Yeah, this is really only an issue for people who have decided from the start that the PCs "must" win the scenario, that they have to come out solving the mystery.
And if that's what you want, then why not do it Gumshoe style? You're playing a fucking railroad anyways, its just a "railroad to victory", but its still a railroad!

On the other hand, if the players fail to pick up vital clues, then what happens next is that either they spend a session doing a runaround and going nowhere, or something else happens.

It depends on what they're trying to solve: a murder from someone who did it as an isolated case and won't kill again?
Well, in that case, just tell the Players they failed. Or have them see some other, better, detective on the news having just arrested the guy they spent months trying to uncover.

A mass murderer? Well, if they failed the first time, the good news is that Dexter will kill again. They'll get to try rolling again, with the second set of clues.

An evil wizard/cult/cthulhu-summoners? Well, this sort of adventure should probably be on a time limit anyways.
If the PCs get close enough that they even look like they might be finding some clues, even if they really aren't, the Cthulhu-cult might try to kill them, and assuming they don't succeed, that might give the PCs some new clues.
Failing that, I guess the PCs will have a VERY clear idea of what's going on once the skies turn red and tentacle things start coming out of chasms in the ground.

RPGPundit

Quote from: Halfjack;334968Run the mystery as a tactical game so that process can be handled piecemeal. Abstract the investigation space as map, establish movement rules based on investigative skills, abstract the solution as a pawn to be moved on the map maybe, and play it a turn at a time, counting turns as time spent. When done, narrate the result by interpreting the board state (who's where, what tasks have been accomplished or failed, how long did it take to get there, etc.)

I am intrigued by this idea.

I'm not sure if I like it, but it sounds very interesting.

I have a couple of methods at my disposal, depending on the game in question.

Call of Cthulhu

The way I do it in CoC is to prepare lots of clues ahead of time, especially as handouts.  There's a story of what's really going on, and I'll put pieces of it in newspaper articles, excerpts from books, or interviews from NPCs.  I'll put it out of order and slightly redundant, so that the players actually get a sense of putting together a mystery.  They have newspaper articles about murders at the old Parker mansion years apart that no one else connected, an invoice from a shipping invoice for several large crates from Eastern Europe just before the first murder, and Old Lady Belry, who says every year on the same week large beasts escape the mansion and hunt people down.  They put together that Mr. Parker has some nasty beast his family brought from Europe that feeds once a year, even if they miss their rolls to find Parker's journal or to see the Evil Beast tracks.

I also try to be open to new ways of getting information - one PC had terrible Library Use, but being a reporter he had a lot of Persuasion.  So if he couldn't find something himself, he'd try to sweet-talk a librarian into looking for him.  Another would use her Accounting skill to figure out who really owns the property that the evil cultists were seen meeting at.

Buffy

I've come up with a system in Buffy that works well.  One player described it as 'fighting with books'.  Where Cthulhu generally involves understanding the history of the place in question, Buffy relies more on identifying the beastie responsible and how to kill it, as well as emphasizing teamwork.  On the show, the kids always helped Giles study, and even though he was an expert sometimes it would take days to find something.

The short version is this (and this will work for pretty much any system with success levels).  Each mystery has a difficulty and an obscurity.  Difficulty is how much material one will have to go through to identify a given creature/spell/whatever.  Obscurity is how hard that information is to find.  So a demon comes into town and the GM decides it is difficulty 3 obscurity 5.  The demon is pretty simple itself, so you need 3 successes to get all the information.  However, its not a well known species, so the first five successes on a given roll don't count!

The other PCs can pitch in too.  The most common is doing a roll that will add your succeses to the main researcher's roll as a bonus.  You can try using different skills, or just go for donuts if you aren't good at research.  The Obscurity is usually the main block, and it is lowered if you get a good idea what you are looking for.  For example, if they do some recon and figure out that this demon prefers caves filled with guano for a lair and eats the liver of his victim, the GM decides that this lowers the obscurity from 5 to 2, making success much more likely.


InSpectres

I've not tried this, but it is a solution.  This is very 'indie', and there's a certain brand of player that will run screaming from the room if you try it.  Essentially, when making a roll to find clues, success indicates not that you found something, but you get to decide what you find.  So a client wants you to investigate creepy lights in his restraunt, you make your roll and declare that your ghost-detector-9000 says it is haunted by a Class IV Poltergeist, the victim of a brutal murder most likely.  If you fail, maybe the GM has the ghost throw a chair at you or something.

This game solves the mystery problem by not having the GM make it up ahead of time.  The players determine what is really going on.  The game balances it all by having the PCs need to make money for their franchise.  Certainly, you could say that every house you visit isn't haunted at all, but you're a ghostbuster.  If there's no ghost to bust, you don't get paid.  So you want to make it bad, but not too bad.

Sounds neat but certainly not right for every game.

The Sherlock Holmes method.

I actually read this idea on The Forge, but it sounds kind of cool to me.  I still need to try it sometime.  In Sherlock Holmes and other mystery stories, Sherlock looks around a room with Watson, perhaps noting some seemingly mundane detail, before concluding the murderer was the butler all along then explaining how it all happened.  This makes for a good story, but is hard to do in a game.

Because on the face of it there are two options.  One, the GM tells the players every mundane detail, including the few that are actually relevent.  But this isn't really fun for anyone involved, and is likely to bog down in play.  Two, the GM calls for notice rolls and for each successful one tells them about one of the vital clues.  This means that either they blow a roll and can't put the links together, or they are handed the mystery.  Might as well do the GUMSHOE solution in that case.

This proposes a third option.  The players are presented with the Scene.  A man in a locked room was shot in the back of the head.  They come up with a proposed solution - Perhaps the assailant climbed out the window!  Now we do Notice checks.  The window is painted shut, and locked from the outside.  Clearly that isn't the solution.  This continues until they solve the problem.

Overall I think a good mystery should be like a good combat encounter.  One that the PCs can win, just barely, and might possibly lose.  If its too easy, as Pundit says, you're just on a railroad.  If its too hard, everyone stares at each other confused.

Quote from: RPGPundit;334999And if that's what you want, then why not do it Gumshoe style? You're playing a fucking railroad anyways, its just a "railroad to victory", but its still a railroad!

GUMSHOE games are about the characters (and immersion), not about the story. Sure, you'll get all the clues... but what do you /do/ with them? And what's the cost?

I can see how this would turn someone off if they thought the story was more important than the characters, though.

a) Don't have necessary clues. Have redundant clues. If the characters miss everything, let them fail.

b) Give them clues automatically, with different levels of detail for successful rolls. Mix in irrelevant clues. The standard lipstick example: "I search the room." "The room has the general disorder of an occupied hotel room. The bed is unmade. There is a wine glass on the bedside table and another, stained with lipstick, broken on the floor. A napkin, also with lipstick marks, is under the bed." On a successful roll, the lipstick is red. On an extraordinary successful roll, the lipstick is a deep, rich red and iridescent. If later characters who got an extraordinary success are interviewing a femme fatale who is a suspect (everyone's a suspect) and don't think to ask what colour her lipstick is, let them fail.

The PCs almost always receive the clues. I never make them roll to get the important stuff, but I do make sure they state how they are gathering clues. If I mention the study has a desk with a locked door and they don't bother trying to get it open somehow, they miss the clue. If they don't ask the NPC the right question, they won't get the answer. Usually, later on I'll have another NPC say something like "You did ask him where the underground lab was hidden right?" and then they slap themselves on the forehead and go back to track down the suspect to interrogate him again, or find another source of info.

Where I run into problems the most is when people don't interpret the clues properly or can't make the right intuitive leaps to find the answer. Then I'll usually have an NPC handy to make a few suggestions to get them going.

Once in a while, I don't have a set end point. I let the PCs gather clues and come up with their own solutions to the mystery and if one sounds good, I just make that the correct choice and throw in a piece of evidence or two to back up their suppositions. Of course, this only works if they have no idea that's what you are doing. I'll flatly deny this technique to my players if they ask.

Quote from: GnomeWorks;335065I am intrigued by this idea.

I'm not sure if I like it, but it sounds very interesting.

We use it as a fraction of Diaspora (one conflict system of four, with general purpose and not specific to investigation) but are homing in on it as a general mechanism in a couple of R&D projects that may or may not see the light of day. It's fun and it solves certain problems -- it's not always what you want for problem solving, but it does break down the barriers at an otherwise paralyzed table, which happens sometimes (infinite planning is sometimes a symptom).

It's an elaboration of an idea from Fred Hicks several years ago, so I'm not claiming originality here.

There is always the Burning Wheel/Burning Empires/Mouse Guard solution...

when they fail a roll for a clue, they find the clue, but they also find some form of nastiness in the process. Sure, reading the plate gives you the information to cope with the Mummy... but it also alerts the traditional guardians that the mummy is up, and now THEY, also, go on a killing spree.

Quote from: Maddman;335072Overall I think a good mystery should be like a good combat encounter.  One that the PCs can win, just barely, and might possibly lose.  If its too easy, as Pundit says, you're just on a railroad.  If its too hard, everyone stares at each other confused.

It sounds like you are describing a riddle instead of a combat encounter, but those two things do have their similarities.  However, I disagree that the difficulty of the riddle has anything to do with it being a railroad.  If you are playing an RPG where the objective of the game is "Solve the mystery" you are playing a game where you must answer one riddle or game over.  That is a railroad regardless of everything else.  In standard RPGs points are given for dozens of potential actions, all of which are riddle solving.  Those include easy ones determined without barely thinking they were even riddles and hard ones missed time and time again and perhaps never recognized by the players.  

Quote from: ThanlissGUMSHOE games are about the characters (and immersion), not about the story. Sure, you'll get all the clues... but what do you /do/ with them? And what's the cost?

I can see how this would turn someone off if they thought the story was more important than the characters, though.

Characters and story are irrelevant in most roleplaying games.  Most mystery RPGs are about the players being the mystery solvers.  If they fail in that role in one situation, then it's best to have another mystery handy.  

Quote from: aramisThere is always the Burning Wheel/Burning Empires/Mouse Guard solution...

when they fail a roll for a clue, they find the clue, but they also find some form of nastiness in the process. Sure, reading the plate gives you the information to cope with the Mummy... but it also alerts the traditional guardians that the mummy is up, and now THEY, also, go on a killing spree.

Per my original post, IMO that's not a mystery solving game at all.  That's a game where the players screw around until enough time passes to when they have received every clue.  And then they answer the riddle or not.  Not that answering the riddle is the point of the endeavor anyways.  If it is, then it is a railroad all over again.  No choice EVER matters in regards to hunting down the clues (re: solving the mystery).