The Search for Planet X[1] is a board game where players solve what amounts to a logic puzzle. The puzzle is generated procedurally and shared via seed amongst the players. The players then interact with an app to extract facts or constraints, keeping track of their own information as well as other players' queries in order to derive the game terminating[2] fact (the location of planet X). The game is, of course, deciding which new facts or constraints to learn.
I got that on Kickstarter, still haven't played it yet, though.
I _have_ played and enjoyed Alchemists, another board game that involves deduction, though with other complex mechanics involved, and pure deduction won't necessarily win you the game.
I think the critique of Return of the Obra Dinn is flawed. The game only seeks two pieces of information for every crew member, but in actuality, you need to crack out a paper and pencil and start "investigating" to get this information without brute forcing the game's mechanics. By asking less of the player it actually provides very rich deductive experience, unlike any other detective game I've experienced.
Wow this is like how in Genshin Impact, a character tells you to go find an item that you already have and when you talk to him again 5 seconds later, he says "woooow you found it after so long! you must have travelled far!".
I've been writing a simple 2d arcade rpg a few years ago. I created a simple graph language for creating quests and started to design the system for specifying arbitrary quest conditions.
It started from simple things like "has player killed a monster of type X", "does player have item Y in inventory", "was the actor X in place Y already", "have player talked to actor Z about Q".
But then I thought it would be cool to add metainformation. "Have actor X seen player killing the monster Y". "Have any actor seen player talking about Q". "is there a chain of actors meeting each other from the actor that seen player killing actor X to the quest giver".
I tried to implement this naively and it turns out the amount of information is astounding. To the point that I'm still amazed our brains can do this.
I was working on something just like this myself as a prototype for a kind of "immersive sim" detective game, and experienced the exact same thing. I still really like the idea and hope to return to it someday, but yeah, it's harder to model stuff like that than one would intuit.
Thanks I'll look into this. I wrote the game in js+canvas so my plan was to add some js prolog or datalog implementation later on for querying. Or just use in-memory sql.
The problem wasn't as much with the querying part as with the remembering part, there was just too much information that needed to be recorded every frame in case I would want to know this 1 fact half an hour later.
logtalk basically takes prolog and bolts on a smalltalk-style OO system and it is kinda batshit but also utterly beautiful and even if you don't use it I hope you'll be fascinated by reading up on it.
(also because it's built -atop- prolog you can potentially get it to run on a prolog-in-X implementation though I've not yet tried that ... though I'm now tempted to so I'm going to finish this comment before I nerd snipe myself)
I think you need the "Aha" moments to let you complete some task, or lead you to some place, or interact with some thing, whilst still hiding enough information to keep you curious. That way the moments of deduction are exciting because it lets you progress the story, and gets you closer to the final deduction.
For games with a grander theme like a whodunit, I think it's totally reasonable to have that just be revealed at the end, and if you got it right you can pat yourself on the back. Books and movies are no different, sometimes you figure it out before the book tells you what it was and you get to have a cheeky grin on your face when it turns out you were right, but that's all. If you really must gamify it, you could have your character accuse people, and if they get it wrong the cutscene shows who did it getting away or what have you.
I feel The Occupation meets the authors requirements. You play the role of a journalist trying to figure out what happened in a bureaucratic scandal, wandering through beautifully atmospheric old institutional buildings collecting documents and clues. There is a lot of freeform deduction and you will definitely not get everything on the first playthrough as some things you won't even realize you were looking for, unless you made those freeform deductions and knew to look for those clues or documents.
They do a great job of gamifying knowledge through meetings and interviews you have with different characters, and if you don't have the right bits of information then you simply don't have those questions to ask, and can't pick the right answers, and you get less information back out of the person.
Are there any good detective games based around interrogating a suspect where you're given a decent amount of freedom in what you can say/ask over fixed dialogue trees?
Her Story can be viewed as something like this in a limited form, where there's a satisfying enough reason for the limitations.
Return of the Obra Dinn was awesome. Like Her Story, the deduction goes on in your own head or on paper in contrast to e.g. an unnatural UI that shows a nodes/lines diagram of facts where you're joining facts together to prove you know what's going on.
Disco Elysium - lots of dialogue trees and people to interrogate. Saying every possible thing is usually a bad idea.
Orwell - not so much about direct interrogation, but you essentially play Big Brother and feed investigators information you find from their Internet activity. What information you pass along or withhold can drastically sway the outcomes.
Consortium - first-person game where you're investigating a murder on a plane. (kinda feels more like a spaceship from Star Trek.) There's some actiony shooting as well, but it can mostly be avoided and isn't the focus.
RPG games are terrible, invalid engineering. Insufferable garbage. They are about as fun as watching ads on 90s television, which is admittedly possibly better than staring at a brick wall. You will never have an RPG where it feels like you're playing the "Role", because what you are really doing is trying to walk in the right spot to trigger a switch. You will always see one way to "solve" a "problem" in any given situation that was just story-told to you, but the author preconceived a finite set of solutions (i.e., one) for you that can only be expressed by stepping in a certain coordinate bound. And this is just the tip of the iceberg concerning a real RPG. Most add all kinds of useless crap on top of this: like leveling systems, grinding, and purchasing things to make your weapon better (not even talking about microtransactions. I mean getting 100 gold to get item x to get item y to get a weapon you need to get to the next area). If a game needs to be so cluttered with emulating an open world with all kinds of details (in practice, these will all be thinly-veield consumeristic nonsense like getting a flower to use to craft or trade), it better be an Epic, and took multiple decades to create, not every single game a studio churns out every year. If it is not worth your time to build an actual interesting open world with an actual meaning (aside from consume product) behind the entities in said world, it's not worth my time to explore it; i will skip through everything, installing hack to skip forced dialog, etc, like every other game.
Granted, in Nethack you "Role Play" as a murderhobo, so it's not quite the narrative focus you are looking for, but if you want multiple ways to solve problems it has an incredible amount of player fuckery built in from decades of development.
Well, the logical nature of the computer is leaking through.
What you are complaining about is that there are discrete states in the game or better yet, that it's a huge mess of sharp edges rather than smooth curves like the real world.
>> I’ve been thinking recently about detective games, and in particular, the “deduction mechanic” — the gameplay system of linking “facts” together to produce a new and interesting conclusions that forward the story.
Pedant's corner: deriving new facts from known facts is logical abduction, not decudtion.
The way that Sherlock reasons is abductive reasoning, although he calls it deduction:
Probably the best (and only one) I've encountered is the Painscreek Killings - a little known gem that I've been pestering my friends to try out. The only system it uses is the old Unity style first-person "interact with item in environment to read note or pick up key" mechanic. However, the design is intricately laid out so that only by making a conclusion in your head, you know where next to go in the small, but sufficient open world town the game takes place in. I think more detective games would benefit by taking inspiration from that game.
32 comments
[ 3.1 ms ] story [ 88.4 ms ] threadIt turns out, you can -- and you can also write code that solves them.
I wrote about the experience here, using the "Cheryl's Birthday" brain teaser as inspiration to create a murder mystery:
https://github.com/shaungallagher/cheryls-murder/blob/master...
The code is a bit convoluted -- it was hastily written during a Hack Days event -- but it's fun to see the end result!
1. (2020) https://boardgamegeek.com/boardgame/279537/search-planet-x
2. Among other ways to win points within the game
I _have_ played and enjoyed Alchemists, another board game that involves deduction, though with other complex mechanics involved, and pure deduction won't necessarily win you the game.
> Obra Dinn achieves this by the simplest mechanism
It started from simple things like "has player killed a monster of type X", "does player have item Y in inventory", "was the actor X in place Y already", "have player talked to actor Z about Q".
But then I thought it would be cool to add metainformation. "Have actor X seen player killing the monster Y". "Have any actor seen player talking about Q". "is there a chain of actors meeting each other from the actor that seen player killing actor X to the quest giver".
I tried to implement this naively and it turns out the amount of information is astounding. To the point that I'm still amazed our brains can do this.
I had the same but for the morality/social system of an apocalyptic im-sim.
I did some notebook sketching and quickly progressed into capsnet territory so I shelved it :(
The problem wasn't as much with the querying part as with the remembering part, there was just too much information that needed to be recorded every frame in case I would want to know this 1 fact half an hour later.
(also because it's built -atop- prolog you can potentially get it to run on a prolog-in-X implementation though I've not yet tried that ... though I'm now tempted to so I'm going to finish this comment before I nerd snipe myself)
For games with a grander theme like a whodunit, I think it's totally reasonable to have that just be revealed at the end, and if you got it right you can pat yourself on the back. Books and movies are no different, sometimes you figure it out before the book tells you what it was and you get to have a cheeky grin on your face when it turns out you were right, but that's all. If you really must gamify it, you could have your character accuse people, and if they get it wrong the cutscene shows who did it getting away or what have you.
I feel The Occupation meets the authors requirements. You play the role of a journalist trying to figure out what happened in a bureaucratic scandal, wandering through beautifully atmospheric old institutional buildings collecting documents and clues. There is a lot of freeform deduction and you will definitely not get everything on the first playthrough as some things you won't even realize you were looking for, unless you made those freeform deductions and knew to look for those clues or documents.
They do a great job of gamifying knowledge through meetings and interviews you have with different characters, and if you don't have the right bits of information then you simply don't have those questions to ask, and can't pick the right answers, and you get less information back out of the person.
Her Story can be viewed as something like this in a limited form, where there's a satisfying enough reason for the limitations.
Return of the Obra Dinn was awesome. Like Her Story, the deduction goes on in your own head or on paper in contrast to e.g. an unnatural UI that shows a nodes/lines diagram of facts where you're joining facts together to prove you know what's going on.
Orwell - not so much about direct interrogation, but you essentially play Big Brother and feed investigators information you find from their Internet activity. What information you pass along or withhold can drastically sway the outcomes.
Consortium - first-person game where you're investigating a murder on a plane. (kinda feels more like a spaceship from Star Trek.) There's some actiony shooting as well, but it can mostly be avoided and isn't the focus.
Granted, in Nethack you "Role Play" as a murderhobo, so it's not quite the narrative focus you are looking for, but if you want multiple ways to solve problems it has an incredible amount of player fuckery built in from decades of development.
What you are complaining about is that there are discrete states in the game or better yet, that it's a huge mess of sharp edges rather than smooth curves like the real world.
Pedant's corner: deriving new facts from known facts is logical abduction, not decudtion.
The way that Sherlock reasons is abductive reasoning, although he calls it deduction:
https://en.wikipedia.org/wiki/Abductive_reasoning#A_Theory_o...
I guess the term never really caught on in the popular language.
Write the story backwards and read it forwards. You get the illusory effect with no effort.