Organizing a mystery with multiple investigatory routes

I don’t have a specific story or game in mind for this, I’m just interested in the community’s opinion.

Let’s say you want to create a mystery story. Something terrible has happened, there are suspects who harbor secrets, and the player is tasked with finding out whodunit. For the sake of simplicity, let’s say a man has been murdered in his mansion, and there are five other guests staying at the mansion who may have killed him.

I suppose it’s helpful to know how the story might end. How important is it that the mystery be solvable during investigation, potentially before the author is ready? Is telling a good story at the forefront, above allowing players the joy of that “aha” moment?

All too commonly, the author tries to be tricky and the solution is something unexpected (as expected). The player was the killer all along, or nobody killed him, or everybody killed him, or he’s not dead after all. Do you consider it a major problem to make players go through the motions when their investigations are mostly doomed from the start? Would this be received poorly?

I feel like implementation makes a big difference on that point. The way a game like An Act of Murder is set up means the murder needs to be solvable. Other methods might be able to sustain drama and make an unexpected reveal feel more natural.

How would you personally organize the gameplay in this scenario, and/or how do you think players would want it organized?

The way I see it, there are several major ways the scenario can play out:

You could leave it open for investigation like An Act of Murder - players can go anywhere at any time and question anyone. This is often going to be a less focused sort of experience, less ripe for scenes and drama, in my opinion. Players will expect to be able to call everybody into the dining room and solve the murder.

On the other side of things, you could make it a very directed experience. The investigation takes place across five days, and each day the player leaves their room and decides to investigate a specific character, making them the target of the day. Perhaps you can only enter that character’s room on that day, or only access certain parts of the house, or only find certain clues due to events of the day. Each section has a clear-cut goal. Vespers, while not a traditional mystery, plays out somewhat like this. This seems like a better set up for a more varied climax.

There are some good intermediate options as well. You could allow the players complete freedom to a point, but once they have seen enough evidence to be suspicious of a given character, an opportunity presents itself to investigate them more closely. Having entered Colonel Mustard’s guest room, you are constrained inside until you have learned something significant, and then you may proceed to the next day. A subtle way of allowing players to choose their own daily focus, but still feel some sense of freedom.

You could also simply present players with a list of characters and ask who they would like to investigate next. That sort of straightforward, “cross-them-off-the-list” mechanic seems like another where players would have a reasonable expectation that, should they choose correctly, they can solve the mystery early. I could also see some players appreciating this method over the above, foregoing the illusion of freedom and offering them a direct choice of where to look next.

It’s the classic question of simulation vs. narrative. Realistically, an investigator could stumble across the damning evidence 10 minutes after walking into the house, but as an author of fiction you’d rather keep the cat-and-mouse going for a while and might want to wrap it all up in an unconventional way. What do you prefer? Are there options besides the ones I’ve listed that are worth exploring?

There’s the iterative model. (Make It Good is probably a good example.) That is, the mystery is set up to be so difficult, and failure such a strong possibility, that winning on the first attempt is near-impossible. The player plays many sessions, building up clues and knowledge. By the time you reach the conclusion, the protagonist can behave as a hypercompetent Holmesian detective, going straight to the most important clues. (The focus is likely not to be about discovering the answer so much as doing something about it).

You can also make the focus be about something other than the mystery itself. This is possibly not really answering the question; but there are plenty of fictions with a mystery structure where the real interest lies in the relationships between the characters, or in the techniques the investigators use, or the wider setting, subject-matter and themes of the story. (People don’t read Nevada Barr or watch Bones because the mysteries themselves are clever.) In that sort of situation, you don’t need to work on making a fiendishly sophisticated mystery design; you need a narrative design with a mystery flavour.

I wouldn’t say that it’s inherently less ripe for scenes and drama, but it does make it much, much more work to accomplish, which will often be much the same thing. (There’s nothing to prevent you from hand-crafting plotted events and scenes in this model; you’re just going to have to hand-craft a lot more of them, many of which the player will miss.)

The big problem with the highly directed approach is that if the player works out the solution ahead of time, and they’re not allowed to go straight to the solution (rather than sticking with your script for the next three scenes), they’ll get very frustrated about it. They will tell you so, in no uncertain terms. So in that situation you need to manage your clue-delivery very well. More broadly, one of the biggest problems with mystery games is that what the protagonist/player knows and understands is really, really important, and the player needs to be able to communicate that information to the game and be acknowledged.

That’s true. Is this sort of game well-received? Despite how satisfying the solution can be, I thought many players were irritated by too much iteration, retreading ground they have already been through, sometimes skipping important things in their rush to pass the old content. The most recent iterative game I played is Lock & Key (finally) but it was very tightly designed to make the iteration easy and not too frustrating. I need to play Make it Good.

This is an interesting way to do it. You could even forego the mystery solving entirely, leaving it up to players to solve it on their own time with no way to resolve it within the story (“The investigator pauses on his way out the door and says to no one in particular, ‘Don’t worry, I won’t reveal your secret,’ and walks off into the darkness.”). You could also leave the mystery with no obvious solution, open to interpretation (which would probably annoy a number of people).

This is assuming the mystery is solvable ahead of time. Is it unfair to present a story as a mystery and make it impossible to work out before the climax? Again, I’m thinking along the lines of surprise endings. A well written game should be able to provide a satisfying enough conclusion that players should forgive “wasted” time. The journey should never be considered a waste anyway!

Another way you could look at it:
A game featuring a lot of clues that the player has to work out to solve the mystery vs. a game featuring a lot of puzzles that don’t give the player the necessary information to solve the mystery, but will nonetheless constitute progress and result in an eventual solution.

Of course any of these methods could be used make a worthwhile game, I’m just interested in what others see as most promising.

Iterative games are well-received when they’re very tightly plotted, yes. They are very difficult to do well. But then, IF mysteries in general are difficult. (An Act of Murder was well-received, in part, because it was a valiant effort at handling a very difficult design problem.)

It’s a very fine balancing act. For a mystery to be a mystery, the player needs to find or receive pieces of evidence over the course of the story. That evidence needs to be genuinely useful, but not give the game away too early. The problem is that the player and the author may end up with very different ideas about what that evidence means.

The player thinks they have sufficient evidence, but they’ve overlooked something. The letter doesn’t actually eliminate any suspects, because it’s a forgery. But the detective doesn’t realise that it’s a forgery, and therefore wants to arrest the widow as the only remaining suspect. The game needs to acknowledge this without ruining the plot: if it simply refuses to arrest the widow, the player will get frustrated.

What’s worse is when the player thinks they have sufficient evidence, and their guess is right. The author has been dropping clues about the murderer’s identity throughout the game; but they’ve hinted too heavily, and the player has a damn good idea about who it is. It might not be cast-iron evidence, but the player really thinks it’s good enough to arrest the guy. They’ll get extremely frustrated if the game forces them to play dumb and act as if they’re still unsure of the murderer.

Conversely, you can get in a situation where the game thinks there’s sufficient evidence, but the player hasn’t worked it out yet. Usually this is because the player hasn’t made the connection between two pieces of knowledge, or (just as likely) because the author hasn’t considered alternative interpretations of the evidence that are equally plausible. This leaves you two options; either the game just tells you what all the evidence means (which takes the fun out of it) or you let the player flounder (also not much fun). Death off the Cuff is a smart little joke about this sort of situation.

And you don’t have as much leeway to gloss over possibilities in an interactive format. A lot of TV mysteries, for instance, will ignore the possibility that the criminal had an accomplice (if they don’t have time to eliminate that possibility, or if it would ruin an otherwise clever-line of reasoning.) You can’t get away with that kind of thing if you need the player to work it out for themselves.

So if your design premise requires ‘impossible to solve before the end’, you need to be very, very sure that there really is no way on earth that the player could work it out, or think they’ve worked it out, but you still need to feed them clues throughout the game. Tough to do.

I’ll pitch in with a couple of ideas that I get when reading this.

You could let the player arrest her and then send him to court to testify, only to have the judge release her again because of lack of proof that the letter is authentic, or even have it proven a forgery. Then the player would be back to square one, through a natural chain of events, and would have to keep investigating from there. He could also be given an extra obstacle or two as some kind of penance or penalty for having been too trigger-happy with the dear old widow.

Maybe the above ploy of having the case thrown out of court would work even if they guessed right.

For what it’s worth, writing Make It Good I decided pretty quickly that any mystery has got to be iterative, and allow the player to fail, otherwise there can never be any true deduction. That means it also has to be really quick to restart and play out differently.


This old forge thread described a fantastic method for building a system that encourages and responds to players’ attempts to solve it. It may give you some good ideas, but I’m not sure whether the whole system would be possible to implement in a prebuilt form like IF, without (re)active human input.