
I’ve done several of these for The Gauntlet: Kuro, Dresden Files Accelerated, Tales from the Loop, Mutant City Blues, etc. It’s also one of the few GMing areas I’m consistently asked about (alongside “Don’t you know Fate sucks?"-style inquiries). Recently, Eduardo L. hit me up with a couple of queries. I thought I’d offer and expand on my original answers to him.
"When you're developing your mystery - are you mapping out the scenes and clues like you might do with a traditional Gumshoe game? In other words, how do you design the clues in your scenarios and individual scenes?"
I held off on an answer because I knew I’d be preparing a fresh mystery for this week’s Bubblegumshoe. So, like most GMs (I assume), I start off with a basic concept. Usually that’s a crime or a major scene. For my Dresden Files game I started with "Someone steals a Chinese Hell." For Kuro I had “Someone’s hacking people’s subdermal nano-tattoos to summon demons." For this week’s Bubblegumshoe, I started with "Someone steals a crop of highly valuable ginseng."
The next thing I think about is the entry point. Who is calling them in, what’s the connection? That’s my lead for figuring out the cast of characters. That gives me a general idea of the kinds of people and circles the PC might interact with, but at that point it’s nebulous.

But in a mystery, those details matter. Engaged players listen and react to those details—looking for patterns and anomalies. Mystery games lean into that. Not everything has to be planned, but I like having a scaffolding. Plus writing those details out helps frame them for me and shows me some of the gaps.
Anyway here’s what I wrote for BGS:
- A & B decide to steal A’s Uncles ginseng crop
- Film group goes out to film an outdoors scene adjacent to Uncle’s field.
- During filming, they catch A&B transporting the stolen goods. Neither realizes.
- A and/or B realizes at school that they may have been caught by camera.
- A talks to C from AV Club trying unsuccessfully to get access
- B tries to break into school; leaves evidence
- D &E from AV Club argue about film. (Red Herring)
- B talks to F, the cameraperson. They used to be friends. B threatens F with blackmail for something.
- F steals hard drive with the video footage.
- B goes to his adult contact.
Then I went through and sketched names and archetypes. That gave me some locations right off the bat: where an AV club member asks the gang for help; the AV club room; the shooting location; the Uncle’s farm; A/B’s hangouts; the reseller’s place.
Usually I then write those places in a kind of flow chart, listing characters at each and drawing lines between them. I try to write out how the PCs can get from A to B, but some of that’s just in my head. I’ll come up with names as well—trying to make sure each one has a distinct sound.
Usually I do this in about an hour just before the session. I only write down core clues and trust myself to have other clues that make sense available. The other day for Trail of Cthulhu I got halfway through the hour, scrapped my initial idea and started over.
Sidenote: For my second attempt, I lifted mercilessly from the 1990’s horror film Lord of Illusions. I changed details, approaches, events, but kept some of the themes and incidents. This is a decent approach, especially if the movie source gives the audience a lot more info than the characters. Because they’ve seen the full, open text, it may be harder for them to pick up the source when they’re in and on the ground.

"How do you make the connection between the clues in a scene, and PC's skills/aspects? In other words, if you are not relying on skill lists, how do you preserve player unique-ness - that sense of competence or expertise in a specific field?"
That’s a tough one, for the most part I trust my players to find their niche and support it. For Bubblegumshoe I keep the character spreadsheet open, just to look from time to time to see what they have that hasn’t gotten hit. I tried to do that with Trail of Cthulhu at the table, but honestly the skill list’s too large.
Generally Gumshoe’s easy because the players know their skills, but I do try to find things which haven’t been hit yet (like Fashion) to present clues with. As I’ve run ToC I’ve gotten less and less worried about hitting all the skills or at least planning for that. Gumshoe modules make things look more intense than they need to be. They made me super cautious the first couple of times I ran modules. Now I realize I don’t have to keep my eyes on all of the details in the scenario.
As I've mentioned, when I run Fate, I use the Discover action. That’s pretty different from Gumshoe and I handle it differently than most Fate skills. When players go to Discover, I have them roll against a modest difficulty. Regardless of the success, I give them basic info if they have some skill in the area. Failures incur costs: time, attention, destruction of evidence, embarrassment, etc. But otherwise Discover’s like Create Advantage with each invoke being a question instead. (This has the benefit of allowing you to create some new Stunts related to Discovery like extra questions for an area or reduced time).
Any skill can be a Discovery skill and I try to make sure we don’t just fall back onto a “Lore" or “Investigation" all the time. Use Social skills for interpersonal, Outdoors for tracking, etc. That’s my general approach to all Fate skills, with the right approach you can use any of them for any action.
I leave those questions open. Sometimes people will freeze so I’ll make suggestions. If they know PbtA I suggest they use those. At first I worried that players would just use that as a tool to cut through straight to the end. But that rarely happens, most players will self-limit based on the fiction. In the rare instance someone “over-reaches," I ask them to explain how they actually get that info—and then use that to bound what’s given. Regardless I ask them to describe their process since that’s an opportunity to bring in NPCs or locations.
I used to worry about that approach being too easy, but it isn’t. Some games have the dictum: Be Obvious and I think that holds true here. I sometimes forget mysteries aren’t just about a solution, but about how to deal with and prove it. If they get the right answer early, that’s cool because I can then complicate things, especially if they’ve tipped someone off.
It’s why Columbo’s such a wonderful TV show. Most of the time the action on the show isn’t about figuring out whodunit. It’s Columbo working out how to get enough evidence to pin them down. That side of things is at least as important as the “mystery" the players wrestle with. I love the moment when players figure things out and feel awesome, but then suddenly realize they have to actually implicate the criminal. Or in the case of most fantasy games figure out how to murder the bad guy.
Of course all of this assume a game with at least a sketched out "mystery" plot. Jason's talked before about using Cthulhu Dark's mystery building tool for structured improv. It can be challenging, and it relies a lot on Trust with a capital T (something I'll talk about in my next post).
For the full backlog of Age of Ravens posts on Blogger see here.