DETECTIVE GAMES - RISKING THE UNWINNABLE

Some thoughts on detective game mechanics.

  • Elder71
  • 05/26/2016 11:43 AM
  • 10176 views
PREAMBLE (SOME WAFFLING THOUGHTS)
(Scroll down to get to the point)

A longish story cut short:

In recent weeks and months, I've become obsessed with creating a detective game, one set amidst the dense rain and glowing neon of a cyberpunk future. It'll be a point'n'clicker, thrown together (prototype style) in VX Ace.

I've been researching the best ways to implement detective mechanics, and opinion seems divided. The recurring problems, as I see it, are streamlining and agency.

First: Agency
aka - The Go Everywhere, Do Everything Method

To truly capture the essence of role-playing as a crime-solver, the player should feel like they are the one pulling the threads together, that they are doing the deducing and the cleverness of solving for X (the culprit).

The problem raises its head when the game's internal logic comes into play. By that, I mean that the entire game is structured in some linear way to shepherd the player towards the correct answer. This leaves a player who isn't following the mystery or doing any solving with the classic fallback solution to any adventure-game puzzle: scroll the cursor over every pixel and click everything, ask every suspect every question and continue in this trial and error fashion until progression is triggered.

The game itself obviously has no way of knowing if the player solved the crime and knew who to talk to/where to look or if they stumbled on the right combination of actions by chance. Agency is lost.

Another problem with this shepherding is that your character, 'the hero', will often be found having done some deducing for you.

For eg.

1. You find a cigarette butt at a crime scene early on in the case and forget about it. In a different crime scene, you find a receipt, which you decide not to look at in that moment because you're still searching the crime scene. Eventually, you forget about it too.

2. A way down the line, after a variety of events have unfolded, you're stuck for a direction and have resorted (perhaps with, perhaps without shame :) ) to just talking to everyone in the hopes that something clicks.

3. You talk to a vendor at a market. Having done deduction that you did not, the hero privately, mentally puts the cigarette butt and the receipt together, realises that the suspect bought cigarettes from this stall and begins pressuring the vendor for a physical description. This happens just because both items were in your inventory - not because you happened to come up with a connection.

In this mode, you the player are just doing the hero's leg work while he, feeding on the game's internal understanding of its own clues and story, takes too much thinking out of your hands in order to keep things moving along. Of course, the player may have made that connection themselves, and may feel satisfied that the game shared their conclusion when speaking with the vendor. But all they have done, in this instance, is keep up with the game - not direct it. Agency is lost.

Second: Streamlining
Too much? Or too little?

A fairly blatant point:
A true detective sim would have to contain thousands of NPCs (each fully realised and uniquely motivated and organically inter-connected) and a vast number of locations to visit, and databases to search to fully and authentically actualise the detective experience in its truest form. Instead of the game handpicking a very narrow range of interesting NPCs to act as its cast, it would have a living population and, accordingly, 1000 ways to get lost, stuck and generally off track.

Apart from the staggering workload that such a game world would entail, it also wouldn't be fun. Real-life detective work isn't Hollywood-cool, and typically involves a lot of dull routine and paperwork. Some form of super-humanly hardcore detective game fan (read: masochist?) might find something to love in that, but that's not the sort of end-goal I'm talking about. I want to maintain agency, while streamlining the project and making it concise AND fun to play.


THE POINT (PULLING MY THREADS TOGETHER)
(Hello to those who pressed X to skip)

Belated disclaimer: I know that there are a ton of detective games out there who have tackled this problem in a ton of interesting ways (some with success, some without) but I'd now like to bear down on the focus of my project:

Those who remember Blade Runner (Westwood, 1997), will already have a general impression of the style and format I'm aiming for when it comes to exploration, interaction and combat. However, while BR is one of my favourite games of all time, it suffered from the problems outlined above - most notably, it was prone to completion via that 'click everything, go everywhere, talk to everyone' trial and error method.

And so, at long (long, long) last, here is a brief outline of the main mechanic I'm planning to use:

Keywords / Ask About
Clues/evidence are not automatically added to an in-game database. The player maintains their own notes and can input as many keywords for future reference as they wish.

NPCs have mood gauges, and will eventually shut down/become impatient/refuse to talk to you if you ask too many obviously irrelevant or nonsensical questions.

Here's the basic structure of the dialogue/questioning system:



The main benefit of this system is that it allows for the game to remain streamlined (not flabby and overloaded with NPCs and locations) but that it insists upon a degree of player comprehension. The limited patience, tolerance and/or willingness to cooperate of NPCs installs a harsh but, in my opinion, fair way of deterring spamming. At the cost of probably frustrating all but the most interested player, I think this system moves the project a step closer to really providing players a sense of being the brains (or lack of them...) behind the case.

(Here, for the sake of a little colour, is a preliminary mockup of how the conversation system will be arranged/presented. Far from finished, obvs, but might be of passing interest to anyone who's just wandering through. It's not a RPG Maker screenshot, either - it's all photoshopped for the time being.)



The case you're working on will involve a lot of clues, leads and red-herrings, so comprehension is doubly important because you'll need to maintain an insight into which pieces of the puzzle fit where - and which can be disregarded.

Ultimately, the most contentious issue will be this:

it will be possible to hit a permanent dead end.


This will be most likely if you try to get ahead by just asking questions at random, trying to strike lucky. It seems to me that if I want to make a game that focuses on investigation, I need to appeal to players who really want to RP a detective and who don't mind a near total absence of hand holding.

Second disclaimer: I accept, without reservation, that the success of this system relies first and foremost on my ability to write it and implement it well while keeping things interesting and solvable - the challenge becomes one of expression, and making sure the player can follow my trail of crumbs!

Anyhow, thoughts and feedback on this idea are heartily welcomed.

(Thanks!)


Posts

Pages: 1
I have a slightly different suggestion - it is not incompatible with what you suggest, but I feel it is a bit more elegant (as it does not force a dead end).

If you don't want the player to just find things at random, you can use combinatorial explosion.

Imagine an "insight crafting" system where you have to put multiple elements together, Clue-style ("Mr X" "bought" "cigarettes" "at the shop"), and if the recipe is successful, you obtain/unlock a new topic/action/something.

Even if you have only 5 characters, 5 actions, 10 objects and 4 places for the whole game, that gives you a thousand possibilities - more than any player should be willing to exhaust - while making the correct combination is obvious once you've inferred what was happening.

If there's more than a single template, or even almost freeform combination (X saw Y kill Z), exhaustive search is even less plausible. The only thing that a player might be able to brute force is one missing element when they're pretty sure of the rest of the formula, and that's really not so bad (even Sherlock sometimes proceeds by elimination!)
author=Hasvers
I have a slightly different suggestion - it is not incompatible with what you suggest, but I feel it is a bit more elegant (as it does not force a dead end).

If you don't want the player to just find things at random, you can use combinatorial explosion.

Imagine an "insight crafting" system where you have to put multiple elements together, Clue-style ("Mr X" "bought" "cigarettes" "at the shop"), and if the recipe is successful, you obtain/unlock a new topic/action/something.

Even if you have only 5 characters, 5 actions, 10 objects and 4 places for the whole game, that gives you a thousand possibilities - more than any player should be willing to exhaust - while making the correct combination is obvious once you've inferred what was happening.

If there's more than a single template, or even almost freeform combination (X saw Y kill Z), exhaustive search is even less plausible. The only thing that a player might be able to brute force is one missing element when they're pretty sure of the rest of the formula, and that's really not so bad (even Sherlock sometimes proceeds by elimination!)


It's a very interesting approach - I'm just stuck on exactly how I'd implement it. The formula system (insight crafting, a good name for it) works in theory - Mr X's name is a piece of in-game information - but where does the word 'bought' come from? Does the player write this in like a keyword? If so, would the game also have to recognise 'purchase' as a correct entry?

It gets unwieldy quite quickly as I've been finding out over and over again, but this insight crafting idea definitely has legs
You could a few generic actions that are always available for use in these formulas. Stuff that is sufficiently broad in meaning that it can be used in multiple contexts.
With stuff like get/take, give, see, say... you can cover a lot of ground ("mr X" "get" "cigarettes" "at the store"), and you avoid the problem with parser interactive fiction where you need to guess the verb the author was thinking about.

One way I'd visualize a system like this: you've got a box where you mix the recipe and you can drop in various elements from an inventory, like "subject: mr X", "action:give" , "object:cigarettes", "target: ms Y", "location:store", "time:yesterday" so that the order of the 'sentence' doesn't matter, and it's easy to add more or less elements (for instance location or time are not always part of the recipe). For elements that can occupy different roles you may need first to combine "mr X" with "subject" or "target" before throwing it into the mix.

If that turns out to be too complicated, you can also have a few templates where boxes have to be filled for each of these roles (a template for direct action, indirect action, action at a location, and so on...)
The first step in this project is building a prototype to entice my programmer friend into getting involved. I have a few ideas on how to implement these systems into VX Ace without any custom scripts, so we'll have to see how well it plays. I definitely like the insight crafting idea - I may very well poach it from you if you don't mind?

EDIT
I should add also that you've highlighted the #1 problem with this sort of project: avoiding situations where the player has a good theory but can't pursue it because it's not what I'm thinking of as the designer. The system, whichever it turns out to be, has to be free-flowing enough and dynamic enough to account for this.
Good luck on making the prototype, and please feel free to experiment with the idea! I've also been frustrated by how guided most detective games (or similarly insight-based games like Phoenix Wright) tend to be, so I'd be super happy to see what you come up with.
Now this is an article I can get behind! I'm working on a detective RPG of my own (although much simpler in scope when it comes to case-solving as I also have other elements of gameplay to incorporate) and have run into similar problems as those discussed. Your guys' conversation is similar to thoughts and concepts I've come up with so it seems I'm on a right track in that regard. Though I've recently felt I'm using too many red herrings (although this seems to be a side effect of being able to name and convict anyone as the criminal with enough evidence).

"Insight crafting" seems like the natural progression of old games like Clue and is definitely a prototype I'd enjoy seeing. And yeah, custom scripts aren't really necessary as far as I've seen in development in Ace(outside of mouse scripts and the like) since there are easy conditional branch and key item commands. Which should be more than enough to get something to show a prospective programmer.

Hope the project comes along in a useful way! I would love to see more authentic detective games. The genre needs some love.
author=dinkledaberry
Now this is an article I can get behind! I'm working on a detective RPG of my own (although much simpler in scope when it comes to case-solving as I also have other elements of gameplay to incorporate) and have run into similar problems as those discussed. Your guys' conversation is similar to thoughts and concepts I've come up with so it seems I'm on a right track in that regard. Though I've recently felt I'm using too many red herrings (although this seems to be a side effect of being able to name and convict anyone as the criminal with enough evidence).

"Insight crafting" seems like the natural progression of old games like Clue and is definitely a prototype I'd enjoy seeing. And yeah, custom scripts aren't really necessary as far as I've seen in development in Ace(outside of mouse scripts and the like) since there are easy conditional branch and key item commands. Which should be more than enough to get something to show a prospective programmer.

Hope the project comes along in a useful way! I would love to see more authentic detective games. The genre needs some love.


I'm discovering more and more than 'true' detective mechanics are a uniquely challenging project to tackle. From my preliminary research and planning, it seems (on the surface, at least) that no matter what route you go down, the relationship between the player's freedom to investigate and the game's internal logic invariably leads to hand holding.

I'm very excited to put these ideas into practice - I'm sure there are pitfalls and challenges around every corner. Does your game have a page yet? I'd like to track its progress. Maybe we could inspire one another?
author=Elder71
I'm discovering more and more than 'true' detective mechanics are a uniquely challenging project to tackle. From my preliminary research and planning, it seems (on the surface, at least) that no matter what route you go down, the relationship between the player's freedom to investigate and the game's internal logic invariably leads to hand holding.

I'm very excited to put these ideas into practice - I'm sure there are pitfalls and challenges around every corner. Does your game have a page yet? I'd like to track its progress. Maybe we could inspire one another?


Yeah, that's for sure. I'm starting to see why games like LA Noire added more vehicle and shootout gameplay during development. Hopefully, I can find a strong balance since I doubt my current project will become a "true" detective game.

I don't have a game page yet for my project yet, although I did release a few screenshots awhile back for the "Release Something" event they had during my spring semester just to get some feedback. I was waiting until I had a demo to make an actual game page. Never really been a fan of starting one without something to show and all that. Also, still trying to finalize a lot of gameplay options. I'll definitely let you know when I get the page up though.
author=dinkledaberry
author=Elder71
I'm discovering more and more than 'true' detective mechanics are a uniquely challenging project to tackle. From my preliminary research and planning, it seems (on the surface, at least) that no matter what route you go down, the relationship between the player's freedom to investigate and the game's internal logic invariably leads to hand holding.

I'm very excited to put these ideas into practice - I'm sure there are pitfalls and challenges around every corner. Does your game have a page yet? I'd like to track its progress. Maybe we could inspire one another?
Yeah, that's for sure. I'm starting to see why games like LA Noire added more vehicle and shootout gameplay during development. Hopefully, I can find a strong balance since I doubt my current project will become a "true" detective game.

I don't have a game page yet for my project yet, although I did release a few screenshots awhile back for the "Release Something" event they had during my spring semester just to get some feedback. I was waiting until I had a demo to make an actual game page. Never really been a fan of starting one without something to show and all that. Also, still trying to finalize a lot of gameplay options. I'll definitely let you know when I get the page up though.


Yeah, please do. I know what you mean about wanting to have something first - but it's difficult to hold back on the excitement factor, isn't it? I mean, when you're buzzed about the thing you're working on, it can be tough not to shout about it asap
The tricky thing about true detective stuff is that there's a discrepancy between the representation of the world that the (linearly written) story presupposes, and the one that is given to the player. There's a degree of guesswork that goes in translating what the story is trying to tell into something that the player can actually do (which is generally limited to walking around, or picking options from a menu). Handholding is just a way to limit the arbitrariness and painfulness of that guesswork for the player.

To avoid it, you must either tone down the story to fit the player's tools (e.g. narratives where it makes sense to just be walking around, like Gone Home or a horror game where you're being chased by a monster) or expand the tools to fit the story. The best is to end up somewhere in between: for instance, as in Clue, have a mechanism that can represent some actions, and limit the story to something that can be represented like that in a clear, unambiguous way.

The closer your mechanism for representing information is to how you actually designed the story, the more likely it will be to let the player express their own thoughts and insights in a productive way.
Pages: 1