Quantcast
Making a Big Budget Video Game Is Riskier and Harder Than Ever. So Why Do It?

Inside the AAA machine with Gears of War 4.

This is part of Motherboard's series on the making of Gears of War 4. Read more.

Rod Fergusson is playing an unfinished video game in front of 3,000 people at the Galen Center in Los Angeles. Another 4.5 million are watching him live at home. It's Microsoft's E3 2016 press conference, and this is Fergusson's opportunity to show everyone why they should spend $60 on Gears of War 4 when it comes out on October 11.

At its peak, around 330 people worked on Gears of War 4 at The Coalition, a game development studio in Vancouver, British Columbia. That number doesn't include staff at Microsoft, off-site contractors, and other studios commissioned to work on different parts of the game. Whether Gears 4 succeeds or fails will have a direct impact on Xbox One, Microsoft, and hundreds of developers and their families.

"I feel 100 percent responsible," Fergusson told me on the phone before E3. He's studio head at The Coalition and its public face in the media and industry events. "If Gears 4 isn't what people want it to be, if it comes out and people say Gears 4 is a bad game, I'll shoulder a huge amount of that responsibility. That would be my fault."

On screen, the characters make their way across the top of a dam that's crumbling under the stress of a lightning tornado. Fergusson shoots a bomb out of the air, and wind pulls the debris into the base of a tall tower. The tower explodes and crashes, with smoke and bits of detritus swept up into the tornado. Like many of the set pieces in the big budget, AAA games Gears 4 is competing with, it looks incredible.

As the characters fight across the dam, one of them pulls out the lancer, a giant machine gun with a chainsaw at the end—a defining feature of the Gears series. "Take it!" the character taunts a slimy, humanoid monster before she cuts it in half and the bloody chunks of meat get swept up in the storm as well. Like many of the AAA games Gears 4 is competing with, it's an overwhelming, juvenile spectacle that will sell millions of copies.

Fergusson demoing Gears 4 at the E3 2016 Microsoft press conference rehearsal Image: Casey Rodgers

I wanted to find out why Microsoft, Fergusson, and the hundreds of people who created Gears 4 still choose to work on this kind of game in a time when there are more alternatives than ever in the games industry. From a creative perspective, smaller indie games give developers more freedom, and from a business perspective, mobile games are set to make more money than console games this year. Game industry research firm Digi-Capital estimates that mobile game revenue exploded from less than $2 billion in 2010, to $35 billion in 2016.

"The games market changed dramatically in the last 5 years, with the rise of mobile growing the market overall and cannibalizing other sectors," Digi-Capital managing director Tim Merel told VentureBeat.

So why do companies spend millions of dollars and why do developers work 12-hour days for months on end to make a digital tower explode spectacularly?

To find out, Microsoft gave Motherboard what it called "unprecedented access" to one of its development studios, including allowing me to embed with the development team during the week leading up to the game's certification, which is often referred to as "cert". This final test of the game by Microsoft, ensuring it meets the company's standards, is the last step before the code is shipped to retail (any minor problems Microsoft or the developer finds during cert will be addressed in a "day one patch"). This phase often also coincides with what developers call "crunch," a time when they work extra long hours to meet the deadline.

Over the course of a week in late August at The Coalition studio in Vancouver, I found rendering technology indistinguishable from magic, and overflowing email inboxes. I found mechanical engineers and filmmakers who abandoned those disciplines to work on video games. I found a lot of people who were tired but determined, as if finishing a three-year marathon with a two-month sprint.

I found that AAA games are patently ridiculous endeavors. As game developers are fond of saying, they're a lot like the biggest blockbuster movies of the summer, only they have to reinvent and build the cameras from scratch every time they make one. Rarely does our society put more time, effort, and money into what is ultimately escapism. But I also found it's essential to the video game industry that we keep making AAA games, not because they are easy, but because they are hard.

Back at the Galen Center, the Gears 4 demo ends and the crowd cheers. When the lights come back on Fergusson is already off the stage and off to the next thing. There's still a lot of work to do.

6 DAYS TO CERT

BUG COUNT: 11

"DON'T SHAKE THE JELLO"

It's afternoon, Tuesday, August 24, six days before The Coalition has to submit Gears of War 4 to certification. Fergusson is holding court in a conference room called BigPark Place where he meets with a dozen senior members of the team he calls the "Dirty Dozen."

The studio, which occupies floors six through eight in a building overlooking BC Place stadium in downtown Vancouver, reflects the messy science and art of developing a video game: Character sketches and project road maps line the walls, heads peek above a forest of monitors, and doodles cover whiteboards with both crucial notes and inside jokes only people who spend way too much time together in the same room will understand.

BigPark Place, by contrast, is more like a dark operating room: quiet, clean, and corporate. The shades are drawn to accommodate a projector that's plugged into one of the developer's Surface Book Pros—it is a Microsoft studio—and staff sits around a long table.

Fergusson is an effective communicator. He maintains eye contact. His speech is devoid of "ums." He answers every question with an interesting story from a previous point in his career, which he started at Microsoft in 1996, providing technical support to enterprise customers. Once, at an Xbox fan event at E3, I shared a table with Fergusson and the five actors who lend their voice to Gears 4, including Liam McIntyre (the lead in Spartacus) and John DiMaggio (Bender from Futurama). It was too much personality for one table, but Fergusson has the kind of presence that dominates and directs the conversation.

Fergusson and the Dirty Dozen have triage meetings in BigPark Place at least three times a day. Once in the morning, once in the middle of the day, once in the evening, and anywhere in between as necessary.

Fergusson and the triage team reviewing a bug. Image: Jackie Dives/Motherboard

At this point of the process, most of what players think of as game development has been over for months. There are a lot of unmanned desks. Level designers aren't designing levels. Artists and sound designers aren't creating new sound or art. Most of the people in the building have rolled off the project and are either on vacation or in the very early stages of the next project. Only a dozen people are even allowed to touch the code. The time for new ideas, as far as Gears 4 is concerned, is over.

The remaining team is working hard to reach what it calls a "zero bug release" (ZBR), or "release candidate zero" (RC0). This build of the game that supposedly has no major bugs is then sent to the small army of testers that make up The Coalition's quality assurance (QA) team, which will put the game through its paces for three days, and inevitably find more bugs. Prince Arrington, director of quality assurance, told me that they sometimes refer to RC0 as RCLOL, since there's no actual chance that the testers won't find more bugs in it. The team attempts to squash the bugs, creating a new build—RC1—which will be sent for another round of testing.

This process repeats itself, reaching RC2, RC3, and so on, until one of the release candidates holds through testing, at which point the team sends it to cert.

"If you want to ship the highest quality product you can, that means making smart decisions, not arbitrary ones."

The triage meetings serve two purposes. The first is to come up with a strategy on how to fix a particular bug. With the heads of The Coalition's multiplayer, single player, production, and technology divisions, plus other critical staff in the room, the team can move quickly to identify what is causing a particular bug and come up with a theory on how to address it.

Read more: A Bug's Life

In one triage meeting, the team watched a video of a tester helplessly stuck behind a locked door. The door was designed to lock behind players as they entered a new area, but after dying in that area, the tester respawned behind the door, which remained in a locked state. The player was trapped.

My assumption prior to attending these triage meetings was that when a development team finds a bug like this, it rewrites the code to reopen the door if the player died. Or, maybe, respawn the player to a point in front of the locked door. It turns out that those solutions require more tinkering with the game than the team likes, as every change in the game could cause new, unforeseen knock-on bugs.

Image: Jackie Dives/Motherboard

Instead, the triage team leaves a note on the bug in the database—The Coalition uses Jira workflow software to track bugs—instructing the relevant staff to address this issue by never setting the door to a locked state in the first place. The locked door will no longer make it clear where the player is not supposed to go, but leaving it open is the easiest, safest fix: no locked door, no problem.

What the average player doesn't realize, but is absolutely critical to the process of making games, is that the second purpose of triage meetings is to identify the bugs the team has no intention of fixing.

Fergusson told me about a bug in the first Gears that covered the mountains in the background of the level in a black texture if the player shot them.

"The team in my office were hand-wringing and saying 'What are we going to do?' and I said 'Punt. No one's shooting the pistol at the background. This is not something we need to worry about, moving on,'" Fergusson said. "The lead programmer looked at me and was like 'What the hell?' But if you want to ship the highest quality product you can, that means making smart decisions, not arbitrary ones."

To that end, punting bugs, or labeling them "wnf" (meaning "will not fix"), is the only way games are finished, and Fergusson is good at this part of the process.

"If you play the game like an asshole there's not much we can do."

The AAA game development cycle can be broken down into four phases. There's the concept phase, where the studio comes up with the idea for the game. There's the pre-production phase, where it creates the prototype and basic art that proves that idea is worth pursuing. At this stage, the publisher (Microsoft, in this case) decides if it wants to take the game into the full production phase, when the bulk of the work of creating it takes place. Last, there's the finaling phase, where the team squashes the last bugs and gets the game in shape for retail.

Most aspiring game developers want to get involved with those concept and pre-production phases, where they get to be creative and come up with ideas. People who actually pursue a career in game development will either learn to appreciate the production phase or burn out, because that's where the bulk of the work takes place. But few people relish the finalizing phase. It's stressful, tedious, and an exhausting way to finish an exhausting few years of development.

This is where Fergusson excels, and why he has a reputation of being "a closer."

Every room at The Coalition has some kind of bug count up on the wall, often with detailed charts that show the number of bugs declining from hundreds to double digits over the course of months, and the number of days left until the team has to submit the game to cert.

"It all comes down to this thing I got from Jim McCarthy, who used to be the head of the C++ group at Microsoft," Fergusson said. "He talks about this idea of 'don't shake the jello.'"

Everyone I talked to at The Coalition has repeated this mantra—"don't shake the jello"— unprompted. When Fergusson first joined the studio in 2014, he got all the producers in a room and showed them one of McCarthy's lectures. The relevant bit starts at the 1:45 mark in this video:

"If your product is a jello mold and what you're trying to do is stop it from vibrating, every time you fix a bug you're more than likely to just create more bugs than the one you just fixed," Fergusson said. "At a certain point, to stabilize, you have to stop shaking the jello. You have to stop touching it."

To do this, the triage team will prioritize bugs from one to five, with one being the most severe category of bugs. As they get closer to deadline, the team will start eliminating categories. They'll look at all the category 5 bugs, and decide which to promote to a category 4, and which to punt, meaning they'll never get fixed. As they get closer to deadline, the team will repeat the same process with category 4, deciding which bugs to promote to a category 3, or punt. The cycle continues until every bug is a category 1, has-to-be-fixed bug, or a bug the team is aware of and is willing to ship without fixing. Whether or not the end user ever notices it, every AAA game goes through a process like this.

"Nobody likes the idea of the triage team getting together, locking the doors for six hours, and looking at 400 bugs," Fergusson said. "It's hard. It's time consuming, but at the end of the day it's worth it because you're making sure people are working on the right things."

Now the triage meeting is reviewing a bug where, after finishing Gears 4, if the player mashed a button incessantly while the credits rolled, it left a small "loading" icon in the corner of the screen when they returned to the main menu, even though the game wasn't loading.

"We can catch a lot of bugs, but we can't catch them all," Arrington, the QA director, said. "Especially with that kind of aggressive user. If you play the game like an asshole there's not much we can do."

Fergusson and the triage team contemplate the bug for less than a minute before they decide to label it "wnf." There are many variables that might prevent you from reproducing this bug yourself, but theoretically it's still in the game that you'll be able to purchase October 11.

"What's next?" Fergusson asked.

5 DAYS TO CERT

BUG COUNT: 2

"MY FAMILY HASN'T SEEN ME IN A LITTLE BIT"

Even die hard fans who have Gears of War tattoos don't know who Max McNiven is. He sits in a corner cubicle on the sixth floor of The Coalition's office. To his left are three other rows of cubicles the QA testers call "the pen," where they test the game's online modes. Some are testing the cooperative Horde Mode 3.0, where four players team up to take on 50 waves of increasingly difficult enemies.

It takes about eight hours to get through all 50 waves on "insane" difficulty, and since some aspects of horde mode will only trigger if no players die, sometimes they'll have to start it all over again. When the team was told at 11 PM last night that they didn't have to run through Horde Mode again, they cheered as they left. But whenever I passed by their desks they seemed in good spirits, calling out dangerous enemies much like players will at home.

McNiven, on the other hand, is quietly going through one of Gears 4's menus.

"The Xbox Store has been my baby," he told me. "It's what I've been focusing on. It's looking great right now."

If you play Gears 4 and buy a character skin through the store and nothing breaks, it's partly because McNiven tested it, over and over again, for months. He gets in at around 8:15 AM. Normally he gets out around 5 PM or 6 PM, but it's crunch time, so today he'll leave at around 9 PM. He's been working these hours for six days a week for "at least a month."

"I'm here all weekend, but that's kind of by choice," he said. "A lot of people are doing that."

That includes people like McNiven's supervisor, QA lead Rob Guwick. He sits behind a wall of monitors hooked up to Xbox Ones. On each screen a character is walking around in circles, looking up at a strange quadrant in the level. Sometimes, all a test requires is running Gears 4 and pushing the joysticks in any direction to prevent the user from going idle, so at the foot of each console is a controller with a rubber band pulling the joysticks together. Even when they're not touching the controls, the QA team is testing Gears 4.

"We've trained people to want to put out the best product that we possibly can," Guwick said. "I'm working tons of hours, it's getting bad, my family hasn't seen me in a little bit. My wife thinks I work way too hard and do more than I should, but that's just who I am. I want to give it my all. That's why it's nice to be part of a project like this."

McNiven at his desk, sneaking a view at an Overwatch match while he works. Image: Jackie Dives/Motherboard

The negative effects of crunch can strain relationships both inside and outside the studio. The most public example of this was the EA_Spouse blog, in which the fiancée of an Electronic Arts programmer chronicled the impact crunch had on their lives. Harsh working conditions at the time eventually prompted a couple of lawsuits from EA employees, which ended in multi-million dollar settlements.

Crunch never stopped, but has been scrutinized ever since. The International Game Developers Association's 2015 developer satisfaction survey found that "crunch is still a problem," with 62 percent of game developers indicating that their job involved crunch time.

There's no rule that says McNiven and Guwick have to work 12-hour days and weekends, but they do. Almost everyone I talked at The Coalition worked the same hours, or more.

"We're still laughing," Fergusson said. "We're not trying to kill each other in the halls. I've been on some pretty crunchy teams where people are pretty angry at each other."

If the pressure caused rifts at The Coalition, the team hid it well. In addition to his usual duties, Arrington spent much of August showing Gears 4 around the world, traveling to China, Germany, and Toronto. When he returned from Toronto and met Fergusson in BigPark Place, I was surprised to see them embrace. Fergusson, Arrington told me later, is his best friend.


Fergusson himself will send an email out every midnight to let the team know the status of the project, and that's not nearly the worst crunch he's been through. When he was finaling Counter-Strike for Xbox in 2003, he spent the last week working 22 hour days, sleeping on an office couch for two hours before going back to work.

Read more: AAA Game Development: A Glossary

That type of crunch, Fergusson said, was the result of a situation The Coalition has been actively trying to avoid by being realistic about the scope of the project from the beginning. That means cutting features and ideas early and often. Based on what I've played of Gears 4, one criticism I suspect game reviewers will raise is that it doesn't do enough new things. It's mostly more of the formula players are familiar with, but on new, modern technology. As Fergusson and other senior staff explained it, building a Gears game from scratch with completely new tools, on a completely new platform, was a challenge that didn't leave them much time to revolutionize that formula.

Even with a pragmatist like Fergusson steering the ship, finaling a game and getting a bug count down from 1,000 to zero is an uncertain process. At this point, The Coalition has been testing the game around the clock for months, and it's still finding bugs it's never seen before. What do you do about that?

"I had a moment where I asked how many of [The Coalition] had not shipped a AAA game before; at least a third of the group raised their hands," Fergusson said. "Either they're young or they're coming from the movie industry or mobile or whatever, and I explained, summer is coming. It's going to be a push because here is where we are and here's where we need to get and there's only a certain amount of time. There's no way to finesse it. You just start working hard. I want you to be prepared for that."

Unlike the EA_Spouse situation, this crunch isn't due to a lack of resources. Gears 4 is one of the most important products that Microsoft will launch this year. The company hopes Gears 4 will motivate players to buy an Xbox One, and inspire other developers to make huge games for Xbox One by seeing what Gears 4 can do. It will launch Gears 4-branded Xbox One bundles the day the game is released. Microsoft spares no expense in the development of its tentpole games.

"Time is a solution, but time is also a problem," Fergusson said. "If we had an extra three months, people would just want to use it to put more features in. Even the most well planned out project will have a flare at the end because you know the door is closing and you're trying to get the most you can."

I believe that's why Fergusson crunches, and I believe that if he were given three more months, McNiven would make that Xbox store the best store you've ever seen. You don't have to be a game developer to feel this way. I spent my days off working on this story, and if I had another week to work on it, I'd take it.

Audio director John Morgan in his soundproof room. Image: Jackie Dives/Motherboard

"North American culture especially is likely to work really hard and not notice they're burning out until it's too late because we normalized the idea that if you care about something, you will give everything to it, until you have nothing left," said Tanya Short of indie developer Kitfox Games. Short also started the "Crunch is Failure" pledge, where developers promise not to overwork themselves or their teams. "I think it's completely natural and kind of beautiful to want to sacrifice yourself to something like this, but it takes a lot of time and experience to figure out when you're doing it for yourself."

I wanted to talk to Fergusson about this, but I had to wait first. He was in a four-hour phone meeting with Xbox's leadership team, discussing the future of the Gears franchise. Next week, he'll have to come into the office at 1AM for a video call to hype a crowd of game store employees in the UK.

Now that he's studio head, he has a lot of new responsibilities that have little to do with actual game development. Fergusson said that every game he's ever shipped went through crunch.

"But nobody sees that word with any sort of grey," he said. "They want a knee jerk reaction, to say you should be working 9-to-5 every day and that's life. It's great if you're working in a factory, it's not so great if you're making art."

4 DAYS TO CERT

BUG COUNT: 8

"WHY DO I KEEP BASHING MY HEAD AGAINST THE WALL?"

"We got 10 pounds of shit and we got to fit it into a five-pound box. How the hell are you going to do that?"

This, Gears 4 creative director Chuck Osieja said, has always been the question at the heart of AAA game development. It was true when he started his career stuffing Sega Genesis games into 512k cartridges, and it's true with Gears 4, which will just barely fit onto a 60-gig Blu-Ray disc.

"Much like your salary, if I give you an extra $60,000 tomorrow, you're going to start living a lifestyle that fills that void," Osieja said. "It's the same way with games. You can create more content, but you're always going to create more content than you can fit on the disc, or want to create more than you have the budget to."

The risky position that AAA games find themselves in today is that the pounds of shit are growing at a higher rate than the box.

We don't have great current and historical data on the cost of AAA game development, but we do know that the vast majority of a game's budget is spent on the people making it, and we do know that fewer than 90 people worked on the first Gears of War in 2006, while around 330 worked on Gears 4. Head of Xbox Phil Spencer told me that Gears 4 is five to six times more expensive than Gears 1.

Patrick Walker, VP of insight and analytics at the video game industry research firm EEDAR, estimates that a AAA game in 2016 costs anywhere between $50 to $200 million. He said the biggest hits in this space can make up to a billion dollars per game, but that overall AAA game revenues haven't kept pace with the cost of development. The margins are getting smaller, which is why Walker thinks that investment in AAA games will eventually crash or flatline.

This is the main reason why we've seen the rise of microtransactions and post-release content that AAA games can sell after launch. Gears 4 does this as well. It sells character skins, cards sponsored by Rockstar Energy Drink that can modify a character's ability, maps, and a $50 "season pass" that offers all that and more. There are two "business intelligence" employees at The Coalition who make sure all of this works smoothly, without annoying players, and they can access Microsoft's data from Halo 5, which sold similar content, to see what works.

Lead campaign producer Zoë Curnoe with the "get shit done" award. Curnoe won it for getting the Gears 4 E3 2016 demo ready in time. Image: Jackie Dives/Motherboard

"At a certain point, the risk is too much," Walker said, adding that, "The revenue made by the top five game publishers on console has stayed the same, from 2008 to 2016."

But during that same time, the cost of AAA game development increased dramatically.

"I think that what it does is it raises the barriers to entry for a new IP [intellectual property]," said Mike Crump, who as director of operations at The Coalition is in charge of Gears 4's budget. "The investment required to make a AAA game, to know whether you got something successful or not, is really high, so you have to have deep pockets and make big bets."

"The analogy I use is that making AAA games is like the Formula 1 of computer science."

In fact, the reason Gears 4 is with Microsoft and not its original creators, Epic Games, is that the latter didn't have deep enough pockets to make it. A company like Microsoft can bet $50-200 million on a game, lose it all, and survive the hit. A smaller company like Epic might not, so it sold the rights and let Microsoft take the risk. Even for Microsoft, Gears 4 is a hedge compared to the completely new shooter The Coalition was working on and shelved in favor of Gears 4. Microsoft has hard data on how many people bought the previous games and how many of them are still playing them online, which makes it easier to project how many copies Gears 4 can sell. That was not the case with the new, unproven game idea The Coalition worked on originally.

Crump told me that the team was sad to let that project go in favor of Gears 4, but that ultimately it was a unanimous decision. Keep in mind that, in a high-risk business like this, working on a franchise like Gears is a relatively safe place to be. It's likely to do well and lead to several more sequels, and since each game takes about three years to make, by taking on the Gears franchise The Coalition essentially signed up for a decade of job security.

It's smaller, independent games like Gone Home and Her Story—made by teams of four and one respectively—that are expanding the definition of what a video game can be. It's games like Flappy Bird, Minecraft, and Pokémon Go that are reaching new audiences and tapping into billions of dollars AAA games can't reach.

I like Gears 4. I understand why the series has a fanatical following of fans that get Gears tattoos, but I also understand why it's everything that people rightfully deride in video games. It's big, loud, and shallow. It's defined by a gun that is also a chainsaw. If that makes people roll their eyes I can't really talk them out of it.

So aside from the rare billion dollar game, why do we still need AAA games?

"You mean why do I keep bashing my head against the wall?" Crump said. "Because it's fun. Frankly, the best people want to work on the best games, and I want to work with the best people."

Technical director Mike Rayner is one of the best people. I'm trying to interview him in his office, but every once in awhile another programmer will pop his head in and ask a technical question I couldn't even begin to understand. The team seems extra busy because there was a bounce—they were at two bugs yesterday, but the QA team found eight bugs in the new build, and now they're only four days away from cert.

"The analogy I use is that making AAA games is like the Formula 1 of computer science," Rayner said. "You're always pushing the bar because that's the expectation and that's great. That's what's exciting about it."

Pushing the bar has benefits that will outlast Gears 4.

Fergusson tracking the bug count. We have blocked out the project codename at Microsoft's request. Image: Jackie Dives/Motherboard

Take, for example, the Tritton audio system The Coalition developed with Microsoft Research, the company's skunkworks-like lab. Many games will change the ambiance to match the environment so that a cavernous space sounds big and echo-y and a small room sounds muffled. Traditionally, this is done with invisible triggers in the environment: once the player crosses a certain threshold, the audio changes. Tritton, as described in this Siggraph paper, models how sound moves through any environment in the game, and creates the appropriate ambiance.

"It's a qualitative evolution of audio that I think you'll start to see in other games because of this paper," Rayner said. "It seems like a natural evolution."

Another example that Rayner couldn't tell me as much about because the technology is still under review had to do with Gears 4's matchmaking. This is the math that insures players of comparable skill end up in the same online games. Matchmaking is a huge, ongoing challenge for competitive multiplayer games, and Rayner believes that here too they will push the standard.

Most importantly, AAA game developers are always investing in the technology that will allow them to keep making games in the future. How do you keep making bigger and bigger games with increasingly complex assets without growing the development team to an unsustainable size?

An emerging solution of the past few years has been outsourcing. Rather than create every 3D model and art asset in-house, game developers will contract outside companies (often located in Asia). The Coalition will give these companies specific style guides and pay them for, say, 10 different burnt-out car models, which is faster and cheaper than building them internally. Any technology that facilitates this workflow will make it easier to make bigger games in the future.

Gears 4 has made strides in this respect as well. It uses a new "geocaching" system unique to The Coalition, which allows it to import visual effects created by the kind of highly specialized studios that make special effects for the latest Star Wars movie.

The amazing tower collapse from the E3 demo, for example, was created by an outside visual effects company, then plugged into the game and replayed in real-time on the Xbox One. By tapping into the technology and talent of blockbuster movies, The Coalition was able to make that scene way more impressive. Other AAA developers use similar methods, but the unique way The Coalition takes advantage of the Xbox One's hardware allows it to create larger scale set-piece destruction moments without slowing down the frame rate.

Whatever advancement The Coalition makes will eventually spread across other AAA developers when they share their findings at events like the Game Developers Conference, and even trickle down to indie developers.

Epic created the first Gears in 2006 in part to entice other developers to license Unreal Engine 3, which it also created. But for years only other AAA game developers could afford it. Today, anyone can make games with it for free in exchange for giving Epic a 5 percent royalty on revenue. Since The Coalition is constantly pushing the Unreal Engine to the limit and is in direct contact with Epic, whatever improvements come out of that relationship will benefit the entire community of developers that use it.

Game studios like The Coalition don't just make games. They're also computer science research labs, and science for the sake of science is worth pursuing precisely because you don't know what you'll discover.

At the same time, the value of science in the name of entertainment will ultimately be judged by a fickle, demanding audience that is continually consuming the most obscene spectacles this industry has to offer. All this work, what The Coalition will accomplish, it doesn't guarantee that Gears 4 will be fun or commercially successful.

"I must have shipped 40 or 60 games, and I've never launched one where I said 'this is a hit,' and I had a number of big hits," Osieja said. "You pour your heart and soul into it and you think you're doing the best job you can, but you don't know until people start buying it and consuming it."

There was one room at The Coalition where I was not allowed. It's a small conference room with a glass wall and sliding door. The glass was covered with large pieces of black cardboard. Occasionally, I would see lead campaign producer Zoë Curnoe or narrative producer Bonnie Jean Mah open the door just wide enough to slide in and out of the room sideways.

Microsoft hasn't announced it, and nobody at The Coalition told me this, but you don't have to be psychic to know that the room is incubating Gears of War 5. The fact that AAA games are so expensive and difficult to make means that they're predictable. The release of Gears 4 and Halo 5 make it more obvious than ever, but this has always been a problem. Gears 5 doesn't exist yet but you can already guess it will feature some bulky dudes in armor who crack wise while they tear monsters in half with chainsaw guns.

That alone doesn't sound like something the entire games industry needs or should aspire to, but since The Coalition is competing on the cutting edge, what it may discover in the process of bringing it into existence always has the potential to change everything.

3 DAYS TO CERT

BUG COUNT: 1

"A BUG ISN'T PRIVATE ANYMORE"

Fergusson is in his office reviewing a submission to Jira.

"That's the craziest bug I've seen in my life," Fergusson says, as if talking to everyone and no one in particular. If a player plants a grenade on a wall early in the game, and then continues to play, four or five levels later, a critical item will spawn where the grenade was planted instead of where it was supposed to. It has something to do with how the game keeps track of items in the world, but as Fergusson said, "it makes no sense."

The original plan was that, by the end of my visit, The Coalition would have a release candidate to submit to cert, and I would get to see them celebrate with a "Beers of War" night, when they buy a bunch of growlers from the bar across the street and drink them on the patio of the seventh floor.

But things don't go as planned.

People are starting to head home and Fergusson is in his office trying to figure out this bug. No one has ever seen it before, meaning The Coalition has to push out a new build of the game, then run it through its QA team, which takes three days. If that build (RC2) holds, the team has just enough time to submit Gears 4 to cert by Sunday night, so Microsoft's certification team in the UK can start working on it on Monday morning.

After spending a week here, I'm not surprised that things didn't go as planned. No single person on the team has a full, perfect picture of the game at any given time. The Coalition spent three years creating an untamable beast of code, art, and design, which they're now trying to cajole onto a Blu-Ray disc without getting hurt in the process. This is a talented, experienced team, but nobody really knows how the entire thing works. Even with Microsoft's resources and Fergusson at the helm, at the end it comes down to a lot of patchwork and tinkering. There's a reason they call it triage.

Image: Jackie Dives/Motherboard

The Coalition, I found out later, met the cert deadline, and after the two-week certification process, Fergusson emailed me to say that Gears 4 maintained his record of passing cert on the first try. The game had "gone gold," and was ready to be pressed on discs.

That doesn't mean the work is done. In addition to the small team incubating the next game, other people are already working on supporting Gears 4 post-launch.

"Back in the day, when it was Counter-Strike Xbox, it was fire and forget," Fergusson said. "I pass cert, ship it, and I'm out. Now it's post-release content, and title updates. It's a services-based thing, and you're getting immediate reactions on Twitter."

"The reality of software development is you never finish a game."

Fergusson said that the studio will lose three days of work the day after release because the entire office will just read reviews, and keep an eye on Reddit and NeoGAF, the influential discussion board frequented by developers, journalists, and players. The immediate reactions also raise the bar for AAA games even higher.

The Coalition celebrating Gears 4 having gone gold. To the right of Fergusson is head of Xbox Phil Spencer, Microsoft CEO Satya Nadella, Windows and Devices head Terry Myerson, and corporate VP of the division Yusuf Mehdi. Gears 4 cuts a line through almost every part of Microsoft's business. Though he is not pictured, I was told that even Microsoft co-founder and board member Bill Gates was there to celebrate. Image: Microsoft.

"That's what made triage so hard now," Fergusson said. "Go back 15 years into that same room and it's a completely different conversation. Take even the grenade bug. 15 years ago someone would see that in their living room and not be able to tell anyone about it. Today, there will be montages of streamers showing their boxes crashing as they're playing. Every potential issue in the game is an attack vector for the whole industry. A bug isn't private anymore. It's out there for the world to see."

As tired as they may be, a lot of people will have to be in the office at 12:01AM, October 7, when the people who preordered the game will start playing and filling the servers. They'll get to see what maps they're playing and if their new matchmaking system works and if people are buying skins. Twitter will tell them right away if there's anything wrong with the game, or if players just don't like it. They'll know within a few days if Gears 4 is on the road to recoup its investment. If it gets a Metacritic score of 9/10 or higher, some people on the team said they would get a Gears of War tattoo, as many fans of the game already have. It would be the second Gears tattoo for Fergusson.

"At the end of the day, the reality of software development is you never finish a game. You abandon it," Fergusson said. "At some point you have to say, I'm letting go."

Get six of our favorite Motherboard stories every day by signing up for our newsletter.