Wednesday, February 23, 2011

Not Bringing Studio 2 Into This

I spoke with Adam last night. Decided after much deliberation not to bring his Studio 2 into this, convenient a source of helpers though it may be. I think he and his Studio 2 class would be better served by coming up with their own project rather than piggybacking on my thesis. The trouble is that Studio 2 class tends to have a very limited scope--not that the scope of any of the projects I'm proposing is that daunting compared with most things I could think to do, but there's a certain problem in the way this type of game is broken up versus the way another type of game is broken up. To give an example:

StarCaster, a pet project I work on in my spare time, a real-time sci-fi action-adventure that can be described as Metal Gear Solid meets Final Fantasy meets Blade Runner, is a very ambitious project. It requires tons of content, a robust physics engine to drive its spell system, and necessarily Euphoria so that the AI can take advantage of the very heavily physics-based gameplay. It's meant to take around 20 hours to complete. This game would require a team of 100+ people two years to make and very likely would cost upwards of ten to fifteen million dollars to make. My estimate is around twenty million. This is a AAA production that I will probably not get to pitch for about five years yet.

La Tarasca is an adventure game with some action in it--it takes place in the Old West, after all, it'd be a disappointment if the player didn't get to pick up a revolver. Overall most of the interactions are dealing with characters, engaging in dialogue, and gathering clues, which are relatively simple things to implement and don't even necessarily require the full power of the Unreal Engine. It's meant to take a comparatively smaller scope, focused around one location and a fairly small set of characters--namely the residents of the town of La Tarasca. In terms of play time I'd estimate the game to take around four or five hours; scaled back to a student project the full mystery could take one to three. This game could be made on a very small budget of under a million dollars in about six months, making it suitable for independent development.

Now here's the bomb. It would be more reasonable for me to try to get that Studio 2 team to create one playable level of StarCaster than to try and create one level of La Tarasca--the reality of a Studio 2 class being that there's only ten weeks to develop an entire project, that everybody is splitting their attention necessarily between multiple classes, and that technical problems would present huge roadblocks; even if our artists can generate the assets for more than one level, we'd only be able to generate one level of actual interactive content, and maybe the AI for one enemy type.

The mechanics of StarCaster are very focused and relatively straightforward, being a combat-based game and all; the length estimation of the entire game comes mainly from sheer content. We could develop just enough guns to make it clear that there's more than one, have the player earn one spell to give them a taste for the magic system, and develop one level focused on this content. It wouldn't be as slick and polished as God of War III, by any means; Euphoria certainly wouldn't be implemented; but it would be a reasonably good demo with a relatively tight asset list. Probably even this isn't a reasonable expectation, but were this ONE LEVEL worth of content completed within the quarter it would be acceptable as a demo and would give both players and prospective employers alike a representative portion of the game.

The gameplay of La Tarasca is completely dependent on the story, however. If we were to generate comparable content to the single level of StarCaster that I outlined above then it would likely be tighter and more well-polished, but it also wouldn't be complete by any means. The demo would end just as the story is starting to ramp up and the intrigue is starting to set in. For the purposes of my thesis it would be more than acceptable as I intend to be working on it for much longer than just next quarter and it would at least get the groundwork laid, but it would not be representative of the entire product and likely not nearly enough to impress anyone.

In other words, La Tarasca overall is a much more doable project, but a portion of it would not be acceptable; as a story-driven experience, however short, it requires the entire story to be seen through to completion to provide a satisfying experience. This is likely to be the case with any game that I try to develop for this thesis, but its' especially true of La Tarasca, which is the game that we've been leaning towards.

I haven't written this plan off entirely just yet, but I've warned Adam that it's probably best to keep my thesis out of his Studio 2. We'll see if I can't use my television writing knowhow to come up with something more self-contained and episodic, but that's the demand this project would have--it would need to be a pilot, not a full game, and that rather distresses me as a pilot for a TV show doesn't necessarily see its themes through to completion.

Monday, February 21, 2011

Presentation Script

It occurred to me that I should probably post the script for my presentation here:

I will posit that by using thematic elements and psychological architecture to develop a consistent structure for creating choice-driven narratives, we can bring more focus and weight to nonlinear storytelling systems in games.

Narrative choice or nonlinear storytelling systems are becoming increasingly prevalent in the game industry. What was initially the purview of choose-your-own-adventure books and role-playing games has now spread through a variety of different genres, including brawlers and shooters--suggesting that our industry sees a lot of potential in this area of the field and that we are eager to explore it.

The reasons why are very clear. Games are spaces where we learn, develop skills, and explore our emotions in relative safety. Stories do much the same thing, allowing authors to explore themes pertaining to the human condition through dramatic structure. Bringing the two together ideally creates a more direct connection between the player and the themes of a story by allowing them to explore those themes and the relationships used to represent them firsthand; essentially, we substitute parables with experiential learning.

The truth is that narrative choice is handled with a great deal of inconsistency between companies and projects, with every studio developing their own scheme and tools for it,
and there's little in the way of guidelines for writers or content builders to refer to aside from a fistful of developer opinions regarding their current line of experiments.
In fact, traditionally writers themselves have very little to say about the content or subject matter of the narrative choice systems that they help build. They think of it more in terms of the control they have to give up to accommodate the player's sense of "authorship," as some put it.

More often it's the lead designers of a project that have the first and final say in the way that a story develops. There's a good reason for this. Gameplay flow has certain demands that require compromise as far as story structure goes--what kinds of challenges, goals, and rewards we want to see and the pacing between those elements. In exchange, the designer's perspective and its risk-and-reward psychology gives us a lot of the tools we need to make decisions meaningful and allow players to explore various styles of play and express themselves.

This utilitarian perspective isn't without its shortcomings, though. Systemic or metagame elements run the risk of overriding players own interests within the narrative and deny them their interpretation of the story. In the worst-case scenario, though, there just isn't anything to interpret because game content designers simply approach the generation of narrative from a point of view divorced from thematic storytelling elements, being more concerned with the introduction of mechanics and challenges than their dramatic weight.
To give an example, one of the many introductions in Dragon Age: Origins casts players as the son of a noble house. In this scenario you run upstairs to deliver a message to your brother, then kill rats in the larder. That night a visiting noble whom you saw for all of two seconds and had no reason to suspect as a traitor stages a coup d'├ętat against your house and you have to fight your way out of the castle. The ingredients here are interesting--a struggle between two noble houses, political strife between families during a time of crisis--but there's little to no exploration of these thematic elements and therefore no dramatic weight to the events that the characters play out. It is a purely utilitarian scenario, constructed purely to introduce the player to the game's systems when it could allow players to explore a more focused story.

To quote Janet H. Murray, an imaginary world is little more than a costume trunk of empty avatars unless it has been called into being by an external author. Just as someone must write the rules for a game, someone must devise the central themes around which a story is focused and the central characters the story is about and an appealing identity for the player to assume. I will posit that players don't want authorship over the experience, as is often suggested, but rather that they want to interpret or explore an otherwise focused narrative--to personalize their experience, to borrow a term from Obsidian Interactive. They want to be posed questions, assume different roles than they usually do, and learn about themselves through the experience--to project a character upon themselves rather than project themselves upon a character.

Traditional narrative structure is thus quite relevant for developing choice-driven narrative content; themes drive a plot, its dramatic turns, and the struggles of its characters; they're what characters learn and where the interesting issues come from; and therefore it's themes that decision-making systems should focus themselves on allowing the player to explore.
The game designer's point of view gives us some models to build a relationship between the player and their participation--an otherwise unknown factor in dramatic structure. We thus have a foundation--a framework. We know what we want out of narrative choice systems: room for interpretation, and discernable changes in dramatic turns based on our interpretations. But, we're a bit short on the information necessary to build a full narrative choice system--just a goal.

This is in part due to the conflict of needs between these two disciplines, so we must look to a neutral party--one that dictates the terms of both for our answer, and that's human psychology.

Specifically, in our case, we are interested in the psychology of judgment and decision-making, which is broken down into two major architectures, each of which has its uses.
The first is rationality, which boils down to the risk-and-reward psychology that we as game designers are familiar with. It's defined by a person's preferences as to what variables they wish to maximize and minimize, the alternative choices apparent, how they expect those choices to turn out, and how risky each alternative is. It acknowledges that we tend to act on probability rather than certainty--hence the importance of risk, an element defined by a lot of factors that we simply don't know, don't understand, or can't predict. Most importantly, Rationality states that we construct the world in our mind as a decision tree, not unlike how we define a branching narrative--although it defines risk in much more abstract terms than we know in games, more in line with what we can work with as writers trying to establish dramatic tension and conflict between characters.

Rule-following, on the other hand, sees decision-making as being more a matter of fulfilling an identity; a concept often explored in role-playing; and it deconstructs this concept very thoroughly in terms of social context and experiential learning. It details several tools we use in everyday life to help understand these identities, which in turn are also valuable tools for guiding players into the identities we craft in our characters.

What we get from all this data we've been crunching is a way to break down individual decisions and their purpose in crafting a nonlinear drama--and thus we are able to develop building blocks by now re-examining the concept of the decision tree from the top down.
At the highest level, we look directly at the way the decision tree is laid out--either in dramatic or constructive events. Dramatic choices are places where the story changes direction in a dramatic way; divergent choices, in other words. Because of the technical constraints we face in how many of these we can allow--we can only paint divergence with a very broad brush--it doesn't allow for a high degree of personalization within the narrative on its own, but it presents clear opportunity for calling attention to dramatic turns and major overarching themes.

Constructive, or non-divergent choices, on the other hand, are extremely flexible, able to be arranged either sequentially along a branch or in a non-mutually-exclusive fashion with optional events. Whatever the case, these create the opportunity for experiential learning and identity-building, which provides context for the more major dramatic turns.

Looking a bit deeper, we next must examine the logic with which a given decision is made. A decision can be direct or indirect; IE, it can be a prompt spat directly at the player, or it can be determined from other actions that the computer reads and catalogues for the purpose of interpreting the player's character. A good example of this in action would be Silent Hill 2, where the game uses the percentage of time the player spends at low health to determine whether or not the player-character is suicidal or values his life.

And finally, we have the dramatic impact the decision itself makes on the narrative within the player's mind, which is where our psychological principles really come into play as we're now dealing with questions of the player's preferences. I break these down with respect to either immediate preference or preference with respect to the whole of the product as well as bias towards either rationality or interpretation. On the immediate scale, we're dealing with rational problems--clear in-game goals, obstacles, and information--or, conversely, just personal preferences; what type of ice cream does your character like, for instance. On the overarching scale, we're talking about wholly practical issues in the player's mind regarding the metagame, or else moral issues regarding overall outlook on the themes of the story.

Bringing all this together we are able to show, from a writer's perspective, clearly and efficiently, the building blocks and logic that makes up a decision tree with respect to dramatic principles. Pictured here is a single scene broken down in terms of the choices a player is able to make using these building blocks and the purpose each one serves in creating dramatic context for major decisions. This scene specifically details the player, as a major crime boss, meeting with a rival crime boss to discuss why he's suddenly horning in on the player's turf. The player is faced with a number of choices that allow them to build their characterization and context, then an overarching moral decision that brings the theme of mistrust to a head as the player must decide whether Fat Larry is too much of a threat to let live, or whether they feel comfortable trying to compete with him and don't want to risk open warfare over him.

This is phase 1 of my visual component, its purpose being to clearly demonstrate each type of narrative choice as I've detailed them--and I have a demo to showcase the scene, simple as it is.

Phase 2 will be a complete narrative at a length of about one hour, also constructed in UDK for the sake of flexibility. Preproduction will begin once my thesis is approved, will continue through March, and then I'll begin production on phase 2, which will continue until August, when I'll initiate a content freeze and focus on bug testing. November will mark the final focus test.

To conclude: Though the needs of different productions between companies will obviously vary, giving writers and designers building blocks with which to structure narrative decision-making systems in terms of dramatic events presents an opportunity to enhance players' experiences as they take part in meaningful, thematically-driven stories.

Thoughts on the Project

So far I've identified a fistful of potential projects in past works I've developed:

1 - "Maddox," a concept for a web sitcom I developed some time ago, as an interactive game. My worry about it is that it will inevitably wander too close to Sam & Max as it's very similar in tone and concept.

2 - A new project based somewhat on the concepts I'm analyzing in Deus Ex, based on neutralizing a terrorist threat or hostage situation. Depending on how this gets defined it may turn out too heavy on gameplay and too light on thematic elements.

3 - "La Tarasca," a story a friend and I developed for a game some time ago revolving around a fellow named Edward Jones, a city slicker who returns to his hometown of La Tarasca on the border of Texas and Mexico to solve the mystery of his sister's death. It's a historical western, one of the important aspects about the town being that nobody in it is who the player would want to have doing their job. The priest is also the town bartender, for instance, the town undertaker is the most exuberant, friendly man in town--and always looking for the player's business--and the telegraph operator is a nice old lady who hogs the telegraph to gossip with someone at the other end of the line. Nothing I pick is free from ambition, but I fear this one may be among the most ambitious projects I can pick. It's also the one I'm most excited about thus far.

The other unsolved question in all this is what platform I'm going to use. I've been leaning towards Unreal, but there's the chance that a ten-week deadline may be imposed on this project next quarter as my friend Adam Price has an interest in incorporating it into his Studio 2 and Unreal... tends to be a little unfriendly. It's the most powerful tool there is for what we're trying to do, but Unrealscript is a very obtuse language with little resources as opposed to Unity, which is extremely open but not as powerful. One way or another I'm going to need to devote a good portion of my time in the next several weeks toying with both of these to see what I can do with these.

Thesis Approved! ... Oh crap, now what?

Passed my review with flying colors. The only comment the board had was that my demo was very standard and didn't entirely show my principles, but again, I was only hoping to re-create the current norm with my building blocks to see if I was on the right track. Now begins the real work, as I have to find a project that's practical, doable, and narrative-driven, figure out a platform for it, organize a team, and make it.

Information Overload

Over the last couple of times I presented my thesis to the class things turned out pretty awful, with a lot of criticisms towards the information I'm presenting and whether I've even actually got a thesis. This time, however, I had the dead-opposite reaction: way too much information to be able to process, all of it very heavy and academic. I'll take this as meaning I'm on the right track--but I need to reorganize my slides a bit. Fortunately that's not such a problem; I can edit slides 'til the cows come home, and I can certainly find ways to cut down my script for time. What is a problem is my demo as it's still only partway to being ready to show. Let's cross our fingers and hope I can make this all come together in time.

Testing -- Meeting with Fat Larry

 Struggling to get my flow chart in here, but my demo for the presentation is nearly complete. I built it around a meeting between two mobsters. It's not Shakespeare, in fact by all rights it's very standard, but I felt it was worth seeing if I could re-create what developers are currently doing with the structure of different types of choices I've outlined.

The scenario features a number of constructive choices that build up to a major dramatic choice--IE, whether to kill Fat Larry or not--by developing context and letting the player make personal choices to develop/interpret their character. It doesn't look like much, but it should be effective enough for the purposes of my presentation.

Dramatic Re-Purposing

I've re-purposed the terms I've set down specifically with respect to building a dramatic narrative, re-defining them as follows:

Divergent Choice = Dramatic Choice
Non-Divergent Choice = Constructive Choice

Both are organized under "Priority."

Direct and Indirect remain as before. Both are organized under "Logic."

Rational/Personal/Practical/Moral are now organized under "Dramatic Impact." Rational/Personal is in the "Minor" or "Immediate" subcategory, concerned primarily with immediate decisions and problems, Practical/Moral is under the "Major" or "Long-Term" subcategory, concerned with overarching issues across the entire game, whether they be major thematic choices or metagame elements.

The most radical change is in the way I view divergence, as I'm now defining it specifically within the context of its purpose in building up to a dramatic turn; IE, a dramatic choice. This means of examining these types of choices, I think, is more effective than the utilitarian/psychiatric method I outlined before as they're defined in terms that are there specifically for writers to understand.

Thesis Statement Take 2!

"By using thematic elements and psychological architecture to develop a consistent structure for creating choice-driven narratives, we can bring more focus and weight to nonlinear storytelling systems in games."

Hashed out the wording a bit with John Thompson; suffice it to say that I feel a lot more confident about this than I do about the idea of focusing on some sub-element of my psychological research.

To explain where this whole thing is going...
First: I decided to focus the presentation purely on the development of a system that writers can use to develop choice-driven narratives--rendering my evaluation rubric, which I've given a ton of focus up until this point, practically irrelevant for the purpose of presenting my thesis to the board. This meant cutting out a lot of content and focusing much more strongly on the different types of choices as building blocks. This has always been the goal of my thesis, I've just been too distracted by the piles and piles of information I've had to dig up and regurgitate to get to it.

Second: I dug up an interview with Bioware lead writer David Gaider, and was surprised to find just how little control writers actually have over these games. The writing process for these games is essentially utilitarian in nature, with designers calling all the shots about what events go on and telling the writers to develop the dialog for them. There's a lot of really good reasons for this approach, as I know from experience. When I developed my design document for StarCaster I ran into a ton of problems trying to develop a satisfying gameplay flow and only quelled them when I approached it from a utilitarian point of view, basing locations and events off of where would be the most fun place to gain specific weapons and powers and whatnot rather than thematic elements--but I knew enough to be able to return later and fit the meaningful thematic elements to that structure I developed.

This is particularly important to narrative choice-driven games as thematic elements and dramatic questions, as I asserted earlier, are what the player is there to explore. They're what give the story weight, what the characters are there to learn, what seeds the most interesting conflicts, and therefore what the player most wants to interact with. It's no wonder I have such a difficult time getting into these stories when developers aren't thinking about that content at all, even when they're particularly strong at it; they're just adding it because they feel it's mandatory to put it in. 

This reflects my experiences with Dragon Age rather well, now that I come to think of it. The opening scenario slates the player as the son of a noble house. The player goes upstairs to deliver a message to his brother, then kills rats in the cellar; after that, there's an attack in the middle of the night as a visiting noble--whom the player met for all of two seconds and has no reason to suspect as a traitor--stages a coup d'├ętat. There's a lot of interesting elements here, certainly: rivalry between two noble houses during a desperate time, political strife, perhaps. It's almost the stuff of Shakespeare, but that's not what the player actually explores.

This, I find, illustrates the need for my system very effectively as we can build much more effective content by focusing on thematic elements and dramatic turns. More importantly, it gives me something I can use to focus my presentation.

Forget risk and narrative choice

After discussing my thesis with a few people and trying oh so hard to figure out a way to make "Risk and Narrative Choice" work for me, I decided to revert back to my old thesis and just try to find a different way to approach it; a different angle to look at the puzzle, if you will. Risk just didn't make sense. The idea of locking the player out of content or having elements of the story diverge randomly doesn't have a lot of clear benefits. It's interesting, but it clashes very directly with the concepts of character/theme interpretation that I've been developing, which have now somewhat become my philosophy.