You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I think that it would be neat to use the mechanics of the board game Clue/Cluedo to try to generate the plot of a murder mystery.
A rough outline of what I'm planning:
The program starts by randomly selecting a solution composed of three parts (probably whodunnit, with what, and where, like in the game) and stores that information for later
Each of 6 detectives/suspects move between locations semi-randomly, collecting clues which are tied to each location. Each clue eliminates one possible person, weapon, or place from consideration for that detective. (the clues being tied to locations are not really a part of the board game but I think this will make for a better/more coherent story)
If detectives happen to be in the same room on the same turn, they can gain a clue through dialogue with each other (the board game's mechanic of "suggestions")
Once one detective has eliminated enough possibilities that they know the solution, they accuse and the solution is presented to all the other detectives
I'm starting by getting the bare bones of those mechanics in place... Once I have some code that can generate the basic structure of the plot, I'm going to write Tracery grammars to embellish the story and make it an actual narrative rather than just a sequence of events. This entry is likely going to be pretty heavy on manual writing of details and descriptions, which the Tracery grammars will then plug into appropriate scenes based on the randomly determined movements of the detectives. That's the plan, anyway! We'll see how far I get. 🙂
Today I set up a repo with some of the logistical stuff: a Github action which runs a python script that will generate the novel text, and then publish it as an html page for easy viewing. So far the only text in the novel is the solution:
The text was updated successfully, but these errors were encountered:
With remarkable equanimity, Mr Green stepped into the billiard room.
The room was empty.
Mr Green called Miss Scarlett into the billiard room.
Mr Green said, "Miss Scarlett, the results are everywhere apparent: it was Mrs White in the billiard room with the wrench!"
Irritably self-conscious, Miss Scarlett answered, "You are certainly not making a reputation for accuracy by some of your statements. Obviously the murder weapon isn't the wrench. The cause of death was not a blunt object."
"Oh," Mr Green said. "Owing to a foolish error, I was mistaken."
I started out trying to write something much more elaborate with an original plot and characters, then realized a couple days ago that there was no way I could finish that by the end of November. :) So instead I stuck pretty closely to the board game mechanics and plot. One notable change is that I couldn't get close to 50,000 words if the detectives played the game skillfully. If they actually made sensible choices about what to suggest, the novel couldn't get much past 10,000 words before someone would figure out the solution. So as a cheap trick to meet the word count, I have given each detective an 80% chance of forgetting each clue the moment they hear it. Hence the title, "The Forgotten Clue."
To add some flavor to the text, I borrowed heavily from Putnam's Phrase Book, a 1919 book which compiles "Over 100 Model Social Letters and 6000 of the World’s Best English Phrases".
I think that it would be neat to use the mechanics of the board game Clue/Cluedo to try to generate the plot of a murder mystery.
A rough outline of what I'm planning:
I'm starting by getting the bare bones of those mechanics in place... Once I have some code that can generate the basic structure of the plot, I'm going to write Tracery grammars to embellish the story and make it an actual narrative rather than just a sequence of events. This entry is likely going to be pretty heavy on manual writing of details and descriptions, which the Tracery grammars will then plug into appropriate scenes based on the randomly determined movements of the detectives. That's the plan, anyway! We'll see how far I get. 🙂
Today I set up a repo with some of the logistical stuff: a Github action which runs a python script that will generate the novel text, and then publish it as an html page for easy viewing. So far the only text in the novel is the solution:
The text was updated successfully, but these errors were encountered: