Hi
A while back i started investigating how to adapt logic puzzles to IF games. Here’s an excerpt of some notes i made. i also developed some tools.
Zebra Mysteries
These are some notes about making a murder mystery game based on logic puzzles like Zebra, but less “logic-like” and more fun as an interactive fiction game;
This is most definitely a “Cluedo” type game and is aimed at people who like that sort of logic puzzle (yes, there are these people).
This is an interesting idea from a developer point of view because this stuff can all be templated. Basically we can almost automate production of the mysteries - at least in a logic sense, graphics are another story!
Introduction
There are many interesting brain straining logic puzzles like this that work as a grid solver puzzle. Here, here and here are some more.
Most of these work by leveraging the fact that each feature is mutually exclusive, for example each person has a different motive or a different murder weapon etc. Then building sufficient negative information until eventually positive facts can be deduced.
The problem with this is that it does not make a proper mechanic for a general mystery, because things (like motive) aren’t all mutually exclusive.
Additionally, assuming you’re the detective, there will be some definite knowns;
- who was murdered
- the location of the murder
- the means of death (eg gunshot, drowning, poison etc.)
- the time of death (within a margin)
Clues
Rather than the clues being dry given statements, they can be ascertained from the game surroundings and text. For example, hair colour will be in the picture or in the textual descriptions.
Talking to people will reveal clues, but these will be mixed with irrelevant information, blind alleys and possibly red herrings.
Since these clues will have to be spotted, the logic nature of the puzzle can be simpler without making the game too easy, but also by making it more fun.
Something not leveraged by logic puzzles is, in games, the villain can be giving false information. However, assuming there is only one villain, any corroborated statement must be false (assume only 1 possible liar).
For the game, we will assume statements given by the characters are true (to the best of their knowledge) whenever they are not the villain. This can be reinforced in the text by saying that they are certain or confident etc.
Mutual Exclusion
Mutually exclusive facts, which are the basis of logic grid puzzles, do however make for interesting puzzles. Some facts naturally fit this model;
- people can’t be at two places at once.
- if everyone is in the house, they must be in a particular room.
- ownership of personal effects; broaches, rings, gloves etc.
Difficulty Level/Score
The game will have an over-determined set of clues. For example, suppose the case can be solved by 8 clues. Not all players will get that. So the game will have more than 8 clues. The player can keep finding clues until the solution becomes apparent. The number of excess clues used will derive the score.
Additionally, if the difficulty level is something that can be set by the player, the game can hide or mask extraneous clues.
Tools
It’s feasible to make solving tools for these kinds of puzzles. This is useful as part of the game design and authoring. A solver tool can be extended into an authoring tool. For example, by adding and subtracting clues and checking the solver tool derives a unique solution.
If the tool takes certain number of steps to derive the solution, this number can be adjusted by varying the clues. For example a positive piece of information could be replaced by a set of negative clues. This will increase the solver steps. This may be useful if the game is to be made longer or more complex. Alternatively, more direct clues could be added to simplify the solve process if required.