GDC Austin: Storytelling in Marvel Ultimate Alliance 2

Marvel Ultimate Alliance 2 marks a new direction in narrative integration and team structure for Vicarious Visions -- lead writer Evan Skolnick and narrative designer Jonathan Mintz discussed its genesis at GDC Austin's Writers Summit on Tuesday. Trying A New Structure The game, which features around 15,000 lines of dialogue (8,000 of which have voiceover) and 55 minutes of cutscenes, required the work of a full-time, on-site narrative designer and lead writer, as well as two off-site contracting writers. The team wanted to create a game with an epic tale that served the core gameplay mechanics, was steeped in Marvel characters and lore, with a wide variety of those characters and plenty of fanservice for the hardcore comic audience that would be picking up the game -- yet still appeal to casual fans and allow them to enjoy the story. At the same time, the team, which did not develop the original Marvel Ultimate Alliance, had to build on its success while updating the franchise for current generation consoles -- the original title led on the prior generation. "In the last game, the story bent around the locations and we wanted to make sure we didn't do that," says Skolnick. They also aimed to create "a human story with dynamic characters... We wanted these characters to feel real and change over the course of the story." Realistic human speech that would sound good as VO was also a priority -- and having VO on the entire critical path was part of updating the game to current-gen standards. On the other hand, design choices, such as allowing players agency in their chosen alliances, conversation choices tied directly to character development, and character selections that matter to gameplay and story, were also top priority. While planning the story out, Skolnick and Mintz broke the audience down into three groups: "skippers", who just blow through the game; "dabblers", who stop and enjoy content if it's presented well, and "explorers", who will find everything the game has to offer. In any case, "satisfying critical path content with more depth for the player" was the goal. Adapting An Existing Work A decision was made to adapt the Marvel Civil War comics storyline, but it didn't serve the needs of the game narrative perfectly. "We needed a stronger buildup than the comics," says Mintz. The decision was made to marry it to the Secret War storyline, which "gave us a real Act I buildup. A lot of games get accused of -- your story gets condensed into -- Act I is your opening cutscene, and your game is Act II, and it's all wrapped up in the ending cutscene. We wanted to avoid all that." Limited details are revealed gradually over the first few acts; "the full story only comes out a few missions into the game," says Mintz. Starting the game in the ordinary world allowed for setup -- setup that's "simple, but deceptively simple, and it allowed us to ease you in." Major Marvel characters fans would expect to see in the game were not in the Civil War comics arc and had to be brought in. "Gameplay needed to trump a strict adaptation of the storyline. We knew the hardcore fanbase would [complain]. We had to make it clear that we know but these are decisions that we made, and we put these facts in the game in the form of dossiers," says Mintz. They also chose some characters, such as Cable, that could fit on either side of the story depending on your play style. From a production process, that was a success, but it also added dynamism to the story and character development for the player. In the comic, Captain America's role was ambiguous. "We needed to lay out early on that players who choose to side with Captain America are trying to convince the public to repeal the registration act," says Mintz. The Dialogue System In the last game, characters shared dialogue, but they wanted to have the characters sound like themselves this time. The scope of writing 24 sets of dialogue (one for each playable character) would be too much, however, so the player characters were sorted into groups of eight character types. The trickiest part, says Skolnick, was the NPC VO response to player-chosen dialogue. Despite the fact that the responses were static and recorded, different hero dialogue was written for each NPC response; there was a surprising amount of wiggle room to write that hero dialogue, and the writers could use it to tease out additional meaning and write for the characters. How Did The Process Work? Modeled on a structure from Ubisoft Montreal, the team put the narrative designer at a nexus of all disciplines. He works with the lead writer, who manages the outside writers, but the LW inside could work with others at key moments. The narrative designer "serve as the narrative champion of the team," says Mintz, "to make sure the developers respect the story and that the story respects their design as well." His tasks included coordinating with the director, tools team, and mission designers; he wrote first draft mission dialogue and data, set up VO recording, managed the narrative team, implemented core narrative content, coordinated production of all text. He also had to make sure to keep the lead writer involved, and worked with the cutscene contractors -- and, finally, supported the audio team with VO integration. The lead writer, on the other hand, "set and [was the] watchdog [for] the narrative tone -- everything that went into the narrative had to go by me," says Skolnick. He wrote all cutscenes, conversations, and scripted VO. He also provided feedback on all story materials, from storyboards to marketing materials. He also selected the right VO takes, organized and submitted narrative materials to Marvel, managed the external contract writers, edited and polished narrative content, and weigh in on big-picture narrative issues as they arose. One advantage, says Skolnick, of having both a lead writer and a narrative designer on-site, "is that with the narrative designer not having to write as much, he could stay on top of the day-to-day, whereas I didn't have to stay in meetings from day to day and could spend more time writing. But because I was there I could be brought in at any time. Great close collaboration was possible at any time, on demand." On the other hand, the team was not always sure whom to speak to about narrative issues, and pressure would lead to one or the other to make decisions without consulting him. Defining roles, then, is clearly key, the Skolnick and Mintz's opinion. "Everything that was on the critical narrative path and would be subjected to design changes," was kept in house, says Mintz. What was farmed out was text "that generally was safer to put in a higher risk area, things that weren't as crucial and would not be affected by game design changes, and theoretically could be cut entirely." Working With The Team "We felt the best solution to get what we needed was to plan around simple tools," says Mintz. The decision was made to use Excel for the game script, including dialogue, tech details, and design details in the document -- which was aimed at three audiences: mission designers, the lead writer, and the narrative designer. The team also used text-to-speech, which auto-generated WAV files of robotic dialogue; this made it easy to know what files had to be replaced, and gave everyone an idea of mission flow early on. "It allowed the designers to pace the encounters," says Mintz.
 Skolnick recognizes, however, that "no game writing plan survives contact with the dev team." The decision was to map out the narrative and decide what was critical; some missions were marked as inessential and could be cut. Says Mintz, "Scope hits and you start cutting things down, but you drew that line in the sand that these are the critical things you need to hit. The core elements are still there. You've kept beating that drum and let the team know why they were important." Getting the team on board with the narrative is not an easy task, particularly as people are disinclined to read documents. Says Skolnick, "We need to dramatize the story to the entire team early and often. We did try this... It's difficult to get people to focus on that when they have their own things to deal with, but it was really important to get them to understand it, and how what they were doing related to it." This doesn't just create interest in the story element of the game, nor is it limited to making it seem relevant to the other developers. "Make sure you get everybody on board with your tone... It's really easy for people to go off course," says Skolnick. Says Mintz, "Communicating to the team, early on, we found flowcharts to be an effective way. Condensing the story -- 'Here's MUA2 in a minute!' 'Here's MUA2 in 10 minutes!' We got people to act out the cutscenes at team meetings." They also played animatics from the cutscene contractor, says Mintz. Mintz was "embedded with the mission team. When it came time to be that bridge between what happens in the mission and what needs to happen in the story, I was able to bring Evan in" to those discussions.