Blood, Sweat, and Pixels

by Jason Schreier

  What was most gratifying for Josh Mosqueira was that people especially loved the console version of Diablo III, which launched on PS3 and Xbox 360 in September 2013 and on the newer consoles (PS4 and Xbox One) in August 2014. After decades of clicking, many gamers almost felt sacrilegious saying so, but playing Diablo III was more fun with a PlayStation 4 controller than it ever had been with a mouse and keyboard.

  In the coming months and years, Blizzard would go on to release more patches and features for Diablo III. Some were free, like a dungeon called Greyhollow Island and a revamped version of the cathedral from the first Diablo game. Others cost money, like the necromancer character class. Although fans lamented the absence of another big expansion pack (which, as of early 2017, hadn’t happened), it was clear that Blizzard had committed to supporting Diablo III for years after its release. Other developers might not have bothered—especially after the launch day catastrophe. “Mike Morhaime, the president of the company, said to us, ‘We want to win and earn the love and trust of our players,’” said Wyatt Cheng. “We had all this work we had done in this game. We believed in the game. We knew it was great, and it would’ve been so tragic, I think, if we were at a company that said, ‘Oh, Error 37, pull the plug.’”

  Josh Mosqueira, on the other hand, was done with Diablo III. In the summer of 2016, Mosqueira left Blizzard, joining Rob Pardo, a veteran Blizzard executive and the lead designer of World of Warcraft, to form a new studio called Bonfire. “Leaving this team and this company was the hardest non–life threatening decision I’ve ever had to make,” Mosqueira said. “I felt that I wanted to take a chance and try to do something totally different.”

  At least he’d left Blizzard with something great. Diablo III was one of the best-selling video games in history, having sold thirty million copies as of August 2015. It also proved a point that would influence numerous game developers in the years to come, including but not limited to the makers of The Division and Destiny (whom we’ll meet in chapter 8): every game can be fixed.

  Often, the developers of a video game hit their stride toward the end of a project, when they learn what their game really feels like to play. For Diablo III and games like it, launching was just the beginning of the development process. “Even with a game that has a really strong vision, a really strong identity like Diablo,” said Mosqueira, “I think one of the challenges is that at the beginning of a project . . . before a game comes out, everybody has a slightly different version of the game in their head. One of the hardest things to do is get it out. But once it’s out, there’s less discussion about that because you can now [see] what it is. Game development’s really hard, but it’s a different type of hard before it’s released. It’s more existential before it’s released.”

  Diablo III was proof that, even for one of the most accomplished and talented game studios in the world, with near-limitless resources to make a game, years can pass before that game properly coalesces. That even for the third game in a franchise, there are still an impossible number of variables that can throw everyone off. That even a game that launches with crippling issues can, with enough time, commitment, and money, turn into something great. In 2012, when Error 37 spread across the Internet, gamers had thought Diablo III was doomed. And then it wasn’t.

  5

  Halo Wars

  In the summer of 2004, Ensemble Studios’ management team flew to Chicago for an off-site retreat that felt unusually somber. They were having what some would call an identity crisis. Ensemble had spent years working on prototypes for video games that just weren’t making any of them happy, and they needed to get away for a few days and talk things through. “We said we should be doing what we’re passionate about,” said Chris Rippy, an Ensemble producer who was at the retreat. “Get back on track.”

  The question was, what would that look like? Ensemble, based in Dallas, Texas, had been making the same type of video game for nearly ten years. They’d built their reputation on Age of Empires, a cerebral game series in which you’d start off with a handful of villagers and gradually build an entire civilization. Like Blizzard’s Warcraft and Westwood’s Command & Conquer, Age of Empires was a real-time strategy game, or RTS, in which the action unfolded without turns or pauses. You’d progress through different technological “eras” (Stone Age, Bronze Age) as you harvested resources, constructed buildings, and trained armies to conquer your opponents.

  To finish the first Age of Empires, Ensemble’s staff went through what Dave Pottinger, a lead designer, described as “a terrible death march which could never be repeated today,” working one-hundred-hour weeks for nearly a year. When it finally came out in 1997, Age of Empires was an immediate hit, selling millions for Ensemble and its publisher, Microsoft. This was followed by a glut of profitable Age sequels, expansions, and spin-offs, eventually leading Microsoft to buy Ensemble. By 2004, the studio was working on yet another game in the series, Age of Empires III, which was due to come out at the end of 2005.

  What had always been unorthodox about Ensemble—and what ensured that people rarely left the studio—was that they saw themselves as a family. Since Ensemble’s founding in 1995, the company had been made up mostly of young, single men who spent nights and weekends together. “On any given Saturday night, a good chunk of the team was at someone’s apartment,” wrote Ian Fischer, a lead designer, in a magazine retrospective.* “Friday after work there’d be a dozen people in the play-test area playing Quake (then Quake 2, then Half-Life) until 3am. . . . If you didn’t actually share an apartment with the guy you sat next to, you were probably going to be having a beer at his place later that night.”

  Even through the early 2000s, as Ensemble’s cofounders grew up and started real families, they shared what they saw as an uncommon amount of chemistry. Every prospective employee had to go through a rigorous interview process that, for a while, included meetings with all twenty-something people at the studio. If just one person said no, the prospective employee was out. “It was really like a family,” said Rich Geldreich, a graphics engineer. “It was a combination family slash a little bit of a frat house.”

  During that summer of 2004, with Age of Empires III in the works, many of Ensemble’s veterans grumbled that they were sick of making Age games. Some were tired of RTS games in general. They’d tried several times to get a second development team off the ground, one that would experiment with other genres, but that team always seemed to fail.

  Ensemble would fall into the same pattern every time. The second team would play around with prototypes and early concepts for a while, then the main Age team would invariably run into some sort of catastrophe, such as when they had to redo Age of Empires II’s entire design after a year because it wasn’t fun to play. Ensemble’s management would then ask the second team to put their own project on hold and come help with the latest Age game. Every time this happened, the second project would lose momentum and peter out, like a steam engine that stopped getting coal.

  “The pattern we fell into, starting with Age II, was that we so overbid and underfinished the games that even though we tried to do other things, we ended up having to pull everybody in just to get the [Age] game finished,” said Dave Pottinger. In previous years, Ensemble’s second team had prototyped an RPG, a platformer, and several other experiments, all of which were discarded. Canceling prototypes wasn’t an unusual practice for a game studio, but the pattern was becoming rough for Ensemble’s staff, who wanted to show the world that they could expand beyond Age of Empires. That was why the management team was sitting in Chicago during the summer of 2004, trying to figure out their next move.

  Chris Rippy, Dave Pottinger, and the rest of Ensemble’s managers evaluated their options. They had two projects in development: there was Age of Empires III, their next flagship game, and then there was a car action game called Wrench that no one was particularly excited about. For two days, they sat and debated. Did they really want to keep working on Wrench? Why were they still doing it? In
ertia? Tons of other ideas kept resurfacing. What if they made a Diablo clone? What if they built a massively multiplayer online game (MMO)? And what if they developed a real-time strategy game for consoles?

  For decades, the accepted wisdom had been that real-time strategy games would work only on a computer. Because of the genre’s speed and complexity, the RTS was best suited to a mouse and keyboard, so a player could zoom around the map with one hand and type in orders on the other. Consoles didn’t have mice and keyboards; they had controllers, whose joysticks and limited buttons proved ineffectual for fast-paced multitasking. For as long as RTS games had existed, nobody had been able to solve this problem. Zealots of PC gaming would point to StarCraft 64 (2000), a subpar port of Blizzard’s StarCraft that didn’t capture the nuances of its PC counterpart, as evidence that an RTS simply couldn’t work on a console.

  Angelo Laudon, a veteran programmer and member of the Ensemble management team, had always believed otherwise. Like most engineers, Laudon relished the act of solving impossible problems, and he thought a console RTS made for one hell of a challenge. “Angelo was fired up about it,” said Chris Rippy. “We had some loose ideas on how to make it work.” During the meetings in Chicago, both Laudon and Rippy pushed hard for a console RTS, arguing that it was the perfect opportunity for Ensemble. They had talent and experience. Gamers respected them as RTS developers. And their parent company, Microsoft, was months away from launching the much-anticipated Xbox 360, a powerful new machine that could be home to the first great RTS on consoles.

  This plan would also allow Dave Pottinger and others on the team who were sick of making RTS games to work on other projects. For years, Ensemble’s CEO, Tony Goodman, had wanted to make an MMO along the lines of Blizzard’s World of Warcraft, which was due to come out a few months later, in November 2004. (Nearly everyone at Ensemble was playing the beta.) The thought of developing a big multiplayer game was polarizing at the studio, but Goodman and some of the other veterans, like Ian Fischer, had always wanted to make it happen.

  By the end of the Chicago trip, Ensemble’s management had green-lit three new projects. First was the console RTS, which they gave the code name Phoenix. Second was the MMO, code-named Titan. The third game, code-named Nova, would be a sci-fi action-RPG and Diablo clone led by Pottinger, once he and his team had finished work on Age of Empires III. If Blizzard could juggle StarCraft, World of Warcraft, and Diablo, the managers thought, why couldn’t Ensemble do something comparable?

  When the management team got back to Dallas, they announced that they were canceling Wrench and starting these other, more exciting projects. On Phoenix, Angelo Laudon would be the lead programmer. Chris Rippy would be the producer. And to figure out what the game would feel like to play, they brought in one of their most experienced developers—a designer named Graeme Devine.

  Devine, a Scottish expatriate with long hair and a high-pitched laugh, had one of the most eclectic résumés in gaming. As a teenager in the 1980s, he’d done programming work for Atari, Lucasfilm Games, and Activision. Before he’d turned thirty, Devine had founded his own company, Trilobyte, and designed a smash hit called The 7th Guest that blended tricky puzzles with cinematic storytelling. After falling out with his Trilobyte cofounder and watching the studio collapse, Devine went to id Software to work alongside one of his old friends, the legendary programmer John Carmack. He spent four years helping develop games like Quake III and Doom 3 before leaving for Ensemble in 2003, where he was put in charge of coding pathing algorithms—the mathematical formulas that told units how to move—for Age of Empires III.

  In hindsight, this felt like an odd choice to the management team. Devine had a legendary brain for game design and storytelling—why did they have him teaching Age of Empires units how to walk around? Recalled Devine: “Dave Pottinger called me up and said, ‘Hey, we’re not really using you properly at Ensemble. We’d really like to try to break this RTS on a console thing—how’d you like to head up a team to try to get RTS on a console?’”

  Sure, Devine told Pottinger. When would they start?

  The Phoenix team was now Graeme Devine, Angelo Laudon, and Chris Rippy, as well as a few other artists and programmers. It was a small, close-knit group whose members were excited to work on something that wasn’t Age of Empires. “Our team of misfits,” Devine liked to say. It didn’t take long before they had a basic concept in place. Phoenix would be a sci-fi game in which humans battled against aliens, sort of like StarCraft. Playing as the humans would feel drastically different from playing as the alien race, which Devine called the Sway. The goal was to evoke War of the Worlds, pitting hulking alien war machines against the scrappy armies of mankind.

  The ideas were the easy part. Far harder, Devine knew, would be figuring out the controls. Most video games, when they first enter development, already have a set of controller conventions they can follow. You know that if you pick up a first-person shooter, the right trigger is going to fire shots and the left trigger is going to aim. The left joystick will always control the character’s movement while the right joystick rotates the camera. Designers working on big series like Call of Duty or Assassin’s Creed might make small tweaks to the core formula, but they don’t have to rethink the controls every time they start developing a new game.

  Phoenix didn’t have that luxury. The main reason real-time strategy games had never done well on consoles was that there had never been a great control scheme. Right away Devine and the team had to start asking fundamental questions that most games would have already solved. How would you move around the screen? Would you control a cursor, as you would in Age of Empires, and use that to select units and give them actions? Or would you control units directly? Would you be able to select multiple units at once? How would you construct buildings?

  Answering these questions would be impossible without testing all the possibilities, so Devine and his team spent months experimenting. “We were burning through tons and tons of prototypes,” said Colt McAnlis, who had been hired as a graphics engineer for Phoenix. “How to move the camera; how that works, moving the camera and grabbing units; how not to make people nauseous in the process. It was a lot of groundbreaking R and D.”

  They cranked through hundreds of different control schemes. They played games like Pikmin and Aliens versus Predator: Extinction, both console games with RTS elements, and tried to pinpoint what worked. “We started incorporating all of their ideas, and skipping the stuff we didn’t care for,” said Chris Rippy. “All of these games did something really smart and really cool, and we just kind of put them all together into one package.”

  Every week—usually every day—they’d hold mandatory playtesting sessions, where the small Phoenix team would sit down and play through the prototypes they’d developed that week. It was a rigorous, focused process that led them to throw out a whole lot of work. “The feedback system at Ensemble was brutal, it was honest, but it forged a good game,” said Devine. From there, they came up with a few control schemes that seemed to feel satisfying, like an “area select” option that would let you take command of a group of units by holding down the A button.

  Now they just needed corporate approval. Although Microsoft gave Ensemble the freedom to pursue prototypes like Phoenix and the other “second team” games they’d canceled over the years, those games couldn’t move forward until they got an OK from Ensemble’s parent company. Until the bosses in Redmond gave the green light to Phoenix, it would remain in purgatory, unable to enter full production or hire new staff.

  When Xbox executives came down to visit, the Phoenix team showed them their prototypes. Microsoft seemed happy with what they saw, telling Ensemble to keep up the good work. Toward the end of 2005, as the rest of the studio prepared to ship Age of Empires III, the Phoenix team kept plugging away, buoyed by the fact that they’d finally made a non-Age game that Microsoft cared about.

  Members of the Age of Empires III team, busy working late nights to finish their ga
me, were far less thrilled. “This was an interesting period for Ensemble because it was the first ‘second game’ that Microsoft got excited about,” said Dave Pottinger, who was the lead designer on Age of Empires III. “And unfortunately for us, it was an RTS. Because we had tried for so long not to make an RTS. We didn’t want to get pigeonholed into just the RTS company. But of course that’s exactly what Microsoft wanted. That’s what we were good at, that’s why they had bought us, and that’s what they wanted us to do.”

  Microsoft hadn’t green-lit any of Ensemble’s previous prototypes, but its representatives had really liked Phoenix. Executives on the Xbox team, who were laser focused on beating Sony’s PlayStation 3 in the so-called console war, loved the idea of getting one of their flagship studios on the new Xbox 360. What they didn’t love was that this was a new IP. In Microsoft’s eyes, this sort of brand-new franchise would be hard to market properly. Real-time strategy games were a tough enough sell in the first place, often performing far worse in the market than the first-person shooters Microsoft loved making. Without some sort of established brand attached to Phoenix, Microsoft’s executives feared that the game wouldn’t move enough copies to justify their investment.

  For Ensemble, getting a game green-lit meant progressing through a series of meetings with Xbox executives like Phil Spencer and Peter Moore, which is what the Phoenix team had to do. During the first meeting, things went smoothly. But when Graeme Devine and his team met with Microsoft for the second green light meeting, Microsoft gave them a new mandate: Make it a Halo game.