[Project] Murder at Midnight
Gudday!
Started a new side-project a week ago. Thought I'd jinx it by announcing it. :) Here goes...
In 1990-ish there was a murder mystery game on the Amiga called "Murder". It was the normal Agatha Christie type setup: guests in a mansion, someone gets murdered, and for some odd reason you have to solve the murder before the police arrive by questioning guests, looking for clues, dusting fingerprints, etc. There were a set number of mansions to choose from, and murders were randomly generated.
Loved the heck out of this game, always thought I could do better, now trying to do so.
I provisionally named it "Murder at Midnight", and then found a 1939 film, as well as an old-timey radio show both with that name, and both with lapsed copyright. So I figured hey it's good material, and am going to be pilfering from those liberally too. Will this streak of originality never end ?!!
Evidently not! Another game I'm drawing inspiration from is the classic Murder on the Zinderneuf. Murder everywhere!
How it's different to the thing I am cloning:
- The original Murder had a fingerprinting system, but I've decided that mine should be more character-oriented. More Poirot, less CSI.
- I'm going to go for more complex motives, and more complex murder setups loosely based on common mystery tropes.
- Dynamic flavour text. The original Murder was quite flat. "Just the facts."
- Red herrings and liars. Murder's guests were mostly just a tad too honest.
- Additional random events. Perhaps in some cases a chance that the murderer could strike again.
What's working so far:
- Procedurally generated (with seed) guests and staff. I want any "random" murder setup to be 100% reproducable by supplying a seed and guest count.
- Modular mansion. I have a room prefab, and can generate a mansion out of a simple data object.
- Basic point & click walking mechanic.
- Guest placement at time of murder.
How it will work:
Guests are generated and scattered around the house at the time of the murder. The murder victim and murderer are set up with means, motive and opportunity ("MMO", lol). A history is then generated backwards from the time of murder to the time the party started, so that every guest can be questioned about what they saw and when. This same logic will then be used going forward from the time of the murder until the body is discovered. From here the game becomes realtime. The exact same logic that generated the history is now used to manage the guests' movement, except now you are interacting with them.
Time management
This game is an exercise in time management. By the time I get home in the evenings it's too late and I am too tired to get much done, so how I am managing is thus: I do my thinking/planning before I get home. On the train, in the car, in my head. So when I sit down to develop, I don't have to think. I'm also forcing myself to spend no more than 2 hours an evening on it. I set a specific goal for myself, small and manageable. For example, the last one was simply "Make sure random guest placement skips the room in which the murder takes place". (this turned out to be a 10 minute task, but hey whatever - it's done :P)
I'll post screenshots as soon as I have anything a bit less placeholder-ish.
Started a new side-project a week ago. Thought I'd jinx it by announcing it. :) Here goes...
In 1990-ish there was a murder mystery game on the Amiga called "Murder". It was the normal Agatha Christie type setup: guests in a mansion, someone gets murdered, and for some odd reason you have to solve the murder before the police arrive by questioning guests, looking for clues, dusting fingerprints, etc. There were a set number of mansions to choose from, and murders were randomly generated.
Loved the heck out of this game, always thought I could do better, now trying to do so.
I provisionally named it "Murder at Midnight", and then found a 1939 film, as well as an old-timey radio show both with that name, and both with lapsed copyright. So I figured hey it's good material, and am going to be pilfering from those liberally too. Will this streak of originality never end ?!!
Evidently not! Another game I'm drawing inspiration from is the classic Murder on the Zinderneuf. Murder everywhere!
How it's different to the thing I am cloning:
- The original Murder had a fingerprinting system, but I've decided that mine should be more character-oriented. More Poirot, less CSI.
- I'm going to go for more complex motives, and more complex murder setups loosely based on common mystery tropes.
- Dynamic flavour text. The original Murder was quite flat. "Just the facts."
- Red herrings and liars. Murder's guests were mostly just a tad too honest.
- Additional random events. Perhaps in some cases a chance that the murderer could strike again.
What's working so far:
- Procedurally generated (with seed) guests and staff. I want any "random" murder setup to be 100% reproducable by supplying a seed and guest count.
- Modular mansion. I have a room prefab, and can generate a mansion out of a simple data object.
- Basic point & click walking mechanic.
- Guest placement at time of murder.
How it will work:
Guests are generated and scattered around the house at the time of the murder. The murder victim and murderer are set up with means, motive and opportunity ("MMO", lol). A history is then generated backwards from the time of murder to the time the party started, so that every guest can be questioned about what they saw and when. This same logic will then be used going forward from the time of the murder until the body is discovered. From here the game becomes realtime. The exact same logic that generated the history is now used to manage the guests' movement, except now you are interacting with them.
Time management
This game is an exercise in time management. By the time I get home in the evenings it's too late and I am too tired to get much done, so how I am managing is thus: I do my thinking/planning before I get home. On the train, in the car, in my head. So when I sit down to develop, I don't have to think. I'm also forcing myself to spend no more than 2 hours an evening on it. I set a specific goal for myself, small and manageable. For example, the last one was simply "Make sure random guest placement skips the room in which the murder takes place". (this turned out to be a 10 minute task, but hey whatever - it's done :P)
I'll post screenshots as soon as I have anything a bit less placeholder-ish.
murder.jpg
1023 x 603 - 71K
Bc_vTSNCIAAT5DF.jpg
1024 x 576 - 22K
Comments
With real time systematic interaction? Whoa!
Looking forward to the game itself :D
Could add an interesting slant if you dropped in a hint of the supernatural into your plot...
Am I done with what I was busy with yet?
Yes:
What new task should I have? (leave, stand around, have smoke, find someone to chat to, etc.)
How long will this task take me?
Did I encounter any divergent incidents during this tick?
No:
Carry on
If you chat to an NPC the timer deciding how long their current task is will be disabled temporarily. And since all decisions are procedural, the sim will play out in precisely the same way every time, unless you butterfly-effect things by talking to people.
The NPC inner monologue sounds like a Scrum standup meeting. :/
I get it in theory but I'm sure it'll be awesomely complex :) Cooooool :)
This game sounds great! Best of luck!
Coming along nicely. At heart the game is basically Data Mining: The Game so most of the work thus far has been about generating a believable history of events. Since the same logic is used for the history generation as well as the real-time sim, most effort has gone into the latter.
Working now:
- Boredom sim - Guests will wander around between rooms, and there's place for all kinds of "what shall I do next?" logic
- Witness accounts - System records every detail witnessed by every guest, and you can click on them to query the full list
Taking a slight detour now to work on some UI/Graphics stuff. The look I want to go for is like these book covers: http://i.imgur.com/WzbqP1q.jpg
Your statement... " Perhaps in some cases a chance that the murderer could strike again." this was the most powerful sentence to me. A murder that happened is nowhere as interesting or terrifying as a murder that could still happen (especially if it impacts the investigation). The possibility of the murderer killing a vital witnesses making solving the murder harder appeals to me or maybe even the player if he/she messes up badly. I sincerely hope you explore that further. I love the randomness (replay-ability and unpredictability is a good thing) and backward generation of the story/history.
This is a great concept you got here.
I loved reading those"Choose your own adventure" books and would like to see a branching mechanic where there are multiple paths to success and failure. Where you only have yourself to blame for jumping to the wrong conclusions and letting the murderer get away. With appropriate gory animation adding salt to the wounds of the would-be sleuth of course ... :-)
http://www.rockpapershotgun.com/2014/02/14/noir-syndrome-is-a-procedurally-generated-murder-mystery/
Oh well - I'll just have to make sure this one is extra-awesome then! :P
Reminds me off Heavy Rain with some Bladerunner mixed in...
I see you guys are 99% done...Deadlines? Don't you love the whoosing noise they make as they fly over your head? :)