Crafty: Corruption

Two schools of thought (used to) define adventure games. One school says, “an adventure game is a story whose conflicts have been translated into puzzles,” while the second believes, “an adventure game is a puzzle described in terms of a story.” The difference is significant.

If you look at “Corruption” through the eyes of the first school, you will see a vastly unfair and agonizingly difficult work of interactive fiction. The game cannot be finished, or even understood, without experience gained through player-character “death.” Much of the behavior required of the player character, like spying on his peers or breaking into his partner’s office, will, initially, seem unmotivated and paranoid until the player gets stuck in a few dead-ends first. What’s more, I can predict, a little smugly, that everyone will discover, just before he thinks he is about reach “Corruption’s” climax, that he neglected to do something or other at the start of the story, and must replay the entire game. For instance, I found out that I should have thoroughly searched the toilet sometime before the 15th move.

If that sounds irritating and tedious, this may not be a game for you. On the other hand, members of the second school of thought will find a mesmerizing, Chinese-puzzle-box of a game. “Corruption” is a giant riddle, and to decipher its meaning, you must play, and replay, each of its parts. Once the player has mapped out the movements of the non-player characters — who are deftly portrayed, and whose characterizations add much to the bitter, cynical atmosphere of this game — he will recognize a web of deceit and betrayal, and be able guide his character to paths that lead to a satisfying ending.

In short, “Corruption” is a well-written, bug-free puzzle fest, and the puzzles are strongly related to an interesting suspense story. Just remember to save early and save often.


Your two schools of thought highlight the significant paradigm shift of classic IF vs modern style. Historically, people wanted puzzle games, but today they want stories.

But in addition, people wanted new and different kinds of puzzles. Most classic IF involved fantasy worlds and key/lock puzzles or disguised versions thereof. But Corruption attempts, to some extent, to replace keys & locks by information. A fundamental driver of its MO, is “tell X about Y”. The game must also track you know about Y, before you can tell anyone else.

A frivolous example is the “joke”. At some point in the game you’re told a joke. it’s not very funny, but having heard the joke, you can then go about telling it to other people. People react differently and sometimes you can use this sort of thing to evaluate personality. Of course, the game has to also remember who has been told the joke as well!

Corruption manages to do some things well, but also things get a bit too complicated towards the end. Essentially the court case is a decision tree where you have to play the right evidence at the right time. This could have been a lot more fluid, although there were multiple paths. Trouble is in 64K you run out of space quickly!

The in-game characters had stack-back goal solving drivers. In other words, they had a stack of things to do. This stack would drive their movements and reactions. This was quite complicated for the time. They even had automatic routing algorithms for their movement around the game.

Character goals get complicated when the player can interfere. For example, you can “pause” a character from their actions by talking to them. Sometimes they’d break off anyway to get on with their duties (because otherwise you could mess up the plot!). Sometimes by telling someone something, you can put a new task into their “todo” list.

Another lark was where a character goes looking for a certain object and the player starts to move it about. If you relocate the object from its original position, the in-game character has to search for it. Then you pop up and drop it on the floor in a corridor and watch if the character walks past it or not. They don’t, they pick up the object. Same if, while they’re searching, you hand them the object. In this case, they have to move on to their next scripted duty recalibrated from their current position and situation.

The somewhat complicated mesh of events in Corruption meant the overall size in terms of location and object count was smaller than was usual. This made it look like a smaller game. In terms of events however, it was larger.


Is this similar to the Reactive Agent Planner that gets used in some TADS and Inform games?

I think role-playing games call this “coupon treasure,” where the player’s reward is getting information to progress in the game, instead of getting experience points.

An interesting trend I noticed in the last few years is that some Twine games are moving in the opposite direction. Twine started with great story-based games (with and some simulations like Horse Master), but has produced better and better puzzle games in the last few years, like
Chuk and the Arena
in the last IFComp.

In fact, in 2016, for the first time ever, we had a parser game take Best Writing and a Twine game take Best Puzzles.

1 Like

I would look a bit askance at this statement about parser IF in any case. Looking at the XYZZYs and elsewhere, there are plenty of puzzly games getting lots of attention; there are things like Arthur DiBianca’s restricted-parser puzzle games (Chandler Groover’s Eat Me as well), and Mike Spivey’s games, and lots of other things. I might also mention Cragne Manor.

What I think is different is that players are (I gather from a lot of comments) not overly patient with puzzles of the form described here, where you have to repeatedly discard your progress and start over, either because you need to discover something you didn’t know, or you realize that it was too late to solve a puzzle you needed to solve at the start of a story. Maybe Hadean Lands is instructive here; you need to restart all the time, but the game automates a huge proportion of getting back to the point where you were before (or where you need to try something else). When you don’t expect a game to be something that you lose yourself in for a long time, you may have less patience for games that make you spend a lot of time redoing your work.


An interesting point you make; the effort of redoing state.

I’ve been wondering for some time whether, given a really easy way to restore and manage game state, whether dead-end puzzles are so bad. Since if state is easy to restore, it’s just a philosophical difference whether you’re playing the same game “flow” or not.

I recently programmed an old text adventure I designed as a teenager, thirty years ago. It had plenty of sudden (but signposted) deaths and one way/points of no return. It was interesting to try and implement the ideas in a way that wasn’t as frustrating, but still maintained the original design intentions… One of the easiest things to do to reduce the frustration was to add in some auto-save checkpoints around all the areas of sudden death… The game state was automatically restored back a few moves where necessary; keeping the original feel of the game design but making it slightly more palatable to a modern audience.

The one way/points of no return were trickier to deal with but they were, in most instances, just either eliminated or accompanied by “nag messages” that checked the player had all the items they needed to proceed.

Games are written for the audiences of their time. Back then you expected that you would have to restart a game many, many times, because of death states, or to check you’d not missed anything. You definitely did not expect to be able to complete a game in a single, unbroken run.

Quite often dead ends, where you have to restore, are a fantastic way of incorporating comedy and all sorts of situations where you’d never be able to go back to normal otherwise;

A great many players like to try silly things in situations, like setting things on fire, breaking stuff and generally causing havoc when obviously the butler is never going to let you back in ever again!

Unless you restore a save game of course.

As mentioned, if undo or even auto-undo is made easy, then the idea of dead-end situations add enormously to gameplay and entertainment.