2015 Workgroup Topic Proposals

Computer horsepower and videogame logical correctness

PART ONE: Over the last decade, I’ve marveled at the increasing computing power we all have available, and I’ve asked many colleagues the question: How can we use this raw computing power in the service of game design? The answers have been dis-satisfying. Most seem confused by the question. Some proudly talk about nightly builds and backups. Others describe how art tools are used.

My mind immediately goes to GA, genetic algorithms, a powerful (and computationally expensive) way to find optimal solutions to specific problems. Unfortunately, GA requires a “fitness function”, and most of the “fitness functions” I can think of (in regards game design) determine something AESTHETIC, which is really the domain of the human mind.

PART TWO: Imagine you’re making an adventure game. In act 1, the player gets a flashlight. In act 2, the player finds a trashcan, which deletes anything the player drops into it. In act 3, there’s a dark room that requires the flashlight to traverse it. Thus, the player can accidentally choose to make an unwinnable situation for themselves.

This is a “bug” in the game design, and there are a thousand ways to fix it. But how do we avoid making the “bug” in the first place? Talk to game designers, they’ll tell you “that’s what testers are for!”. But testers can be expensive, especially for indie developers, and they aren’t a magic solution.
But also, since this problem is a “bug”, I could essentially be asking for a game language that has “prove-able correctness”, a computer science problem that has never been solved (for general cases). So, way too hard to solve.

FUSE THESE: With the new computer horsepower we have available, I suggest we revisit the idea of testing games “brute-force” (examining every possible input combination) for correctness. I also propose that we use text adventure games as a first step, because of the limited and discrete choices those games are known for.

So I’d like to spend time examining the innate structure of text adventure games, and building a representation of text adventure games that facilitates computer “brute force solving”. ¬†Understand that I’m not trying to “revive” text adventures/interactive fiction; I’m thinking about all videogames, really.

Leave a Reply