How to write a mystery?
- Thread starter 229
- Start date
Kaitou1412
Moderator
First and foremost: Game Mode. It generates a check list for physical evidence by itself with the boolean variables, can be used to measure relationships with suspects and witnesses to impact how forthcoming they are with the numerical variables, hiding the progress variables would make it possible to map the thoroughness of the investigation and interrogation without actually revealing whether or not all the answers have been found, and using if-then text to control how the results of the information present themselves. For the progress specifically, I'd say combining it with invisible boolean for critical clues and missed opportunities. For example, let's say one of the suspects saw something that proves the culprit's identity, but doesn't realize it immediately due the limited information it already has and also has a powerful greed. If we trigger the correct conversation, then we get the clue, triggering the boolean "true" and can use this testimony to nail the suspect - whether or not it's hidden is dependent on whether or not we too know the importance of the testimony. However, if we don't press hard enough or reveal the information sloppily, then the suspect pieces the truth together itself and clams up so that it can blackmail the culprit, thereby triggering a boolean "false" instead. This is just a basic premise for the moment, and it can expanded upon in greater detail later, but use this as the base outline.
Second, I see many short threads. Maybe a couple long ones, but some wouldn't be much more than a paragraph or two. It all depends on how much information we acquire at a given time. There's far more interactivity with this than a normal story. Typically, we can get away with saying "this is a logical extension, just a little dressy for the sake of character or plot development." In this case, however, "logical extension" is code for "giving away the game." The results of choices would have to be incredibly narrow. Similarly, I'd also say the setting at first has to be very drab and generic. Enough to know what's there to begin with (Ex. It's a room with a single bed pushed against a wall between two windows, with a TV in the corner between a bookshelf and a dresser, and a closet opposite the work desk with a computer on it), but not so much that we think we could get away with no hard looks. Generous descriptions (such as books on the shelves, things atop the dresser, state of the bed, etc. for my previous example) would have to wait until a certain area was set to be investigated. By consequence, threads get shorter and the total number of threads gets large very quickly.
That's what I've got so far. Feel free to weigh in on how effective some of that sounds and pitch your own ideas.
Best of luck, regardless.
Second, I see many short threads. Maybe a couple long ones, but some wouldn't be much more than a paragraph or two. It all depends on how much information we acquire at a given time. There's far more interactivity with this than a normal story. Typically, we can get away with saying "this is a logical extension, just a little dressy for the sake of character or plot development." In this case, however, "logical extension" is code for "giving away the game." The results of choices would have to be incredibly narrow. Similarly, I'd also say the setting at first has to be very drab and generic. Enough to know what's there to begin with (Ex. It's a room with a single bed pushed against a wall between two windows, with a TV in the corner between a bookshelf and a dresser, and a closet opposite the work desk with a computer on it), but not so much that we think we could get away with no hard looks. Generous descriptions (such as books on the shelves, things atop the dresser, state of the bed, etc. for my previous example) would have to wait until a certain area was set to be investigated. By consequence, threads get shorter and the total number of threads gets large very quickly.
That's what I've got so far. Feel free to weigh in on how effective some of that sounds and pitch your own ideas.
Best of luck, regardless.
LizardGod
Really Really Experienced
One way I could see it working is that you have a list of people you can interrogate. Which are represented by a thread. You ask them questions and go down the thread, these give you "clues" which unlock questions you can ask others or maybe opens up a new line of questioning with that person.
So you are purely learning things from conversations with the suspects with maybe the "Main" chapter that you come back to gaining new details about the crime as you talk to the people. You could then have a thread for accusing one of the suspects but you can only accuse those who you have some kind of reason to. So you couldn't just instantly go to that thread and accuse everyone to see who actually did it.
Off the top of my head, that's how I could see it working.
So you are purely learning things from conversations with the suspects with maybe the "Main" chapter that you come back to gaining new details about the crime as you talk to the people. You could then have a thread for accusing one of the suspects but you can only accuse those who you have some kind of reason to. So you couldn't just instantly go to that thread and accuse everyone to see who actually did it.
Off the top of my head, that's how I could see it working.
Wow, great replies and ideas! I should have mentioned in my original post that I'm thinking of trying it out with a more silly scenario, although I don't have anything concrete yet.
Kaito 1412, I've got to admit to not really understanding the second part of your post.
Lizisbest, you're right, there does need to be a way to force players to not just guess until they get it. I don't think having it unlock them based on which clues you had though because then every time you unlocked enough clues to accuse someone you would try it.
Kaito 1412, I've got to admit to not really understanding the second part of your post.
Lizisbest, you're right, there does need to be a way to force players to not just guess until they get it. I don't think having it unlock them based on which clues you had though because then every time you unlocked enough clues to accuse someone you would try it.
gene.sis
CHYOA Guru
You could unlock accusing of all suspects after an amount of time.
Maybe the protagonist could accuse someone when they have a number of clues...
Though, if the clues don't include enough evidence, the protagonist can't solve the mystery. Then, getting more clues pointing to that suspect could be more difficult.
After being able to accuse suspects, there could also be another time limit to accuse the suspect. If the protagonist isn't fast enough, the suspect could destroy evidence, flee or do something else to make it a bad end.
LizardGod
Really Really Experienced
Well if you decide which clues are the ones that you think would be enough for someone to accuse someone. So instead of it just being "if X number of clueY have been collected then show this chapter" rather being "if clueX && clueY are true then show this chapter"
You do still have the trouble of picking which clues fit into that. There have been more games that have tried to do a mystery than I can count across RPGs, board games, video games and even VHS games. I would say that the best thing to do would be to maybe look at a couple of those for some ideas.
I also wouldn't worry too much about tone either, that doesn't matter as much as how you actually make the game work. I would also just say that you should most likely not do it in a "adhoc" way, by which I mean it would be best if you wait until you have the whole thing ready and done before making it public.
I don't like the idea of the time limit, I'd like to encourage players to find as many clues as possible and put them all together.
What I mean is that if you pick up a clue, and see that the option to Accuse Bob appears, most people will click on it just to see what happens, and if they're wrong, they can go back.
Do you have any suggestions for mystery games that you think did it well?
LizardGod
Really Really Experienced
off the top of my head there is the PC game Consuming Shadow. Which has three "suspects" you get clue that read "X is the Z of Y" and you use those to build up a picture of who the guilty party is. Although in this case "Guilty" means they are an elder god trying to take over the world.
L.A Noir used a system where you pick up clues from a crime scene and then use those to interrogate, someone, later. However, it was not a perfect system mostly because, to steal a joke, it when you accused someone they would say things like "I bet you can't prove I am married to Hitler" and all you would have to do is provide the Marriage certificate at that point.
There is also the game The Vanishing of Ethan Carte that has an interesting take on the problem. In that, it gives a timeline of events and asks you to put things in the right order. Not sure how you could do that in text but it might be worth a look.
Thanks! I looked up all those games. I think Consuming Shadow's system would only work for stuff that's randomly generated though.
LA Noir's system would also be harder to do in this, for a few reasons. For instance, in the example you gave, you'd also have an option in the next chapters that would say "show marriage certificate" and then they'd know it's somewhere to find.
Obviously the Vanishing of Ethan Carte isn't something I could do in nearly the same way, but I could potentially have something where you hear a bunch of witness accounts of different events, and have to piece together the order.
LA Noir's system would also be harder to do in this, for a few reasons. For instance, in the example you gave, you'd also have an option in the next chapters that would say "show marriage certificate" and then they'd know it's somewhere to find.
Obviously the Vanishing of Ethan Carte isn't something I could do in nearly the same way, but I could potentially have something where you hear a bunch of witness accounts of different events, and have to piece together the order.
Nemo of Utopia
CHYOA Guru
To expand on what the others have said, all of which is accurate:
So, basically, you're trying to write THIS?
https://www.gamefaqs.com/ds/925589-phoenix-wright-ace-attorney/faqs
Or, to put it in terms that even a ten-year-old can grasp: (which is the target audience for your average mystery story.)
So, basically, you're trying to write THIS?
https://www.gamefaqs.com/ds/925589-phoenix-wright-ace-attorney/faqs
Or, to put it in terms that even a ten-year-old can grasp: (which is the target audience for your average mystery story.)
Last edited by a moderator:
Kaitou1412
Moderator
Well, probably the best example would be the Jisei series from sakevisual, in particular Kansei. By outward appearances of the house, there doesn't seem to be any suspicious objects lying around. However, investigating the boxes in the basement reveals documents pertaining to some of the victim's previous work, investigating a cabinet after we learn about it being in a blindspot in the cameras enables us to crack open the cabinet, and completing that reveals a second lock that needs a key and supplies a cryptic description of the hiding place, which in turn allows Kangai to notice a second blindspot in a different part of the house if we investigate the hiding spot. Furthermore, going back to my first point, Kangai can only unlock either the key or the documents, not both, depending on what he chooses to do when they need to override the office door's security. Additionally, there are five endings, all of which are based upon both the previously mentioned choice, as well as how he interacts with each suspect: if he's less than truthful with Sophia after choosing to reset the power himself, Sophia will confront the killer herself with the critical evidence; if he instead levels with her after making the same choice, she'll give him the critical evidence and allow him to identify the killer himself. Though it's the same killer every time and much of the actions overlap, the story isn't really linear, the background forces players to be more imaginative, and it gauges relationships. If it wasn't so heavy on witness testimony, placed more value in physical evidence, and also allowed players to use evidence by choice (rather than making it automatic), it'd be the perfect model - Jisei in particular suffered for these choices.
However, for how to handle the choices, Ace Attorney isn't a bad example either. Frequently, the player is forced to justify its choices, clarify what about our choices creates contradictions, balances the importance of evidence and testimony wonderfully, and even left plenty of red herrings (and Logic Chess from AAI2 is the perfect example of how to approach testimonial matters like my example from my first point). However, even this franchise has fallen flat here occasionally; the first example that sticks out is the game skipping over whether or not Moe failing to see white roses was important. Additionally, we can see fabric sticking out of a muffler or a shift in the paint on the floor if we're observant enough, so whether or not these locations are worth investigating is based on what we see. In my opinion, a purely verbal CYOA should be more heavily based on how deeply and imaginatively we're willing to think, and the details are only revealed if we choose to investigate more closely as is done with the Jisei series. Finally, and most critically, Ace Attorney is linear. Doesn't matter how many times we screw up, we still get the same information and the end results, we just have to double back more often if we make a wrong choice. There have been some bad endings, but these only pertain to single cases. The rest all progress the same way, and no amount of apparent incompetence is going to change the sequence of events.
Overall, I say blend sakevisual's use of unassuming backgrounds, relationship meters (though their meters are concealed, but this is ultimately to your discretion), and love of alternate endings with the Ace Attorney method of presenting evidence by choice, defending its importance, and constant red herrings - possibly even allow the player to build a strong case against the wrong suspect.
Doing this requires threads to become shorter. Less description until we ask for it, less answers unless we ponder the possibilities, and more need to defend one's choices. Short, straightforward, and no embellishments unless they ask for it. This is what my second point means.
I'm worried that it would sort of give things away. Like if you are in someone's room and one of the options is "look under the bed" you're going to do it. I'd like to make it so that the player has to actually figure things out. Which I'm beginning to think might mean focusing more on deductions than evidence.
Kaitou1412
Moderator
Well, you said no time limits. Both of my example games exist without time limits, and everyone ultimately investigates everything anyway. They do more heavily emphasize deductions and how conversations are handled than the actual investigations. However, that also results in a number of investigative options that go nowhere. For a simple example, Farewell, My Turnabout had Phoenix use a crude bug sweeper to figure out why his transceiver suddenly cut out, and the sheer number of things he had to look through before he found it was remarkable; of particular note was Phoenix disassembling a phone receiver with his only discovery being that he couldn't put it back together. The question of "How thorough are they willing to investigate?" is then dictated by their own time constraints in the real world. And in both series, these useless investigations are a prominent source of humor.
A time limit would change this. Utilizing progress variables and making it more like a Don't Escape game from scriptwelder (which are free on Newgrounds if you're interested) would balance it off. Sure, we could investigate the bed until it's completely disassembled, but would we be able to investigate any place else? Would we have to forfeit looking through the bookshelf for hidden switches or hollowed out books?
Without invoking a time limit, well, I circle back to numerical variables as relationship meters and if-then conditions for the best you could hope to do. If the relationships get too low, the if-then conditions give us lockdown answers, denying all opportunities to extract testimony, at which point the investigation grinds to a halt and we're pulled from the case. If a single certain relationship gets too low, we have attempts on our lives or characters choosing to work without us and creating new problems. We can also use booleans to lock out certain options once a certain threshold is crossed. The main deterrent in over indulging is the Game Over.
You can say a backtrack still circumvents all of it, but that's the basic hurdle of the "save, try, and reload if it fails" method all games come with. The only way to bar that is to either host a theater performance or become a real detective.
LizardGod
Really Really Experienced
I don't think it is possible to do that in a text-based format outside of having tons and tons of "Dead end" options for any given place but that still just means the player clicking through things.
I don't think there is really a way that you can stop players from just brute forcing the whole thing. However I would not worry about that kind of player too much.
gene.sis
CHYOA Guru
There is a suggestion to be able to hide unavailable chapters, but this won't help much at the moment.
Though you can avoid that the chapter answers give too much away. I explained that here.
If there is no downside of asking again and again, the reader would most likely do it until every piece of information is revealed. Possible downsides could be to gather wrong information (a witness gets nervous and fabricates information) or to make the suspect more circumspect. Or it could be the way you ask or investigate something. When searching for a secret drawer, being too careful would result in not finding it, being too forceful could lead the drawer to break and damage the evidence or result in trouble with the owner.
Kaitou1412
Moderator
229, I see what you mean now. Forgive my missing something so obvious.
gene.sis we're on different pages. You're right, the threads can make plot choices like that, but that's not what I mean. I was talking more specifically about leading the story. Basically, with a normal CHYOA story, it's not uncommon for authors to drop a few ideas before ending the thread; for example, someone mid-coitus with another man's wife about to be caught may first contemplate the distance to the closet, the fall out the window, and the gap under the bed before making a choice. For a normal story, no big deal since we can just add more choices later. For a mystery, we're leading the readers, implanting in their heads the idea that at least one of the listed options is correct. While this is okay for characters to do in dialogue, the narrator should be a little more impartial. Sticking with your drawer, we can mention the contents of the drawer, and then ask as a follow up, "Investigate more closely? Or investigate something else instead?" If we want to cut down on number of threads, we can make the options descriptive, such as "Investigate the drawer for a false bottom by measuring." and other such ideas.
No suggestions for the reader. We can screw then all we want with different, irreparable choices, but they have to make their choices without influence. That's all I meant. The logical extension of this is we have to draw back how much we describe and break threads up more frequently.
Let me put it this way. If a choice, in only one or two options, is obviously bad, (For example, if we have a time limit and investigating the drawer gives nothing) players can always just hit the back button. In order to be impartial, you'd need an absurd amount of chapters that go nowhere. For instance, there's a piece of evidence under the bed and in the desk drawer. Your options are to look at the window, the bookshelf, the bed, the floor, the light, the desk, the closet, and the dresser. A lot of those will lead to dead ends, and end up just wasting the player's time, as they click through every one to see if there's anything useful, and then go back when there isn't.
That's why right now, I'm thinking of letting the evidence stuff be fairly easy. You'll get some choice of where to go, and who to follow, but once you're in the right place at the right time, you'll see everything you need to. Maybe every now and then give you the option of making a deduction, like if you've gone to a man's house and found a picture of him marrying a brown haired woman, and then follow him and see him having sex with a blonde, you can hit "he's having an affair" or "his wife dyed her hair blonde" or "he must have remarried".
That's why right now, I'm thinking of letting the evidence stuff be fairly easy. You'll get some choice of where to go, and who to follow, but once you're in the right place at the right time, you'll see everything you need to. Maybe every now and then give you the option of making a deduction, like if you've gone to a man's house and found a picture of him marrying a brown haired woman, and then follow him and see him having sex with a blonde, you can hit "he's having an affair" or "his wife dyed her hair blonde" or "he must have remarried".
gene.sis
CHYOA Guru
Yes, indeed.
I thought that there is already a suggestion about that but I didn't found one. (only a line or two in the old suggestion thread)
So maybe you want to create a suggestion to let the story owner decide how many chapters a reader could go back (like 0, 1, 3, 5, 10)
With what you can use at the moment you could always give information about the investigation like "You find a torn off piece of paper with three phone numbers scribbled on it." which could be a good clue or is completely worthless.
Kaitou1412
Moderator
Even if we start talking about closing "previous thread" during game mode, all we replace it with is people who reset their game completely.
So now I'll open the discussion on contextual investigation and if-then parameters again. A simple example is a super early AO episode of Detective Conan: "Unexpected Visitors"/"Facial Mask Murder Case." Conan's only interested in why a woman expecting visitors is dressed for bed, but Ran calls attention to the fact that her facial mask should have cracks and wrinkles since the strangulation forced her mouth wide open. Now that Conan knows the facial mask could only have been applied after the victim died, suspicion fell to the victim's daughter since she was the only suspect who had both the opportunity and ability to greet the security officers at the door by using a facial mask on herself (along with a wrap in her hair, a bathrobe, and genetics) and convincing them that her mother was alive and well. However, the detectives didn't realize that this changed the window for the time of death, which was based entirely on the video footage of the "victim" greeting the officers at 9:10 and the third suspect's testimony of finding her dead at 9:20. Once Conan realized the victim could have died much earlier than 9:10, he deduced that the first suspect being an hour away at 9:41 no longer qualified as an alibi for the murder. Only then did Conan go looking for more evidence to discern which of these two suspects could have killed the victim.
CHYOA's if-then variable check enables pretty much the same line of work. Only by activating certain boolean's do we find the means to enable certain investigations. This could even theoretically change what I have to say about short threads and limited leading. Still some small ones as build up to ensure people don't get a variable without earning it, but this would eliminate people casually backpedaling since they're going to be wrapping around the same threads over and over anyway.
