There is something that drives all stories, whether the dramatic tension of a inter-character relationships, the adrenaline pumping of an action adventure or the shrouded enigma of a mystery.
In every story, there is something that stands in the way of one party achieving their goals...if the goals had no obstacles, they wouldn't be goals...and if they were easily achieved, then there wouldn't be much of a story.
Every type of story has obstacles, the genre of the story simply defines the type of obstacles expected.
For decades roleplaying has done action-adventure, where the obstacles are monsters, traps and magical curses.
For years, recent games have started to master the notion of interpersonal drama, where the obstacles are rivalries, love triangles, and situational triggers that prompt a characters primal drives in a given direction.
But recently, I've been hearing questions about ways to run a good investigation or mystery. These have been popping up in a few forums and social media outlets.
I had thought that investigation was just as well covered in the various products filling the roleplaying market...but obviously this isn't the case. Perhaps the 4th edition D&D notion of pulling everything back to combat has tainted the current generation of novice gamers. I even think back to one of my worst roleplaying experiences at Gencon Oz a few years back. D&D 4th had only just come out, but here was a GM who really couldn't handle the concept of making a satisfying investigation story.
So, how would I run this sort of thing?
The way I do it is very similar to the way a session was run at Eyecon this year. We were playing Castle Falkenstein and there was basically a sandbox of potential leads. Players could use a variety of skills to justify their information gathering techniques, then plug together the clues to form a hypothesis about the deeper issues at work.
An investigation works purely on the notion that the players need information to reach the climax of the story. They don't start with the information, and some clues might only become viable paths for investigation once other clues have been revealed.
Consider this in the terms of vector game design theory I was working through a couple of months ago (years ago). The basic premise of gaming vector theory is that a roleplaying game works in two modes; "Story Mode" where a narrative path heads in a straight line and "Game Mode" where a path has the option to deviate along different courses due to player decisions, dice rolls or other forms of intervention.
A traditional mystery twists and turns, it doesn't follow straight lines for long before a protagonist discovers something new and turns things on their head. A good mystery does this and revisits previous clues in a new light with the information gained later in the story. It's like foreshadowing, and while a clumsy GM might make it look like railroading, an adept GM creates something like a beautifully multifaceted diamond that dazzles as more light is shone on it.
Try to develop a mystery to suit the characters involved, provide some potential paths of inquiry that suit the strengths of the characters, and others that are simply beyond their means. Use everything, from conversations, to clues that can be revealed through sheer brute force, obscure knowledge or the traditional Sherlock Holmes style of intuitive deduction and logical reasoning.
Mysteries really aren't that tricky to run.
...and remember "Occam's Razor".