Dragon Age II, like its predecessor, is played on two levels. On the one hand, it's a party-based dungeon-crawler, full of stats and levels and weapons and upgrades and so forth. However, the more unique and compelling aspect of the game is its decision-making component: every interaction involves complex dialogue choices that let the affect how plots evolve down the road. There are hundreds of such interactions across the game, which means players end up making hundreds of choices.
"The guiding principle," says Gaider about how all these choices are handled, "is degrees of consequence. You can actually split choices up into several categories. Many are going to be flavor. You're asking the player to make a choice, but either there is no 'real' effect, or it's immediate. The player doesn't necessarily know that, however, and for them the fact they're being asked to decide something gives it weight.
"Other choices are going to be local in their effect; you see a repercussion after the choice is made, but it's confined to the plot or the region the choice occurs in. Then a smaller number of choices are global; they have effects down the line, affecting the critical path in small and large ways."
"What's important is that there's a balance," he continues. "Not necessarily balance in the number of choices, but more that you have a spread among the types of choices. Too many flavor, or local, choices, and the player begins to lose a feeling of agency."
The sum total is this: in the Dragon Age games, decisions matter. The player knows that not every choice will have a lasting impact, but any choice could have one. Combine this with a deep back story, a gritty world of thorny conflicts, and a cast of nuanced characters, and you get a game that, for better or for worse, develops different for each player, delivering a highly individualized gameplay experience.
The challenge, of course, lies in the implementation.
Once the concept for the sequel was ready to go, it was time to bring on a staff. Gaider led a five-person writing team (consisting of screenwriters, prose-writers, and modders) that would develop all the written material. That includes everything from dialogue trees to cinematics to background chatter to the hundreds of collectible codex pages that appear throughout the game.
Writing a game like this is a top-down process. "We start with what we call a one-pager," explains Gaider, "which lays out the flow of the plot in general, and then slowly begin expanding on that in more and more detail. We end up with a very detailed overview, break it up into segments, and split the work among writers who in turn break up those segments into workable sections. Any required assets (such as a codex or a description) will be identified as the work is ongoing, and eventually all this information simply accumulates."
Every plot line – whether it's the spine, a particular chapter, or a subplot – has what Gaider calls its own "narrative overview." The overview lists all the beats for that section of story, including the choices involved, the gameplay elements, and the resources needed – along with the required budget. Once approved, the narrative overview serves as a blueprint that can be shared by writers, combat designers, level artists, and so on.
"The trick is to keep track of it all," says Gaider. "We have a wiki on our internal network that houses our accumulated lore, and our in-house editors devote a fair amount of time to trying to keep it all organized and searchable. You get a lot of outdated legacy information in there, which can make it a real challenge to keep up-to-date as things inevitably change over time."
Ensuring the tone and consistency of what gets written is an endeavor unto itself. Each completed game section first gets sent to one of BioWare's in-house editors, whose job is to check dialogue for errors and consistency. These people see every line in the game, and end up being the de facto "masters of the lore" who point out when a word is misused or when information is incorrect or inconsistent.
Then comes the first review, which the team at BioWare calls "white box." "Nothing's recorded, models are placeholders, level art is all simple geometry," says Gaider. "All we're looking for is whether the broad strokes make sense. Is the pacing good? Are there enough choices? Does the player receive enough information to know what to do? From that review, I'll work with the writer (unless the writer is myself) to come up with a list of changes."
Further down the line, story segments come back for an "orange box" review. "Now we comment on more specific things," he says. "Does the story feel right? Is it exciting? Did the player get the chance to ask the questions they wanted to? Do specific lines of dialogue make sense? This is where we would start looking for things to add polish. Maybe a follower should comment on that strange statue they pass by, or here's this section of the level where nothing happens and we should add something there? Things like that."