Blood, Sweat, and Pixels

by Jason Schreier

  There was just so much work to do. They still had levels to finish, set pieces to tweak, art to polish. Catastrophes would occasionally impede their progress, like the time the studio’s servers kept crashing because Naughty Dog was uploading hundreds of thousands of humongous asset files per day.* Even with the bulk of the Uncharted 4 team working extended hours for five days a week (if not six or seven), the finish line seemed miles away.

  As more of the game came together, it became clearer which parts needed improvement, but it was difficult for team members to know what to prioritize. “There were a lot of leads meetings at the end where we were just reminding everybody, ‘Perfect is the enemy of good,’” said the writer Josh Scherr. “You’re polishing something that’s at ninety-five percent while this thing over here at sixty percent needs a lot of love. So that was what made crunch hard, because [when] you’d get down into it, you’d have trouble seeing the forest for the trees.”

  In the last months of production, Neil Druckmann and Bruce Straley decided to cut one of the team’s favorite set pieces, a sequence in Scotland during which Drake would climb a giant crane, then try to escape as it collapsed, fighting off enemies as he went. The designers had made an elaborate prototype for this crane sequence, and they all thought it’d be an exciting spectacle, but they just didn’t have enough time to get it finished. “We had that all working in a prototype form,” said Evan Wells. “But then just the sheer amount of polish it takes to take that prototype to completion . . . You hit FX, and sound, and animation. It hits every single department in so many ways, and it might take you a few days to put that prototype together, but it’s going to take you a few months to really complete it.”

  Druckmann and Straley also started ramping up their focus tests. Participants in these tests—usually a diverse group of LA residents with varying amounts of gaming knowledge—would stroll into the office and assemble at a row of desks, each equipped with a pair of headphones and the latest build of Uncharted 4. As these testers played through multihour sessions of the game, faces glowing from the pale blue light of their monitors, Naughty Dog’s designers could watch everything. They could gauge testers’ physical reactions through a camera hooked up to the testing room, and see exactly what each tester was playing at any given time. Naughty Dog’s designers could even annotate the footage, sticking in relevant notes like “died ten times here” or “face seemed bored.”

  During that final stretch, these focus tests became even more important to Druckmann and Straley. As Uncharted 4 morphed from a collection of prototypes and gray-boxed levels into an actual game, the directing pair could start pinpointing big-picture problems like tone and pacing. After nearly two years with the game, Straley and Druckmann had lost all semblance of objectivity, which was why testers were so important. “They might not get certain mechanics,” said Druckmann. “They might be lost for a while and lose all sense of pacing. They might be confused by the story and not pick up on certain nuances that you thought were pretty clear and wound up not being clear at all.” Bits of individual feedback weren’t that helpful—what if that bored tester was just having a bad day?—so instead, Naughty Dog’s leads looked for trends. Were a lot of people getting stuck on the same encounter? Did everyone think one section of the game was too boring?

  “The first few focus tests are really humbling,” said Druckmann. “Bruce and I now get excited by them because it just puts everything into perspective. They’re brutal and you feel defeated, and a lot of times the designers hate watching levels being played and they’re pulling their hair out. ‘No, turn around, the handhold is right there, what’re you doing?’ And you can’t do anything.”

  By the end of 2015, Uncharted 4’s deadline looked terrifying. To hit its scheduled release date of March 18, 2016, Naughty Dog would have to submit its gold master in mid-February.* Fixing all the highest-priority bugs by February felt like it’d be an impossible task, and everyone at the studio was consumed by anxiety that the team might not get it done. “You always have that sinking feeling in your stomach, that you just don’t know how it’s going to pan out,” said Evan Wells. “You just know, well, we’ve got three months. And you start looking back at old bug reports that you had in your last game. ‘Three months out, how many bugs did we have? How many were [A-level bugs]?* How many were we fixing every day?’ OK, we were fixing fifty a day, so we start running all those calculations like, ‘All right, we’re not totally screwed.’”

  It helped that they were running focus tests every week. As they fixed more and more bugs, which ranged from major (“game crashes every time you shoot in the wrong place”) to minor (“it’s hard to tell that you should jump here”), the focus testers would give the game higher and higher ratings. And Naughty Dog’s staff kept pushing themselves, working longer hours than they’d ever worked before to get Uncharted 4 in shape. Problem was, they were running out of time.

  “We started saying, ‘OK, well, we’re going to have to patch,’” said Wells, referring to the increasingly common practice of releasing a “day one” patch to fix bugs that have made it into the gold master. “When we put something on a disc, it’s not going to be a completely polished, Naughty Dog experience, and we’re going to use those three to four weeks between printing the discs and it landing on the shelves to put that polish in.” In other words, anyone who bought Uncharted 4 and played it without connecting to the Internet and downloading the patch would get an inferior version of the game. “We started trying to prepare Sony: ‘OK, we’re going to have a pretty big day-one patch,’” said Wells. “Things are not looking too smooth. We’re going to come in pretty hot.”

  As Sony’s engineers started preparing to expedite the patching process for Naughty Dog, word started getting around the company that Uncharted 4 was on shaky ground. And eventually, the news reached the very top.

  One night in December 2015, Wells was sitting in his office and playing through a build of Uncharted 4 when he heard his cell phone buzzing and saw a San Francisco number he didn’t recognize. It was Shawn Layden, the president of Sony Computer Entertainment America and the man in charge of all of Sony’s development studios. Layden told Wells he had heard Uncharted 4 needed more time. Then he dropped a megaton.

  “[Layden said], ‘So what if you guys had a ship date in April?’” said Wells. “And I [said], ‘That would be so awesome.’ So he said, ‘You’ve got it. That’ll be your new ship date.’”

  Instead of hitting gold master by mid-February, they now had until March 18 to fine-tune, fix bugs, and ensure that Uncharted 4 felt polished even without a day-one patch. It was, as Wells would later describe it to staff, a “Hanukkah miracle.”

  A little while later, a representative of Sony’s Europe division came to Wells and asked if Naughty Dog could actually submit the gold master on March 15. Sony was overhauling some manufacturing plants in Europe, the representative said, and in order to print Uncharted 4 on time, they’d need the game three days earlier.

  “And we [said], ‘Really?’” said Wells. “Those three days are huge for us. We really need that time.” But Sony Europe had no flexibility. If Naughty Dog needed those extra three days, then they’d need to delay Uncharted 4 again, this time to May. “It was so frustrating,” said Wells, “because we were just [thinking], ‘Oh my gosh, people are going to blame us again,’ and it was not our fault this time. It was not our fault. But hopefully people have forgotten about that, and just remember the game.”

  Getting all that extra time was indeed a Hanukkah miracle for Naughty Dog, but every extra week of development meant an extra week of crunch. It was rough, said Bruce Straley, “especially when your brain and your body are saying, ‘I’ve got just enough for one more week,’ and you’re like, ‘Wait, I’ve got three more weeks?’ That was really hard at the end.” But they made it happen, thanks largely to Naughty Dog’s collective experience and knowledge of when to stop working on a game. “It’s the old phrase, ‘Art is never finished, it’s just aba
ndoned,’” said Straley. “The game is just shipped. That’s our motto in that last quarter of production; I walk around just saying, ‘Ship it’ to everything I see. ‘Ship it. Ship that.’”

  And on May 10, 2016, they shipped it. Uncharted 4 sold 2.7 million copies in a week, received stellar reviews, and was, without a doubt, the most impressive-looking photorealistic game that had been made at that point. In the coming weeks and months, a number of burnt-out developers left Naughty Dog. Others took long vacations and started prototyping for their next two projects: the Uncharted 4 expansion pack, which later morphed into a stand-alone game called Uncharted: The Lost Legacy; and a sequel to The Last of Us, officially titled The Last of Us: Part II.

  For these projects, Naughty Dog would have proper preproduction time, and the team wouldn’t have to jump right into the fire. “That’s why you see everybody happy now,” said Bruce Straley when I visited the studio in October 2016. “They’re going home at decent hours, coming in at decent hours. They go surfing in the morning. They’re going to the gym at lunch.” But just a couple of months later, after Naughty Dog announced both games, news came out that Straley wasn’t returning to codirect The Last of Us: Part II. The official word at the studio was that he was taking a long sabbatical.

  At the end of Uncharted 4, after nearly losing their lives in Libertalia but somehow escaping, Nathan and Elena realize that maybe they need a little bit of adventure after all. Maybe there is a way to find a balance between their work and their personal lives. Elena explains that she’s just purchased the salvage company where Nathan works, and that she wants the two of them to go on journeys together—minus the near-death experiences. From now on, they’ll be hunting for artifacts in a far more legal capacity. “It’s not going to be easy, you know,” says Nathan. Elena stares at him for half a second before responding. “Nothing worthwhile is.”

  3

  Stardew Valley

  Amber Hageman was selling pretzels when she first met Eric Barone. She was about to finish high school, he had just started college, and they both worked at the Auburn Supermall, just south of Seattle. Barone was handsome, with dark eyes and a shy smile, and Hageman was drawn to his passion for making things—little games, musical albums, drawings. Soon, the two started dating. They discovered that they both loved Harvest Moon, a series of serene Japanese games that put players in charge of restoring and maintaining their own farms. During dates, Hageman and Barone would sit side by side and play Harvest Moon: Back to Nature on PlayStation, passing the controller back and forth as they befriended villagers and planted cabbages for profit.

  By 2011 the couple had grown serious and started living together at Barone’s parents’ house. Barone, who had just earned a degree in computer science from the University of Washington Tacoma, was struggling to find an entry-level programming job. “I was just kind of nervous and awkward,” Barone said. “I didn’t do well at the interviews.” As he scuffled around the house, applying for any job he could find, Barone started thinking, Why not make a video game? It’d be a good way to improve his programming skills, build some confidence, and maybe help him land a decent job. He’d tinkered with a few big projects before, like a Web-based clone of the action game Bomberman, but he hadn’t completed any of them. This time, he told himself, he would finish whatever he started. He told Hageman he’d be done in six months or so, just in time for a new wave of job openings.

  Barone’s vision was well defined, if unglamorous: he wanted to make his own version of Harvest Moon. The original series had grown less popular thanks to a trademark dispute and a sharp decline in quality, and it was tough to find a modern game that evoked the tranquility of the original farm simulators.* “I just wanted to play another game that was exactly like the first two Harvest Moons, but just with different people and a different map,” he said. “I could’ve played that forever, different iterations of that same thing. But it didn’t exist. So I just figured, why hasn’t anyone made this? I’m sure there’s a lot of people out there who want to play this kind of game.”

  He also wanted to do it solo. Most video games are built by teams of dozens of people, each of whom specializes in fields like art, programming, design, or music. Some games, like Uncharted 4, employ staffs in the hundreds and use work from outsourced artists across the world. Even small independent developers usually rely on contractors and third-party game engines. Eric Barone, a self-proclaimed introvert, had a different plan. He wanted to write every line of dialogue, draw every piece of art, and compose every song on the soundtrack by himself. He even planned to program the game from scratch, eschewing established engines, because he wanted to see if he could. Without collaborators, he wouldn’t have to talk through ideas or wait for approval to get things done. He could make decisions based on what he—and only he—thought was best.

  Barone planned to put his little Harvest Moon clone on the Xbox Live Indie Games (XBLIG) marketplace, a popular storefront for independent developers. Unlike other digital distributors in 2011, XBLIG had few restrictions and would feature games from any developer—even a college graduate with no experience. “My idea at the time was it would take a few months, maybe five, six months, and then I would put it on XBLIG, and sell it for a couple bucks, and maybe make a thousand dollars off it or something,” Barone said. “And it would be a good experience and then I would move on.”

  Using a rudimentary set of tools called Microsoft XNA, Barone started writing basic code that would let his characters move around on two-dimensional screens. Then he ripped some sprites from Super Nintendo (SNES) games and taught himself how to animate them, manually drawing different frames to create the illusion that the images were moving.* “There was no methodology at all,” Barone said. “It was just completely haphazard and scrappy and random.”

  By the end of 2011, Barone had given up on finding a day job. He’d become obsessed with this new project, which he called Sprout Valley (later retitled Stardew Valley), and he wanted to finish it before joining the grind of full-time employment. The premise of Stardew Valley was simple. You’d create a character and customize his or her appearance, from hair to pants color. At the beginning of the game, your hero would quit his menial office job at a giant corporation to go live in an idyllic village called Pelican Town, where he’d inherited an old, disheveled farm from his grandfather. As this hero, you’d be tasked with growing crops, cultivating relationships with the villagers, and restoring Pelican Town to its former glory. Barone wanted to make it feel satisfying to do mundane chores like planting seeds and clearing debris in Stardew Valley, just as it was in Harvest Moon. You’d even be able to team up in online multiplayer with your friends.

  Barone’s daily habits rarely changed: every morning he’d wake up, make coffee, and stumble over to his computer, where he’d spend anywhere between eight and fifteen hours plugging away at the game. When Hageman got home, they’d eat dinner and go for a walk, where they would talk about Stardew Valley and mull over important questions like “Which characters should you be able to marry?” and “Which characters should you be able to kiss?”

  Living rent free helped Barone get away with this pattern for a few months, but soon the couple wanted to move out on their own. They’d saved some money while living at Barone’s parents’ house, which helped, but it couldn’t pay for everything, especially if they wanted to live in downtown Seattle. Barone’s video game project brought in a whopping zero dollars a month, so Hageman, who was finishing up her undergraduate degree, had to support them both. Once they found a place, she started juggling two jobs, working as a coffee shop barista on the weekends and a caretaker after school. “We just kind of lived a humble life and it worked for us,” Hageman said. As the months went by, they settled into this routine: Barone worked on his video game; Hageman paid for food, expenses, and rent on their tiny studio apartment.

  A less patient girlfriend might not have tolerated this arrangement, but Hageman didn’t seem to mind it. “When we were living at h
ome it wasn’t as hard, and then when we moved out to Seattle, the reality of having to support him became more real, but it really wasn’t a problem ever,” she said. “He was working so much that it was impossible to be frustrated.”

  It was true: Barone was working a lot—but he wasn’t working very efficiently. Because he was making Stardew Valley by himself, there was nobody to hold him accountable or force him to stick to a schedule. He had no employees or expenses. No producers were hovering behind his computer chair, telling him to stop overscoping and just ship the damn game. Any time Barone thought of a cool feature or an interesting character for players to befriend, he’d add it. Every week, the game grew exponentially bigger.

  Of course, it’s not very hard to distinguish between a game made by hundreds of people and a game made by one. The more realistic a video game looks—the higher the graphic fidelity, the more polygons in each 3-D model—the more likely it’s the product of a humongous, expertly trained staff that can handle highly technical aspects of art and engineering. Games like Uncharted 4 demanded massive teams (and tens of millions of dollars) because they needed to make people’s eyes pop.

  For Barone, sitting alone in his studio apartment, game development meant something else entirely. His game didn’t have high-end 3-D graphics or a fully orchestral soundtrack. Stardew Valley used hand-drawn two-dimensional sprites and music that Barone composed himself with an inexpensive audio production program called Reason. Although Barone had little experience making games, he did know how to write music from years of playing in bands. (In high school, he’d wanted to be a professional musician.) He’d learned how to program in college, and he was gradually teaching himself how to draw the simplistic backgrounds and sprites that would make up Stardew Valley. By reading pixel art theory and watching guides on YouTube, Barone figured out how to compose each sprite by drawing individual pixels. He knew nothing about complicated video game lighting techniques, but he learned how to fake them, drawing semitransparent white circles that he’d place behind torches and candles to evoke the illusion that they were brightening rooms.