When someone talks about BioWare games, whether we’re talking about »Dragon Age« or »Mass Effect« or even earlier games such as »Knights of the Old Republic« and »Baldur’s Gate«, chances are they’re going to mention the party members. These are the characters that accompany the player on their adventures, the teammates with whom the player interacts the most, and could arguably be considered a hallmark feature of BioWare games that few other developers have attempted to replicate.

Why not? Aside from the fact that having story-based party members requires a game with a team-based combat mechanic, it should also be considered that this is a feature which is both content-heavy and incredible complex to construct. The average party member in a BioWare game is going to have upwards of 2,000 lines of recorded dialogue specific to that character … and that’s solely for the dialogue used in their personal interactions with the player. Dialogue used elsewhere in the game, such as their combat soundsets or any role the character plays in the main plot, adds to that number. This is all dialogue which is going to be recorded, and much of it accompanied by cinematic cutscenes, and thus makes for a very expensive proposition.

The average party member – like the dwarf Varric – has upwards of 2,000 lines of recorded dialogue solely used in the players’ personal interaction. Combat soundsets or main plot cutscenes add to that number.

How much to write?

The question that often gets asked, then, is how the writers at BioWare go about constructing all that dialogue. How do we create a relationship arc for each party member that offers the player agency in determining how that relationship unfolds?

The first issue we need to resolve is one of budget — specifically the word budget with which we need to work. Every line we write must be recorded and have cinematic work done on it, and thus has a cost associated with it. This cost is expressed both in a dollar amount, but also in time… as in »how much time do the downstream teams need to work on what we create?« We come up with a word budget which both reflects how much we can spend, but also how many lines we can write which the downstream teams can feasibly handle in the time they have before release.

This is going to be an overall budget, one which the writing team will divide up:

• the critical path: the amount of dialogue needed to create the main story
• side content: story which the player encounters during the game, but which is not part of the main story, such as minor quests or exploration content
• ambients: dialogue which brings the world to life, such as a conversation between two peasants overheard as the player passes by, but is not part of a story
  party members

Splitting up the budget is going to be a lengthy process, one that requires each part of the game’s story to be concepted enough that its initial dialogue needs can be ascertained. Those numbers are going to be run by all team members for a gut check before they are sent up to project leads for approval. Once we have our final budget, the writers can proceed — each major chunk of work, such as a major plot or a party member, is assigned to an individual writer as the one who will not only do the initial design work and writing, but will be the point of contact for other team members working on that same section.

Constructing the Arc

So let’s say I’m a writer who’s been handed a party member to design. At this point in the development process, we very likely have a good idea of who this character is and how they fit into the bigger picture. That’s all part of the concept which was put together in the pre-production phase of the project (and done in conjunction with other departments such as Concept Art), but that doesn’t give much instruction on where to go next.

If the budget for this character’s personal arc is 2,000 lines, our first goal will be to split it up into how many full cinematic scenes that will support, and how much of it will go into »extra« dialogue. Traditionally, the player gets an opportunity to ask the party member questions about their background or their feelings on the ongoing story — these aren’t absolutely required, but can definitely add to the player’s sense of familiarity with the character outside of the major story beats. That »extra« dialogue is word-heavy, but lighter in cost compared to anything which requires cinematics.

We’ll simplify and say we’re going to split the budget up into half for the »extra« dialogue and half for the cinematic scenes. In Dragon Age: Inquisition, a short conversation is going to be about 50 lines. A long conversation can be anywhere up to 200 lines, and that can expand if it’s a dialogue which offers opportunities for questions or has branching paths within it. If we assume an average of 100 lines per cinematic conversation, that means I have room for 10 during the game.

But how do I turn those 10 conversations into an arc? First thing to do is to split them into three types:

• First conversation: We generally don’t need to worry about the specific party member’s introduction. That will be handled as part of the critical path — meaning the player will meet that character during the main story, and while the writer will generally write the portion of that which involves this party member, it’s not really a part of their arc. That doesn’t begin until the character has already joined the player’s team, and the player’s going to expect that, once that’s done, they get at least one full conversation the first time they go and speak with the character. This should kick off the »getting to know you« phase and lead into the rest of the character’s story arc.
• Reactivity conversations: These form the majority of the arc, and constitute the character »reacting« to certain major plot points or actions the player has taken. It’s a form of gating, to prevent the player from getting all the conversations in the arc one on top of the other, and provides dialogue off of which the player can develop their relationship with that party member. An example: the main story has an event happen, such as a war starting. When the player goes to speak to the party member later, the party member expresses their concerns over the war and the player’s part in it, and the player can either anger or please the party member with their responses. These lead into …
• Approval conversations: This is another form of gating. In Dragon Age, party members have an »approval« integer which is tracked, and which tells us how they feel about the player. The player does something they like? The integer goes up. The player does something they hate? The integer goes down. We’ll mark certain levels of this integer as »milestones« which will prompt a recognition of the character’s growing friendship or hatred (or even romance) towards the player. Since a lot of approval is going to come from the reactivity conversations, these conversations are going to inherently be gated by requiring the player to get through the main plot in order to »develop« their relationship.

Party members have an »approval« integer that tracks how the characters feel about the players’ actions.

Take a look at Chart 1. I’ve taken my ten conversations and split them up as follows: the first conversation, four reactive conversations that follow the main plot, and five approval conversations that respond to the player’s approval integer.

Chart 1: An example how conversations with a party member can be split up in different types and possibly progress over the course of the game.

You’ll note the grey circles on the chart—those could be considered possible »endings« to the player’s arc with that character. For the reactive dialogues, there will come a time when the story is over (or nearly so), and thus a final dialogue that follows, one which wraps up the player’s relationship to the character. Likewise there is a natural culmination to the approval conversations, a point at which the character admits to friendship or hate (these latter ones we call »crisis moments« where the party member could leave the player forever or even get into a fight), and that state is afterwards reflected in dialogue elsewhere.

So, from the player’s perspective, it proceeds like this: they make their way through the main story. They get conversations with the party member that recognize the story is progressing, and these conversations will drive approval up or down. Occasionally this approval will prompt the approval conversations, and provide us states we can reflect elsewhere in the story—if the player and the party member have had a talk about how they are now friends, for instance, we have a scripting hook to occasionally drop in lines reflecting that in the main plot or reactive dialogues that follow.

Chart 2: A specific party member plot changes the approval integer in a large way, depending on how the player completes it.

An added complication as depicted in Chart 2? In Dragon Age: Inquisition, we also have a quest that occurs specific to the party member, a plot which initiates about halfway through the main plot which is incredibly important to the party member (presented as a »could you help me out with this?«). Depending on how the player completes it, it changes the approval integer in a large way … and, furthermore, gates the »ending« for positive approval. Meaning that, in order for the player to get to the end of the positive approval track and have the party member consider them a »true friend«, they must not only get the approval integer to a certain milestone but must also complete the plot. This has the added benefit of allowing us to refer to the events of that plot in the ending conversation.

But what about romance? I’ve mentioned it before, and it’s one of the most well-known aspects of BioWare’s party members — characters that, in the course of befriending them, might begin a romance instead.

Chart 3: If the player gains enough positive approval from a party member, flirting becomes available with certain characters that can be romanced, depending on the player character.

If you look at Chart 3, you’ll see the point at the right at which the »romance track« would begin. If this is a party member who can be romanced, at some point in the buildup of positive approval, we allow the player to flirt with the character (or the character flirts with the player, and the player responds positively). This is, in essence, the player telling us: »I would be interested in romancing this character.« At this point, the normal positive approval conversations would divert to conversations specific to the romance. The player would still have to build up positive approval, but the »ending« of that track is a culmination scene for the romance arc — either a scene where the player’s character sleeps with the party member, professes their love, or otherwise indicates »this is now a serious relationship.«

But wait … with the two added charts, didn’t we just blow the budget? I now have thirteen conversations instead of just ten. Indeed, this happens all the time during development. We end up deciding to add on features, or add scenes as we feel them become necessary … that would mean revisiting our original budget and decide if it needs revision. Do we cut other scenes? Do we reduce the amount of »extra« conversation? Do we move the budget up, with the approval of the project leads?

In this case, knowing that a party member has a romance arc means the budget for that particular character would be a bit higher. The party member plot, meanwhile, would be accounted for by the word budget assigned to side content … which seems like a shell game, as we’re simply moving numbers from one category to the next, but that is indeed how we track our project allocations. It’s incredibly important to know where all our words are going, as not knowing means an inevitable bloat in all categories which — inevitably — just brings misery to those departments further down the pipeline.

In Conclusion

It’s not the only way to do this kind of character writing, of course — not even for BioWare. We tend to change things up from one project to the next, depending on the team culture and the things that team wants to try. The »Mass Effect« series, for instance, doesn’t use an approval integer or even truly have positive/negative approval conversations. Instead, it will use a larger number of reactive conversations and a more involved party member plot. In »Dragon Age II«, we had the romance arc being accessible off both the positive and negative approval tracks and had three separate side quests for each party member.

If a company other than BioWare were to do something similar, they also needn’t copy our exact method. The most important requirement is that it has to feel organic to the player. The player, after all, is not going to have any awareness of the delineation between the various types of conversations, or even the difference between a critical path conversation that happens to include the party member and something that’s specific to the party member’s relationship arc. To the player, it should seem as if the party member is actively responding to input … they remember how they feel, reply in kind, and later conversations seem to grow out of earlier ones. That’s what provides the player a feeling of agency, and what makes many players look on these relationships with fictional characters almost as if they’re real ones.

David Gaider