My experiences at GDC stirred up a lot of my own thought on game design, and what I think I know about it, and the directions that progressive game design are going. I'm talking about the functional-academic stuff you've heard of, like agency, emergence, intentionality, affordance; the kind of thought on game design that grew out of the Looking Glass school, and has been picked up by the guys at Ubisoft, if Friday at GDC was any indicator. The train of thought I'm on right now has come from the points made by Clint Hocking and Patrice Desilets at the conference, and reading I've done since by the Smiths, Doug Church, Matt LeBlanc, et al.

Here's the foundation of the progressive school of game design (as I've termed it,) as far as I understand: to create a single-player, character-focused video game is to place the player in a world which is populated by a variety of discrete affordances, each with its own distinct set of significant properties. Through his own judgement, the player deciphers the possible relationships between these affordances based on their properties, and forms intent based upon this range of perceived possibilities, defining his own hypothetical paths to the goals either imposed by the game's designer or created by the player. When the player follows through on this intent by manipulating the affordances in the world, he has acted upon his own intentionality; when a chance interaction between affordances results in a dynamic outcome unforseen by the player, they have experienced an emergent behavior, which may subsequently become an element of their strategy in approaching the gameworld. Gameplay becomes an extension of the scientific method. The player is placed into an unfamiliar space, and learns how best to navigate it by using observation, experimentation, trial and error. An effective game designer from the progressive school successfully creates and arranges the affordances in the world to provide the player with maximum functional agency to address the conflicts in the gameworld using his own judgement and creativity.

What does this translate to, in practice?

Ideally, it allows the player to address ingame obstacles in ways that the designer didn't specifically plan, and possibly never foresaw. In many games, the designer dictates both the "what" and the "how" (What: a locked door that blocks progression; How: the door requires the yellow keycard.) A successful progressive designer only dictates the "whats," which lead to a natural accumulation of "hows" (What: a locked door that blocks progression; weapons; enemies; robots; a keycard; lockpicks; special player abilities; bot control base station; How: the door may be unlocked with the yellow keycard, blown down with an explosion, picked with a special tool, or bypassed entirely by discovering a hidden route. Explosion option: An explosion of necessary magnitude can be created with a remote-detonation explosive charge, a proximity detector explosive charge, a rocket launcher, by detonating an explosive canister, or by detonating a damaged robot. Robot option: a robot can be detonated by draining its HP with straightforward attacks; hacking its control base and ordering it to self destruct; or setting it to a stunned state and hitting it with a single attack. Stun option: a robot can be put into stun state by hitting it with an EMP grenade, or by taking remote control of it and then exiting remote control, at which point it will be stunned briefly. Hereby, if the player has been observing the robot and door's properties, he knows that, while one bullet cannot destroy the door directly, he has the option of remotely controlling the bot, navigating it to the door, releasing its control, then hitting it with one bullet, causing it to detonate and breach the door. This is one of the classically cited examples of intentional play. On the other hand, maybe as the player is examining the door, he is discovered by the patrolling robot, engages it in combat, and happens to destroy it as it stands next to the door, breaching the obstacle and clearing his path. This is an example of an emergent occurence.)

It's worth mentioning that, in the above example, the player still has the option of approaching the stated obstacle the same way they would in the game that offered only a binary interaction. They can still search for the yellow keycard and use it to open the door. The hallmark of progressive design is the unspoken presence of any number of alternative approaches. And, as above, the most often-cited examples of progressive design are always extremely complex, and only the rarest of players will ever actually perform them unprompted. The beauty of the progressive design approach is that creating the conditions for such a solution is much less complex than the solution itself. The designer doesn't need to predict every possible solution and design it in specifically, but only populate the world with independent affordances that bear the necessary properties for that solution to occur. The key is to create only affordances which have significant relationships to one another and to the gameworld. Each affordance should be capable of affecting any other affordance it might come into contact with in at least one significant way; no affordance must be an island.

So, successful progressive game design presents the player with a series of challenges, and then provides them with a variety of pieces to build their own solutions, simple or complex, to these challenges. The solution can be as complex as turning a key in a lock, or simple as a Rube Goldberg device.

Character-based games on the opposite end of this spectrum--those linear, restrictive, or closed in nature-- are derided as too simple. Chris Crawford: "Explosions may be sexy, but as an intent, it turns the player into a trigger." But it's not the explosion, or the trigger that is the problem. It's what happens between those two concepts, or more accurately, what can happen, that is important. "To create an explosion" is in no way an invalid intent, and no matter how many elements lie between the player and the explosion, the player still fills the role of trigger. Like setting off a chain of dominoes or engaging a Goldberg device, the player is always the trigger; a successful progressive game design allows the player to choose and set up the dominoes himself, before knocking them over.

I think the major impediment to this school of design is the human factor. For a progressive game design's domino effect to work, all affordances must behave in a systematic, predictable way--a player can only employ his perceived agency effectively if all affordances perform in a manner consistent with his expectations. Since all games that have come from this school revolve around individual-level interactions between living characters, humans in the gameworld necessarily become affordances, and must also act in a systematic, predictable way. This obviously runs counter to our own lifetime of experiences interacting with other people; they don't all act the same, nor do they consistently react the same way to the same stimuli. I think that this objectifies humanity to serve the game's systems, to the detriment of the entire endeavor's sense of believeability. How can believeable humans--believeable in large part because of their unpredictable nature--exist in a system that demands they act in a predictable way? This is one of the primary conceits of video games now--that people have a limited set of possible reactions to a limited set of possible stimuli, and that their reaction to any given stimulus will always be reliably predictable. It's a limitation of AI technology, and of content creation--an AI can't be expected to respond believably and dynamically to every possible stimulus, nor can the game's development team be expected to include voice samples and animations to account for every possible situation. If the goal of progressive level design is to create a more dynamic, less linear, more natural, more believable gamespace for the player to inhabit, I think that the complexity of human behavior as a viable affordance is the next major obstacle. It's worth attempting to tackle, and many games have devised convincing ways to mask their AI's limitations, most often by restricting the scope of situations in which the player encounters other people to binary, life-or-death hostile conflicts, or by restricting the options that the player has in addressing other humans. Even within the most effectively crafted system of inanimate affordances, I wonder if there is any way to break out of the inherent artificiality of the game experience without actually bringing other, real human beings into the play. I wonder if the logical endpoint towards which the progressive school of game design strives is achievable in a single player setting. Again, not that it isn't worth striving for-- the steps along the way are very entertaining. But realistically, is this approach to game design a dead end? Or am I just extrapolating it out too far?

I guess that's the nature of the progressive school. It's the most interesting and rewarding segment of the current game design dialogue, but also perhaps the most problematic. It is the school that is essentially trying to work against the nature of single player games as they currently stand. They directly oppose the design philosophy of some of the games I love most, such as the Monolith titles I cited a couple posts down. That type of game, that aims to give the player a specific received experience, is a completely different beast-- largely linear, extremely focused, and essentially identical for each player who experiences it. It is almost entirely author-focused, which is an extremely conservative, if not outright outdated, design philosophy considering the current climate. Is it a contradiction to find them more enjoyable than their more progressive alternatives, then? Is it because they're "easier," conceptually? Is it because they're more movie-like, more passive, more directed than titles from the progressive school? I suppose my own preferences really says more about the player than the product.

If you ask me, the ongoing functional-academic dialogue among the progressive school is the most interesting avenue of game design theory going... Maybe I'm just not convinced that it's always the most fun. But perhaps it's precisely that oversimplified, assumed goal of "fun" that these games are trying to progress beyond.


Savid Daunders said...

Hi Steve,

This is probably the best description of "progressive" (as you aptly call it - I know of no better term) game design that I've read so far.

As you know, it's still an ongoing discussion, and will continue to be so for many years to come. But the questions most interesting are the ones with more than one answer.

As with the cutscene/no cutscene discussion, I can see two very divisive schools of thought forming around this subject. There will be those who aggressively adopt the emergent gameplay concept, and those that aggressively stay faithful to the tailored experience. Luckily, both have a place in the ever-growing market.

And of course there are already those that walk the line between the two, such as Bethesda.

In your opinion, how has the landscape changed since you wrote this in 2006, if at all? Clint Hocking and his team seem to be at the forefront WRT producing emergent gameplay, and it was met with mixed, tho generally favorable, results. Any other teams come to mind?

Anyway, keep up the good work!

David S.

p.s. If you haven't already, check out the in-development MMO Love (http://quelsolaar.com/). If you're into emergent gameplay, it'll blow your mind.

IQpierce said...

Great post.

Minor typo: "The solution can be as complex as turning a key in a lock, or simple as a Rube Goldberg device." Seems like it should be the other way around!

Matt Braddy said...

I think that that "typo" was intentional.

Loved this article. I've been having a difficult time figuring out how open world designers go about setting up their scenarios, and now the light bulb just switched on.

Thanks, Steve!