Advice on running a mystery? Clues?

[NeoTiamat]

Okay, I've managed to back myself into a corner, so I find myself in need of help. Hopefully someone here can assist me.

Basic situation is that I am running a chat RP and I need to construct a mystery for my PCs to solve. Issue: I have no idea how to make a mystery situation. What's more I'm worried that since I've been raised on a steady diet of Agatha Christie and other whodunnits I won't be able to tell something ferociously obtuse from something very easy. At the same time I want the PCs to work for it.

Background:
The setting is a homebrew arabic-themed domain, lots of djinn, marid, efreet, and the rest interacting as kind of greater nobles to the genasi and humans that make up the populace. One of the PCs, an air genasi, is the heir to the throne who finds himself the soon-to-be-crowned Caliph after his dad croaks. After another assassination attempt and a second murder, PCs have to find the killer. Issue: I've laid the background but don't know how to proceed from here. The killer is a disguised Efreet who is trying to stir up a civil war.

The cast of characters so far (I can add a couple more at need):

Chancellor: Hated the PCs but is seemingly loyal and has come around. Very competent. Was regent till the PCs came.

Steward: Ditzy, sycophantic character the PCs have been ignoring, and who I think will be the eventual killer. PC's finding the real stewards body would give the game away, I think.

Herald: Traitor who worked for the Efreet, but was poisoned in the last session after botching an assassination attempt, so the Efreet killed him to cover his tracks.

Servant: Herald's servant who assisted in the assassination attempt (leading the PCs into an ambush), but managed to get away. Finding him might be another approach.

Foreign Ambassador: Was slain by the Efreet as part of his plan to foment a war. A riot was blamed. PC's haven't realized it was a murder yet.

Deputy Ambassador: Acting ambassador, is convinced this is all some power play, very suspicious, pre-disposed to disliking the PCs.

PCs consist of a rogue, a fighter, and a druid. Druid is the heir.

Current angles I can see following: Find and interrogate the servant (what should they learn?), search the Herald's room (find what?).

Advice on specific clues, etc?

[Pamela]

The first thing to find out is whether any of these players have also had a similar diet in whodunnits, be it book or film. Have you also had puzzles or any form of investigations in your previous campaigns with them? Were they curious and running about asking pertinent questions, or content to be led to the next leg of the adventure?

[NeoTiamat]

I haven't done any mystery/puzzle situations before, which is why I'm uncertain of myself right now. The one PC I spoke to read a bit of mystery, but it was a while ago.

[Igor the Henchman]

When running mystery adventures, I found two approaches that seem to work:

Player-driven method:
The adventure progresses as the PCs uncover new clues and interact with the adventure setting. Works best for plots restricted to one or a couple very evocative locations. The PCs have free reign to solve the mystery in any way they see fit - but make it rather easy. Easy enough that any kind of serious observation or NPC interrogation uncovers the answer. Expect the PCs to come up with surprising ideas, or do things you haven't planned for in advance. Don't bother with red herrings, as they might confuse the players, or get them going in circles. Make all the clues straightforward, and keep things moving, moving, moving. The second the mystery is solved, drop an exciting climax encounter, where they get to stop the bad guy once and for all. The NPCs should act amazed by the PCs' cleverness.

DM-driven method:
This kind of adventure is chronologically-driven. The DM prepares a number of interesting mysterious events that happen around the PCs, and riddles the setting with clues that - if investigated - give the players a more complete picture of what's happening. Red herrings can be used to spice things up, but each "dispelled" red herring should result in the PCs getting a "real" clue for their effort. Each "piece of truth" the PCs uncover should feel as a minor victory for the players, but don't expect them to solve everything. Whatever they do, the adventure ends with a big climatic scene where all masks are dropped and the villains move out in the open. No matter how much or how little the PCs have uncovered up to this point, they get to confront and defeat the villain in that final encounter.

Few players are as smart as Hercule Poirot, but all players enjoy feeling like they are. When in doubt, make it easy instead of tough, simple rather than complex. Above all, cultivate the impression that the PCs are making real progress, whether or not it is indeed the case. If the PCs feel like they're stuck, drop an obvious clue on them, even an improvised one.

[Igor the Henchman]

Steward: Ditzy, sycophantic character the PCs have been ignoring, and who I think will be the eventual killer. PC's finding the real stewards body would give the game away, I think.

If blindly stumbling upon a body feels too obvious, how about stumbling upon the steward's faceless ghost? If they investigate the mysterious haunting, the PCs just might uncover the corpse.

[NeoTiamat]

That is extremely useful, thank you Igor.

Anyone else have suggestions?

[Archedius]

a good technique that goes along w/ Igors' advice is to use the occasional subtle look of suprise on your face as you DM when your players find out something- if you use the right discretion when you do this your players might feel like they pulled one on ya. And there's nothing wrong w/ givng your players the occasional victory (even if you knew it would happen).

[Jason of the Fraternity]

I think that Igor's approachs really cover the basics of what you need, although there is one bit of advice that I would highly recommend: Never assume that what is obvious to you is also obvious to your players. A DM, who I used to game with, tended to make that mistake.

I would recommend that you tend to provide multiple clues that all could lead to the same conclusion. Or, like Igor had mentioned, provide obvious and/or improvised clues, if your players are getting stuck. It is easy to see how a clue solves the puzzle when you know what you've looking for, but the piece might be too obscure when you don't. Having multiple clues scattered about can ensure that missing one detail in a particular place doesn't stump your would-be detectives.

Likewise, be careful about throwing too many red herrings into the mix. A couple can keep your players on their toes, but it can become frustrating if they feel that they're constantly running in circles. Again, having additional clues can help them dispell the "fake" clues for the real ones...

[Dr Ed]

In my experience, the best way to form a mystery is to set the players in a situation that requires them to do a great deal of research and planning in order to solve the mystery/survive -- i.e. pit them against a nemesis that has a very powerful, very secretive edge over the players and has every point of he/she/its existence shrouded from public knowledge. You can further expound on this concept by making the culprit's identity a secret protected by a powerful organization or government agency, requiring a large amount of subterfuge to investigate safely. Place small clues within the information that the characters research, leaving a trail of breadcrumbs that leads from one location to another until the big picture is revealed.

If you're trying to create a basic murder mystery scenario, the way I'd go about it is by visualizing every detail about the event as if I was writing a scene in a movie, then covering it up completely save for one minuscule detail (a thread, a hair, a footprint, etc.). If you have particularly inquisitive and imaginative players, they'll ask questions on their own and come to all sorts of wild conclusions, which you can run with as they make them and shape the mystery into something truly epic. However, if your players aren't used to that sort of roleplaying, you might want to have them make an appropriate knowledge check and fill in the details yourself so they catch on.

Example: "Joe, roll knowledge underground... (success) You remember that the Dead Dogs gang down by the wharf uses the same type of blade that was found at the scene of the murder."

As odd as it may seem, sometimes the best way to come up with a good mystery is to put yourself in the place of the culprit and run through how you would commit the crime. From there, it's only a matter of adding little clues -- little oversights on the culprit's part -- to make it a case worth solving.

[NeoTiamat]

Well, I've started. So far so good.

Thanks to everyone for their advice, I'll let you know how it turns out.

[NeoTiamat]

Huh.... despite my misgivings and utter certainty I'm messing up, my PCs seem to love it so far. Thank you to everyone who contributed!

And thanks Roti, the essays were pretty helpful.

[Rucht Lilavivat]

Player-driven method:
The adventure progresses as the PCs uncover new clues and interact with the adventure setting. Works best for plots restricted to one or a couple very evocative locations. The PCs have free reign to solve the mystery in any way they see fit - but make it rather easy. Easy enough that any kind of serious observation or NPC interrogation uncovers the answer. Expect the PCs to come up with surprising ideas, or do things you haven't planned for in advance. Don't bother with red herrings, as they might confuse the players, or get them going in circles. Make all the clues straightforward, and keep things moving, moving, moving. The second the mystery is solved, drop an exciting climax encounter, where they get to stop the bad guy once and for all. The NPCs should act amazed by the PCs' cleverness.

*snip*

Few players are as smart as Hercule Poirot, but all players enjoy feeling like they are. When in doubt, make it easy instead of tough, simple rather than complex. Above all, cultivate the impression that the PCs are making real progress, whether or not it is indeed the case. If the PCs feel like they're stuck, drop an obvious clue on them, even an improvised one.

This is very solid advice. And well put together as well.

The only thing I would add is that in Ravenloft, every clue or set of clues is an excuse to entice the PCs to travel into a Bad Place. In regular, D&D, you don't need much of an enticement to go into the Bad Place. "Look! It's a dungeon! There's treasure! Let's go check it out!"

However, in Ravenloft, where things are particularly creepy, you need a strong enticement to get your PCs to travel into places they shouldn't be. "I'm not going in there. Did you see that?"

So, if you leave strong indications that the clues necessary are in one of your Bad Places, then your PCs will be strongly enticed to go. If they know that the only way they can find the killer is to travel into the Haunted Tower, then they have a strong "hook" as it were. That's how you can set up your great horror scenarios.

When I run my games, I already have a few Bad Places laid out. As the mystery in the campaign progresses, I occasionally will have a strong reason for the party to travel to one of these locales. The best part is, because they are looking for clues in the Bad Place, they have to explore most of it. It's not something they can just run in and grab. Also, they can't burn down the whole place and just kill whatever runs out.

The downside of this, of course, is that you cannot do this too many times, or it will feel like they are being led by the nose to go to creepy places. Sometimes, you have to make the creepy places come to them.

Anyhow, it sounds like you are off to a great start.