Blood, Sweat, and Pixels

by Jason Schreier

  As Brennecke scrambled to produce a regular stream of Kickstarter updates, the team began to add stretch goals—features they’d add to the game if they hit certain funding thresholds. It became a tricky balancing act. There was no way to know how much each feature would cost before the game had entered production, so they simply had to estimate. Feargus Urquhart wanted to put a second major city in Eternity, for example, but how could he and his team possibly know how much time it would take to build a second city when they hadn’t even started making the first one? So they guessed, promising that if the Kickstarter hit $3.5 million, they would add a second city to the game.

  On October 16, 2012, the final day of their Kickstarter campaign, the staff of Obsidian held a party to celebrate their success. The whole company gathered in a conference room—much as they had seven months earlier, when they found out Stormlands had been canceled—and hooked up a camera so they could broadcast their reactions on an Internet stream. They drank, sang, and watched the final dollars pour in. When the countdown hit zero, they had raised $3,986,794—nearly four times their Kickstarter goal and double what Adam Brennecke, the most optimistic of the bunch, had hoped they could get.

  With additional backing they received on PayPal and the Obsidian backer website, their final budget was around $5.3 million.* Urquhart’s company had gambled that fans would show up and pay big for the type of classic RPG that publishers would never fund, and those fans had delivered. In just half a year, thanks to Armored Warfare and Project Eternity, Obsidian’s trajectory had reversed. The company was no longer on the edge of collapse. They finally felt free to do what they wanted to do, rather than what a publisher wanted from them.

  Now they just had to make a game.

  A week after the Kickstarter ended, once the dust had settled and the hangovers had eased, Josh Sawyer released a short video update to backers. “The Kickstarter’s over,” he said. “Thanks. But now it’s time to work.”

  Work, for Sawyer and the rest of the team, meant entering preproduction, the period of development when they’d try to answer fundamental questions about Eternity. They knew they wanted to make a game that felt like Baldur’s Gate, but if they couldn’t use the Dungeons & Dragons rule set, what would their characters look like? What kind of skills would each class have? How would combat play out? How many quests would be available? How big would the world be? What sort of graphic techniques would the artists use? Who would write the story? How big would their team get? When were they going to release the game?

  Josh Sawyer, a history buff and an avid cyclist with striking tattoos of poetry running up both arms, had taken the role of project director, the same job he’d had on Stormlands. Colleagues describe him as a director with a strong vision and a tendency to stick to that vision, no matter whom he rubs the wrong way. “Josh by far has the highest percentage at making the right decision the first time,” said Bobby Null, the lead level designer. “Eighty or eighty-five percent [of the time] he makes the right call. . . . When he needs to, he’ll fight the fight all the way up as far as it needs to go, which is not an easy fight sometimes to have when we’re talking about fair amounts of money.”

  In the opposite corner for many of those fights was Adam Brennecke, who took on the dual roles of executive producer and lead programmer. Brennecke, an easygoing soccer player with expressive eyebrows, liked to describe himself as the “glue guy”—the person in charge of ensuring that Eternity’s pieces would all fit together. He was also responsible for the budget. During the first few weeks after the Kickstarter ended, his job was to plan a schedule and figure out exactly how much money the team could spend on each part of the game. It made him a natural foil for Sawyer, who was looking to squeeze in as many ambitious ideas as possible.

  Early in preproduction, Sawyer stood his ground on big decisions; most significantly, the size of the game. The old Infinity Engine games, which the team had been playing during spare afternoons for inspiration, had worlds that were divided into individual map screens, each full of objects and encounters. The largest of these games, Baldur’s Gate 2, had nearly two hundred unique maps. To plan Project Eternity’s scope, Sawyer and Brennecke would need to figure out how many maps the game would have. Brennecke had settled on 120, but Sawyer disagreed. He wanted 150. And Sawyer wouldn’t relent, even though they all knew it would cost more money. “The way I view the relationship between production and direction is there intentionally should be a little bit of antagonism,” Sawyer said. “Not hostility—but the director at some level is saying, ‘I want to do this, I’m writing this check.’ Production holds the checkbook.”

  Thanks to crowdfunding, Brennecke had over $4 million in that checkbook, which was a lot of money for a Kickstarter project. But compared with modern major video game budgets, which can cap out at hundreds of millions of dollars, it was minuscule. Using the standard $10,000 per person per month burn rate, the Eternity budget could maybe feed a team of forty working for ten months. Or a team of twenty working for twenty months. Four million dollars could even fund a team of two people working for two hundred months—although Kickstarter backers might not enjoy waiting seventeen years for their game.

  That was all on paper, though. In real life the math was never that clean. A development team would expand and contract based on what it needed every month, and the budget would adjust accordingly. Brennecke would have to build a living schedule—a schedule that could change at any given day based on how they were doing. In game development, these schedules are always malleable: they have to account for iteration, human error, and the fact that creativity can come and go in bursts. “On top of that,” said Brennecke, “there is an expectation that we will get better and more efficient as we make more, and as the [development] tools get better.”

  The number of maps made a huge difference. Building 150 maps instead of 120 might mean several extra months of work, which would expand the schedule, which would cost Obsidian a lot of money. But Sawyer wouldn’t give in. “I think, in retrospect, that’s why our games are what they are,” said Brennecke. “Making crazy decisions like that, where it’s just, ‘We can do it. Let’s figure it out.’”

  As Brennecke and Sawyer battled over Project Eternity’s scope, the art team was running into its own set of problems. For years, Obsidian’s artists and animators had used a program called Softimage to create 3-D graphics, but by 2012 it felt obsolete, lacking pivotal features compared with its competitors. (By 2014 it would be discontinued.) To modernize, some of Obsidian’s owners and Rob Nesler, the art director, decided to switch to Maya, a more popular 3-D graphics tool that would work more smoothly with their game’s engine, a modified version of Unity.*

  Long term, Nesler knew this was the right choice, but there would be growing pains. It would take the art team weeks to learn how to properly use Maya, which meant that early production would move far more slowly. “People like to say, ‘Oh, it’s just another software package, you’ll learn it,’” Nesler said. “But to really learn how to achieve mastery and competence, to be able to adequately schedule how long it’ll take you to do something, you need to get to a level where you’re able to problem-solve in these packages. . . . It takes time, months or years to become so good at something that when someone says, ‘How long will it take you to do something?’ you can say, ‘It takes me this long.’”

  Without the ability to estimate the length of basic art tasks, the producers couldn’t put together an accurate schedule. Without an accurate schedule, they couldn’t determine how much money the project would cost. Four million dollars wouldn’t go very far if it took the team six months to build each map. If they were working with a publisher, they might have been able to renegotiate their contract to eke out more funding, but for Eternity that wasn’t an option. “The budget was the budget,” said Justin Bell, the audio director. “We couldn’t go back to the backers and say, ‘Give us more money.’ It just wouldn’t look good. So there was very little negotiation room.”

>   Because of the switch to Maya—and the art team’s lack of experience making isometric RPGs—it took a long time before Eternity’s early prototypes started looking right. For a while, it was too dark, too muddy, and too different from those old Infinity Engine games. After some heated arguments and long periods of iteration, the art team started to learn that there were certain aesthetic rules they needed to follow for a game like this. They shouldn’t have tall grass, for example, because it would hide the selection circles that appeared under Eternity’s main characters. Short grass would allow the player to better keep track of his or her party. Another rule: floors had to be as flat as possible. Maps with different levels of elevation were particularly tricky. You’d usually start from the south or west section of a screen and move north or east, so each map had to escalate accordingly. If you walked into a room and there was a set of stairs going upwards to the south, it would feel disorienting, like you’d just walked into an M. C. Escher painting.

  In the months following the Kickstarter, the ever-expanding Project Eternity team battled over dozens of these creative decisions, reducing the scope and cutting features as they tried to figure out the optimal way to build each area of the game. “For a game, especially with the ‘fun factor,’ you don’t really get it until you start playing it and seeing it,” said Brennecke. “You [think]: ‘There’s something that doesn’t feel right. What doesn’t feel right about this game?’ That’s where Josh and I come in and we sit down and we really analyze what is actually wrong with this.”

  After building a few technical prototypes, the team’s first major goal was to hit “vertical slice”—a small chunk of the video game designed to resemble the final product in as many ways as possible. During traditional, publisher-funded development, it was important for a vertical slice to look impressive, because if the publisher didn’t approve, the studio wouldn’t get paid. “When you’re focusing on a publisher, a lot of the times you’ll just do things the wrong way [on purpose],” said Bobby Null, the lead level designer. “It’s smoke and mirrors, hacking stuff in and trying to impress the publisher so they’ll keep paying the bills.” But with Eternity, the team didn’t have to fool anyone. The checks were already deposited. They could approach their vertical slice in what they liked to call the “right” way, drawing models and designing areas using the same methods that they would then use to make the final game, which helped save time and money.

  There was no big publisher demanding progress reports, but Obsidian did feel obligated to offer regular updates to the 74,000 Kickstarter backers who had funded Eternity. The upside of talking to fans was that they could be open and honest without worrying about a publisher’s iron-tight PR strategy. The downside was that they had to be open and honest all the time.

  Every week or two, the Obsidian team would put out a new update filled with details on what they’d been doing, sharing extravagant concept art and snappy chunks of sample dialogue. Some of these updates were incredibly in-depth, including photos of spreadsheets (spreadsheets!) and extensive explanations of systems like combat and character building. That meant getting instant feedback, which could be tough to handle. “You grow a thick skin pretty fast,” said Kaz Aruga, a concept artist.

  Eternity’s developers, like most of the people who make games, were used to developing in isolation, getting feedback from the outside world only when they released a new trailer or wandered around a trade show. With the Kickstarter approach, they’d get criticism in real time, which could help make the game better in a way that just hadn’t happened on previous projects. Josh Sawyer, who read the Kickstarter backer forums nearly every day, would constantly read and absorb fan feedback, to the point where he scrapped an entire system they’d planned after seeing a backer’s compelling explanation for why it shouldn’t be in the game. (That system, item durability, would’ve been tedious and boring, Sawyer said.)

  Some backers were outspoken and demanding, even asking for refunds when they didn’t like how Eternity was shaping up. Others were energetic, constructive, and supportive. A couple even sent packages to Obsidian full of treats. “It was actually really cool,” said Darren Monahan, one of Obsidian’s co-owners. “It felt like we had maybe three hundred or four hundred other people working on the game who weren’t really working on the game.”

  Toward the middle of 2013, the Eternity team finished the vertical slice and shifted from preproduction to production, the phase in which they would build the bulk of the game. The artists had grown familiar with the tools and pipelines; Josh Sawyer and the other designers had laid out systems like spellcasting and crafting; and the programmers had finished up fundamental features like movement, combat, and inventory management. The level designers had built outlines and sketches for most of the areas. But the game was still very much behind schedule.

  The biggest problem was Project Eternity’s story, which was coming together far more slowly than anyone on the team had expected. Sawyer and Brennecke had entrusted the main narrative to Eric Fenstermaker, a writer who had been at Obsidian since 2005. What made things complicated was that Fenstermaker was also the lead narrative designer on South Park: The Stick of Truth, a game that was going through a publisher change and its own development hurdles. Eternity was shipping later, so South Park became the priority.

  Fenstermaker had come up with some high-level ideas for Eternity, and the team had already established a lot of the background lore, but it was clear they needed help to finish all of the story and dialogue. In November 2013 the team brought in Carrie Patel, a published novelist with no previous video game experience, to be Eternity’s first full-time writer. “Narrative being a little late to the party was [a challenge],” Patel said. “The way the story was assembled was: there were a bunch of treatments put together by different people in preproduction, and the result was trying to take the best of all of those and assemble them into a story. It created some challenges that we might not have had if we said, ‘Let’s just write the story and figure it out.’”

  Patel found the transition to video games fascinating. Writing a video game was a totally different experience than writing a novel, in which a story moved in one linear direction. A game like Eternity would force writers to look at their stories less like roads and more like trees, where different players moved along different branches. Almost every conversation in Eternity allowed players to choose what to say, and the story had to account for each possibility. Late in the game, for example, the player had to decide which of the world’s gods to support on their quest to track down the villainous priest Thaos. Patel and the rest of the writing team had to dream up dialogue for each deity’s scenario, knowing that each player might see only one of them.

  At the end of 2013, Obsidian decided to put out a teaser trailer to give the world a taste of what it had been building. Adam Brennecke sat down and started editing the trailer, while the concept artist Kaz Aruga was tasked with designing the logo. For Aruga, another industry newcomer who had worked on Star Wars cartoons before coming to Eternity, this was a terrifying prospect. He’d been working at Obsidian for less than a year, yet now he had to come up with one of the most important parts of the game, the image that would be plastered across Eternity and all of its marketing materials for years to come.

  It was grueling. Every day, Aruga would get different bits of feedback from different department leads at Obsidian. Often these items of feedback would contradict one another, leaving Aruga wondering how he could possibly make everyone happy. “I was in the pressure cooker,” he said. “That was a learning experience.” It took more than one hundred different sketches before Aruga finally came up with a logo that pleased the whole team.

  On December 10, 2013, Brandon Adler, the production lead, posted an update on Kickstarter. “Through the hard work of the Project Eternity team we are proud to present our first in-game teaser trailer,” he wrote. The trailer zipped between snippets of gameplay, to the accompaniment of epic choral chanting. Wizards flun
g flames at giant spiders. An ogre slammed his hammer down on a party of adventurers. A massive dragon spat fireballs. At the end, Aruga’s logo appeared: towering onyx pillars flanking the project’s official new name, Pillars of Eternity.

  People were thrilled. This Pillars of Eternity game looked right out of the early 2000s, resembling Baldur’s Gate and all those other old Infinity Engine games that people had missed so much over the past decade, but the graphics were sharper and prettier. “Oh my gosh!” wrote one backer. “The indoor scenes look absolutely amazing!” wrote another. “And, you know, the outdoor ones too.”

  Exhilarated at the positive reactions, the Eternity team entered 2014 feeling like the project had momentum, though the workload in front of them was intimidating. Brennecke’s schedule had them on track for a November 2014 release, and there was still a great deal of game left to be finished. Even without a fancy 3-D world, Pillars of Eternity had grown humongous, thanks mostly to Josh Sawyer’s edict that they ship with 150 maps.

  On many games, once the schedule starts feeling too tight, the producers might look to cut features or areas that don’t seem critical. But with Pillars of Eternity, Adam Brennecke had a unique problem: Obsidian had already promised many of those features to fans. During the Kickstarter, the developers had made a big show out of having a huge, fifteen-level optional dungeon that they now had to construct, no matter how many late nights it would take. And then there was that second big city.

  Obsidian had already finished the first city, Defiance Bay, and it looked fantastic. It was complex, layered, and pivotal to the story. And now, after the developers had gone through the convoluted process of designing and modeling Defiance Bay’s various districts, the thought of building a second city was making them all queasy. “Everybody said, ‘I wish we hadn’t done that,’” said Feargus Urquhart. “Ultimately it was not necessary.”