision criminals who have no past criminal record, never talk about what they do, who meticulously erase clues and who have no friends or family to become suspicious.  Furthermore, all of their accomplices are willing to die to keep the secret.  By artificially elminating "realistic" types of evidence, GMs often make the mystery much harder to solve than it should be.

It happens that as I'm working on Multiverser: The Third Book of Worlds, one of the worlds, The Vorgo, contains a mystery. It's a robbery, not a murder, but the principles are the same. Thus I'm playtesting it. One of the things that I particularly like about the design is that among the myriad of suspects, any one of six could have done it; the referee has to make minor adjustments in the information to select between them. So far it has worked fairly well.

I think it's important to recognize in designing a mystery as an adventure that there are three places that the challenge can arise:

[*]Finding the clues that matter.[*]Recognizing that certain pieces of information are in fact important clues.[*]Solving the mystery based upon the clues.[/list:u] The question is, where do you want the mystery to fall?

You also have to ask yourself why the player character(s) need to solve the mystery. In Vorgo, the answer is because he is the prime suspect, and someone is going to pin this on him if he can't work out what really happened. Of course, if they are playing the detectives or police, you've got a built-in motivation.

In this regard, you've called attention to Poirot and Holmes and the contrast between them and their sidekicks. Of course, how do you do that when you've got a group of player characters? Making the clues obvious to one player and not to the others is near impossible; you'd either have to communicate extra information secretly to the one or have to have the collusion of the others. Multiverser solves this by tying up one player character in the mystery while the others are involved in separate adventures; that's not always an option. Your best bet is usually to let the player characters corporately play the sharp observant detective, and let the non-player characters be the dimwitted sidekick role.

If you make all three of the above challenges part of the mystery, you've created an extremely difficult puzzle, and probably one which is more difficult than the genre literature. That is, if you read a typical Agatha Christie book (and I've read many), the problem is not working out what the clues all mean, but what the clues in fact are. If you had a complete list of the clues, in most cases you would know who did it. On those rare occasions on which having all the clues spelled out in front of you doesn't make it seem obvious, the clues are made very plain by the author to the reader ("And of course the interesting matter of what the dog did in the night." "The dog did nothing in the night." "Exactly.")

If the puzzle is about gathering the clues and recognizing them, it's fine to keep them a bit secret. Interestingly, I'm running a few articles in the Game Ideas Unlimited series at the moment on adventure design, and particularly quest-style adventures, and it is often the case that mysteries follow these models--either the Treasure Hunt model (in which each clue tells you where to look for the next) or the Scavenger Hunt model (in which you're looking for many clues which could be anywhere, but there's no particular order in which you have to find them). In this case, though, I suggest this balance: to the degree that a clue is obviously important, it may be concealed; to the degree that it is likely to be overlooked, it must be placed in the open. Thus as players look for clues, they will be certain to see all the bits of information that they don't recognize as clues right away, but anything for which they have to dig will obviously be important as soon as they see it. There are two other important considerations with this model. One is that having most of the clues should get you most of the way there. The other is that if you have all the clues, there should be exactly one solution to the mystery, and it should be obvious to a third grader. That's because in this model of a mystery what matters is finding the clues.

The other model assumes that the puzzle is figuring out what the clues mean. In this case, it is absolutely essential that the players get all the clues, one way or another.

Try doing a search on the phrase "moving clue", and see what you find. This is an idea Ron has put forward specifically for running mysteries. The basic concept is simple enough. If the players absolutely must have a certain piece of information, they will get it, even if they never look in the "right" place.

The example of this that I always use take the typical English mansion murder mystery. The master of the house left at three in the afternoon, and that piece of information is absolutely essential to solving the mystery. However, he left rather quietly, and only one member of the household knows that he did.

In a traditionally designed mystery, the design would tell you who has that piece of information, and the character who knows it does not realize it is important. Let us say in this case the chauffer saw the master in the garage and asked if he wished to be driven somewhere, and the master said no, he was just going to drive himself, and left. If the players fail to elicit that information from the chauffer, they can't solve the mystery. The moving clue prevents this. Exactly one character knows that the master left, but the referee has not decided which character knows it. That decision is made at the moment the players have asked a character a question that should elicit the answer. Perhaps the cook saw him pull out of the garage, or the maid was resting in the upstairs sunroom and saw the car speed away, or the gardener waved to him as he headed out the drive, or the butler offered to get his jacket for him--one character has the information, and provides it, but which character has it is not predetermined.

Using illusionist techniques of this sort, the referee skillfully ensures that the players are given all the clues; they can't really not get a clue. That leaves them to sort through what they know and derive the solution, which is the real challenge of the game.

I hope that's helpful. There have been some excellent ideas here, but I think this line will get you closer to the traditional mystery literature in a lot of ways, particularly if you want the players to have the experience of solving the mystery, as opposed to writing it.

--M. J. Young

Hiya all,

Lots of great ideas coming forward.

Thanks especially to Rob for finding those Cthulhu links for me.  I'd read all those threads before, but reading them again today helped alot.

To cover a few points,

My thoughts on Watson vs. Holmes and Hastings vs. Poirot were not intended to put forward the idea that one PC should get the clues and others should not.  Indeed, I only wanted to illustrate the idea that was bouncing about in my head that perhaps _finding_ clues isn't the only option.  Maybe thoughts on Deduction will lead us to a more satisfying mystery game.

Further, it's Illusionist techniques that I'd love to avoid.  Ron's ol' Floating Clue is old-school stuff to my group.  I couldn't tell you where we got the idea from originally, some old module I'm sure, but we've been playing off the Illusionist techniques for a while.  As a GM, I'm starting to feel a bit dissatisfied with them.  Not sure why.  They're just starting to feel flat to me.  Maybe I'm just looking for something new.

I rolled Sheryas' suggestion around in my noggin for a bit.  The idea that the players get the solution right off the bat, then RP their characters through the clues.  I've gotta say, that sounds pretty unsatisfying.  Like reading the last page of the novel first.  May be fantastic for some groups, but I think my players would boo me out of my GM's Comfy-Chair if I even suggested such a thing.  It's a good approach, and I'm glad it got brought up, but I just don't think it's for us.

What I do really like is Zaidaco's suggestion of the R-Map.  Now, I've never read Sorcerer, but I think I've got a pretty good idea how they work.  And yea, I think just writing up everyone's connection in the story and letting the PCs explore it would be terrific.  I think perhaps this would be one first good step in creating the type of mystery game I'm looking for.  However, I imagine it just cannot act alone.

I sat down a little while ago and tried my hand again at writing a bit of an essay on what I'd like to get from a good mystery game.  I started considering all the mystery games I've ever had 'fall apart'.  The ones where the PCs missed a clue I'd worked in as essential, and I just couldn't come up with a satisfying way to get the ball rolling again.  Amongst other things, it occurred to me that I've been habitually guilty of PC clue-blocking.  But I'm not sure how to correct this.

Lemmie see if I can explain what I mean.

I could probably give a hundred examples of my own guilt in this manner, but what it comes down to is; The players come up with a nifty method for finding information.  This method happens to be one that I, while writing up the scenario, didn't fathom.  Unless it's blatantly obvious that there's no reason why the PCs shouldn't get the information they seek, my brain finds every possible reason why that path won't give the information the PCs want.  This eventually leads to two possiblities.  The first being that the PCs manage to find the clue-path I wrote in the first place.  The second being that the players just flat-out run out of ideas.  Poopy.  I'm so guilty.

So, having recently purchased Dogs, I've found that one little rule that's likely to change an awful lot of how I run my games;  Say yes or roll the dice.  Wow.  I'm sure that that idea is as much of a 'duh' situation to you guys out there as illusionist techniques are to my crew, but I'll tell ya', it was a lightbulb over my head to be sure.  

So now, in an effort to break that terrible old habit, I re-consider my entire philosophy in mystery writing.  Where I once wrote the classic 'Treasure Hunt Model' (as MJ put it) where one clue led to another, and the loss of a clue killed the momentum, now I'm looking for another way to do it.  One where I don't need regular reliance on illusionist techniques, and yet don't have the tragic probability of complete story-line derailment.

Whew.  Lots of my chest.  

-Eric

Quote from: Technocrat13I could probably give a hundred examples of my own guilt in this manner, but what it comes down to is; The players come up with a nifty method for finding information.  This method happens to be one that I, while writing up the scenario, didn't fathom.  Unless it's blatantly obvious that there's no reason why the PCs shouldn't get the information they seek, my brain finds every possible reason why that path won't give the information the PCs want.  This eventually leads to two possiblities.  The first being that the PCs manage to find the clue-path I wrote in the first place.  The second being that the players just flat-out run out of ideas.

I'm not sure exactly what you're talking about here, but there's a saying for high-level d&d play which might be applicable: "don't ban powerful abilities, require them." That is, once the wizard has the teleport spell, don't try and trap them in a cabin in a snowstorm with a killer on the loose, because the wizard will just teleport out. Instead, have the killer be on the loose in two cities on either side of the continent and anyone who can't teleport is going to have no way of tracking them down.

So you don't need to move the clue around to make sure the PCs will find it -- put the clue in some hard-to-get spot, let the PCs know it's there, and then wait for them to think up a clever way to get to it. You don't need to have come up with a solution as long as they can.

Hi Eric,

I find the best way to push mysteries is to make every NPC very talkative.  Each one has lots to say about the situation, which is almost always a mixture of truth and falsity.  Most characters spend more time talking about other NPCs rather than themselves, and tend to put certain people in good light and others in the bad.  Whatever bits of truth fit with their bias(that isn't going to work against themselves), is the parts to filter through.

When pressed about something they feel bad or shameful about, or about a person they're trying to protect, either they admit part of it(and lie about the rest), tell a BIG truth about someone else to deflect the focus, or else completely break.

The players get a big tangled mess, and try to follow the lines and figure out what is true, what isn't.  And it works perfectly with R-maps.  Most of the preparation goes along the lines of "What happened" and "How everyone feels about everyone else".  After that, the players can talk to anyone, and they WILL get information, although discerning what is what is the challenge.

Chris

I like the point about Watson and Hastings (and all those other second-fiddles).  Seems to me that one possible way to produce this effect is to have a narrative power that says, "X is a clue," with the caveat that the same person cannot say what it means.  So if you had three PCs working out a mystery, the identification of clues cannot come from the same person who works out their meanings, and so long as these offices are traded around moment by moment you should be able to produce the effect you're talking about.

A really brilliant example is Nero Wolfe, by the way, since Archie Goodwin is anything but a mindless, "Gosh you're clever!" sort of sidekick.  In essence, he does all the hard work, and Wolfe does all the stuff that requires genius.  That's the theory, anyway, and it makes Wolfe look better to give him a sidekick the reader can admire.

For a more fantasy spin on the non-clueless sidekick, a good inspiration would be the Lord Darcy stories.

Quote from: BankueiI find the best way to push mysteries is to make every NPC very talkative.  Each one has lots to say about the situation, which is almost always a mixture of truth and falsity.  Most characters spend more time talking about other NPCs rather than themselves, and tend to put certain people in good light and others in the bad.  Whatever bits of truth fit with their bias(that isn't going to work against themselves), is the parts to filter through.

When pressed about something they feel bad or shameful about, or about a person they're trying to protect, either they admit part of it(and lie about the rest), tell a BIG truth about someone else to deflect the focus, or else completely break.

The players get a big tangled mess, and try to follow the lines and figure out what is true, what isn't.  And it works perfectly with R-maps.  Most of the preparation goes along the lines of "What happened" and "How everyone feels about everyone else".  After that, the players can talk to anyone, and they WILL get information, although discerning what is what is the challenge.

You know, that's almost exactly the same approach Sophocles used in Oedipus Rex.  To a lesser extent, Shakespeare used this in Hamlet as well.  I can't really offer any ideas or advice more substantial than to read or look at various media that have mysteries that interest you.  Consider The Usual Suspects, Chinatown, LA Confidential, or one of M. Night Shayamalan's films and watch very closely how they set things up, especially how the relationships between characters influence how the truth comes out.

One thing that I'd consider to be important is what happens after the crime is discovered.
A lot of mystery fiction seems to revolve around people trying to cover their tracks, and not just the actual perp. The kind of American style mysteries that I like ( Chandler [ironically, and englishman by birth], Mosley) tends to have a ton of suspects, all of them with dirty secrets.

Inevitably, there are people running around after the fact trying to divert attention away from themselves and cover up their skeletons. In fact, more often than not, a mystery novel seems to cover several minor investigations, each one that in turn eliminates a line of inquiry to the main mystery.

One thing that I might watch out for is creating a mystery best solved by deep understanding of the game system. I recall an old Dragon magazine AD&D locked room style mystery that I read many moons ago. Frankly, I doubt that the group that I ran games for would have ever solved the thing, as it involved so much magic in certain very specific ways that I could barely decipher it, even with the solution given to me. Does anyone recall the adventure I'm thinking of?

A couple of things you might want to check out:

Murder in Harmony ( TSR: Gangbusters)- I saw this one up somewhere for dowload. An actual attempt at a 20's era mystery. Worth checking out to see someone else's attempt. Some physical evidence, but a ton of motivational evidence, skeleton filled closets, and papertrail evidence. Also includes a side motivation for just about everyone involved, leading to several minor mysteries to muddy the trail. Oh yeah, and gunplay. What's Prohibition era gaming without a shootout or ten?

The Simple Art of Murder ( Raymond Chandler)- Chandler's essay on mystery writing and differences he saw between English detective fiction and his own and others' American style.

The Big Book of Conspiracy Theories (Paradox Press)- Check out the section on the JFK assassination. It screams multiple party cover up, with all kinds of folks attempting to cover up either their involvement or their incompetence. Heck, even working from the idea that ol' Lee Harvey was a lone nut that got off a couple lucky shots, there's still tons of mysterious after the fact going-ons.

Incidently, has anyone played a sort of- "Blame it on Jeeves" scenario? Basically, the players are the suspects, one or more of whom actually committed the crime. The goal here would be more how to solidly pin the rap on some poor sucker than to solve the mystery. Of course, the explanation needs to be pretty air tight, because even the flatfoots aren't completely stupid, and the victim was high profile so they aren't likely to let the case drop as if it was a dead hooker report. Messing things up further, several of the would-be patsies know dirt on our prime suspects and might not be afraid to let it out if it could save their own neck.

Come to think of it, that would probably a good set up for either Universalis or Chris Engle's Matrix Game (feeling competitive?), both of which allow for lots of imput by all participants...

Quote from: inkySo you don't need to move the clue around to make sure the PCs will find it -- put the clue in some hard-to-get spot, let the PCs know it's there, and then wait for them to think up a clever way to get to it. You don't need to have come up with a solution as long as they can.

I completely agree. In several Nobilis games centered about some kind of investigation, I had not clues, but some story and facts. PCs discovered these facts with their powers (ask a tree, usa a Domain miracle, etc.). Before the game I had no idea *how* they would have discovered them. This keeps a bit of railroading (some clues *must* come out in some way) but players are more involved in their creativity.

L.