What goes right and wrong when games like Diablo 3 take a decade to make

What does a decade’s worth of work look like? You can start and finish almost any form of education; establish a career; raise a child; build a thriving business. For many, a decade of work could represent a house, a page on a resume, the title on a business card, or something less tangible — a feeling, even.

For some in the video games industry, though, a decade might produce a single product. Ten years of late nights, dashed-off ideas, and half-functional prototypes for a bundle of living, breathing code.

A decadelong development cycle is far from the industry norm, yet high-profile examples abound if you know where to look. The infamous Duke Nukem Forever charted 14 years from its announcement in 1997 to its final release in 2011, with many internal reboots in between. Fumito Ueda’s ambitious The Last Guardian was released more than 10 years after his 2005 classic Shadow of the Colossus, due in part to technical hurdles associated with creating the protagonist’s leonine sidekick, Trico.

Some developers who toil on one game for that long regard their experience with benign bitterness, speaking frankly of the missteps that led to the innumerable delays: the product of lofty expectations, technical hurdles, mismanagement, or a lack of direction. For others — sometimes independent game makers building their dreams in their garages — the sentence can be a source of pride, carried out due to decisions made early in the development process, overly ambitious design docs, or sheer perfectionism.

These are some of the members of the decade club.

Illustration: A game developer hides inside a calendar

“A decade isn’t nearly enough”

Some developers in the indie space don’t see a 10-year development as a stigma, or even a mark of failure. Robert Kurvitz is the lead designer at ZA/UM, the studio developing Disco Elysium, an eccentric computer RPG in the tradition of Baldur’s Gate that seeks to overturn decades of conventional role-playing traditions.

“We are trying to make a new kind of RPG that is almost unrecognizable even to people who grew up on these games,” he says. “That kind of work takes serious time. To me, it takes as long as it takes.” A self-styled provocateur, Kurvitz compares his endeavor — which has taken 15 years so far — to that of one of his literary heroes, the epoch-defining Russian novelist Leo Tolstoy, who famously spent years polishing his major works, like Anna Karenina.

Disco Elysium began life as a tabletop game that Kurvitz and a handful of friends ran as a shared world-building exercise, before transforming it into a video game once they realized they wanted to surpass the limitations of the pen-and-paper space. Kurvitz admits that his team’s approach to game development runs counter to what most studios consider the central tenets of the field. But, as he puts it, most studios are trying to improve slightly upon their heroes, make a minor contribution to a well-trod genre, or merely keep a giant company from falling to pieces. He likens the development process of a big-budget game to a war — both sides draw their lines, the stakeholders create a blueprint, and the machine begins to chug forward, loudly and publicly, and rarely according to plan. To Kurvitz, a game like Disco Elysium is more akin to a revolutionary conflict, with ZA/UM taking the side of the guerrilla fighters ducking in the jungle.

The team has had a lot more freedom over the course of Disco Elysium’s development cycle by charting its own path, eschewing external funding to tinker about with mechanics in the garage. But now that it’s ramped up production, which necessitated taking money from investors to build out the game, the team is a lot more vulnerable than the hypothetical war machines. “Once you’re out of the basement, running out of money becomes a serious concern,” Kurvitz says with a laugh.

At first, as with most pen-and-paper experiences, Disco Elysium took place in a fantasy world that the developers built out as they played. A year or two in, however, they came to an upsetting conclusion. “At a certain point, we just realized that the name of every city in the world sounded really stupid to say,” Kurvitz says. “That’s when we decided to move up to a more steampunk-ish setting, and that’s a major part of why that’s the setting of Disco.”

Though ZA/UM has been very cagey about the game’s details so far, even 15 years into development, this is clear: Disco Elysium casts you as a police officer in a decrepit city teetering on the verge of collapse, investigating odd crimes including grisly murders. According to Kurvitz, the game’s layered world and “fail-forward” approach to conflict resolution — inspired by design trends in the tabletop space that allow players to continue to contribute to the story even when they lose a fight or fumble a task — is what sets it apart from the raft of CRPG-style fare that has flooded the market in recent years. Unlike in most video games, absolute control over your player character isn’t a given. Your “willpower” is a stat, and when the invisible dice roll below the target number you need to succeed, you might do something a bit embarrassing, such as smoking an entire carton of cigarettes or licking a liquor-stained carpet.

Your character’s build doesn’t just determine your capability to succeed in the choices you make; it determines what choices you can make, or how you perceive your environment in the first place. For example, if you’re big and burly with a high “Fysique,” you’re more likely to view violence as the only solution to a problem, while a more intellectually inclined character might lack empathy or self-control. This all goes back to Kurvitz’s stated goal of making player characters feel less like avatars for wish fulfillment and more like actual human beings, with their own thoughts, feelings, and limitations.

To him, the task of trying to move the genre in a less escapist direction is a task of decades, not mere years. “I think with billions of people on the planet, to some degree, if a thing is possible for humans to do, we will eventually do it,” he says. “In the past, when the technology used to make games was less developed, it was possible to make a smaller thing that changed things, or moved us forward. Now, with so many people making games, I think it takes many years to make something that is actually revolutionary, which is what we’re trying to do. Both the spear itself and the act of throwing it must be honed to such a fine edge. If I were to make a sequel to Disco, that would seem to be a 20-year project or more, to me. Soon enough, I believe art will move in a direction of works that span generations. It is not just a hypothetical.”

Digging down deep

While it’s clear that a certain breed of hyper-obsessive auteur will continue to push the limits of capital and even their natural life span with a labor of love like Disco Elysium, some indie developers who worked full time for nearly a decade with no funding have a different view: never again. Just ask Jo-Remi Madsen and Simon Stafsnes Andersen, the duo behind the intricate and beautiful 2D platformer Owlboy, which took nine years and change to find its way to Steam, and jumped over the decade line before it hit consoles.

Development on Owlboy might have officially commenced in 2007, but Andersen notes that he was working on prototypes of the game’s flight mechanic with a team of unpaid hobbyists before even the original Wii was revealed to have motion controls in late 2005. When several members of that team decided to drop out to pursue full-time jobs in the industry, Andersen decided to stick with the XNA-powered engine that his programmer friend had put together for the game. After a chance meeting with Madsen at an industry event in their native Norway, they started to work on the game in earnest. “I was on the verge of making a major life change, moving abroad for a traditional development job,” says Madsen. “But I just fell in love with Simon’s art, so I decided, ‘This is what I want to do with my life.’ Of course, I didn’t know it would take nine more years at that point.”

Both developers say they’re tremendously proud of what they accomplished with their decadelong endeavor, but they admit that they’ve learned vital lessons that they’ll apply to their next project. Andersen ascribes Owlboy’s titanic time in the pipe to several key decisions that the duo made early in development, most of which are common for first-time indie studios. For one, continuing to push a game engine that a talented but nonprofessional programmer put together in a year proved to be a source of constant anxiety, necessitating endless adjustments and fixes. (At one point, Andersen’s brother Henrik was brought onto the project as a second programmer to help ameliorate these issues.) For another, the duo say that they naively let their anxiety over fan expectations have too much of an impact on how they put things together. When they released the first demo for the game in 2011, the deluge of positive feedback overwhelmed the team.

“What we basically heard was that people expected the game to be about 18 times longer than we thought it was going to be,” Andersen says. “And that was definitely an ‘oh, shit, back to the drawing board’ moment. We decided to completely rethink the basic structure of the game.”

Having blown up the backbone of their project, the duo then worked 12- to 15-hour days for the next six years in order to make the game they thought people wanted, moving into the home of Madsen’s parents in order to save money. As the years passed, the team found itself having to make surprisingly tough decisions in order to make the project feasible, even on its macro time frame. When the pair started testing builds on modern platforms in 2014, they realized that their art assets looked stretched out and discolored in the then-standard widescreen format — “[Owlboy protagonist] Otus looked like he had some sort of disease,” Madsen says, laughing — so they had to spend several months figuring out how to display Andersen’s intricate pixel work in a different aspect ratio than the one they had developed in.

There were some decisions that were more consequential. Andersen spent several months designing characters and architecture for a city that was supposed to serve as the game’s hub world. Before completing it, however, the duo decided that traversing a large location would detract from the kinetic gameplay that they saw as the game’s main strength, and it would take too many man-hours to put together anyway, so they decided to drop it as a core beat of the main plot. “When we talk to fans, they always say how much they love that part, how unexpected it was,” Andersen says. “Well, that’s great, but that reaction is very unexpected to me. We came to that decision after looking at it and just saying, ‘It would add another year to the project to make this the way we want it, and it’s not really fun anyway, so let’s just blow it up.’ Clearly, though, it worked.”

Since the game’s initial release in late 2016, Owlboy has sold well on several platforms. While Andersen and Madsen say that they appreciate how the success of the game means that they don’t have to immediately start crunching on a new project, the process of returning to a normal life after working on the same title for a decade is challenging in itself. In Andersen’s experience, when you work on a project for that long, you start to fantasize about the blissful moment when you’ll be able to release it into the world and soak up the praise of all the people you admire. The reality is far less invigorating: “You imagine that moment again and again, and there was one where everybody on the team was crying, but then it passes, and you’re looking at sales projections, bug reports, and marketing emails,” Andersen says. “That’s the reality of releasing a game. You never get off the treadmill. […] I know this sounds a bit dramatic, but it’s basically the process of learning to be a human again. How to manage a team where you’re not working every waking hour, how to go back to normal living conditions, how to work a healthy number of hours per day. It’s a learning process.”

His advice to fellow game developers is familiar and honest: Focus on the process, not the finished product. “The two forms of reward that you get as an artist are the joy that you get from working on a project, and the feedback that you get from your peers when you finally release it,” Andersen says. “If you have the first, you’re set for life, because that’ll always be there. The second is never exactly what you expect it to be. Even if you release something that people love, there will still be people who hate it, and you have to deal with that. I’ve had depression since I was 10 years old, and though my dream was releasing a game like Owlboy, shipping a game doesn’t fix that. It’s actually gotten worse. So I try to enjoy the process, because that’s what’s always there.”

Illustration: A piggy bank that has been smashed and repaired many times

Cogs in the machine

Of course, it can be harder to enjoy the process on the big-budget side of the industry, where lengthy development times are more likely to be the result of sky-high expectations, shifting resources, a lack of direction, or just mismanagement. That’s not to say that the games produced by this process are necessarily subpar, but the experience of working on them can burn out even battle-tested veterans. This pressure can be especially intense when trying to craft the sequel to an epoch-defining game like Blizzard’s Diablo 2, which pioneered the colored-loot concept that has blossomed to encompass every genre. According to a former senior-level staffer at Blizzard who worked on Diablo 3 for most of its 11-year development cycle — who requested anonymity for the sake of speaking frankly — the immense legacy of its predecessor overshadowed every aspect of Diablo 3’s production.

As with many big-budget follow-ups that linger in “development hell” for a period of many years, one of the central problems with Diablo 3 was that the team had differing ideas of what the franchise actually stood for. Rather than coalescing around a central idea or aesthetic, senior staff kept second-guessing themselves, which meant that the rest of the team had to spin its wheels by making several different versions of the game, all of which “dead-ended,” in the words of the anonymous staffer. Since the first two entries in the series focused on combat over storytelling, there was a lot of room to build out the world beyond the cast of demons that made up its core mythology. But some on the team resisted that urge, saying that it ran counter to their vision of the franchise.

The enemy designers would put love and care into a terrifying insect monster, only to hear back from more senior members on the team that since Diablo 2 lacked enemies like that, they should go with more traditionally “demonic” designs for Diablo 3. Some staffers also resisted attempts to inject some levity into the game’s environments, saying that Diablo should stick to the grimdark environs that made it famous. “I think if we’d listened to every ‘that’s not Diablo’ that was said, there’d be a very small, very bland game left over,” says the staffer. “If Diablo 2 is a perfect game, with no faults, then it was a mistake to make a sequel.”

Upon release, Diablo 3 was immediately engulfed in a major controversy surrounding players’ ability to buy and sell their hard-earned loot using real-life currency, which echoed throughout the gaming world. That cloud eventually cleared when Blizzard removed that functionality, and players found that Diablo 3 hewed far closer to its famed predecessor than even the fans expected, which made for a mostly positive response that has only increased in pitch to this day. To the staffer, this indicates that the essential aspect of a successful sequel isn’t what they term a “sexy new mechanic”; rather, it’s replicating what made the game such a hit in the first place.

“People try to sell sequels on big new features, but none of the big new features in Diablo 3 were the best things about it,” the staffer says. “At the end of the day, Diablo is beating up monsters and collecting loot. The thing 3 did right was combat. That’s 99% of the game, and I’m very proud of how it turned out.”

While the staffer describes their time at Blizzard as “a privilege,” they personally feel that firmer deadlines would’ve likely improved the experience of making Diablo 3 for the entire team, to cut out the endless second-guessing that handicapped the project from the start. “I think ‘when it’s done’ is a great commitment, but it’s a terrible way to run a design team,” they say. While they’re adamant that Diablo 3’s lengthy development time wasn’t the sole reason that they left the company, the staffer admits that it was a significant factor in that decision. “Being a major part of Diablo felt like managing an empire,” they say.


In the world of commercial games like Diablo 3, since developers are either collecting a salary or living off a budget to produce their work, the decision to keep working on a project in deep water is often motivated just as much by material needs just as much as artistic ambitions. As Owlboy’s Andersen puts it: “Once we hit the four-year-mark, it was like, ‘We’re too far in.’ We never considered turning back. It was either ship this game, or work at a grocery store.” When it comes to the somewhat underground realm of free fan games, however, the will to continue has to come from elsewhere, at least according to Milton “DoctorM64” Guasti and his collaborator Steve “sabre230” Rothlisberger, who both garnered a measure of internet fame for working on a remake to the Game Boy title Metroid 2: Return of Samus for over a decade, dubbed AM2R.

Though Guasti is careful to note that he primarily worked on AM2R in his spare time while trying to start up a studio in his native Argentina, he readily admits that the fan game was a shelter from the slings and arrows that life threw at him during its long development, which started around 2005. When he released the first demo of the game in 2011, he was absolutely flabbergasted at the positive response, which contrasted strongly with his own efforts to get original indie projects off the ground in a country that didn’t have much in the way of a game industry. After posting the demo to a Metroid fan forum, he was flooded by requests from talented enthusiasts who wanted to contribute sprite work and design ideas to the game. When Guasti started to make requests of his newfound community, which included Rothlisberger, he was stunned — not only that they would fill them, but at the quality of their output. “I never even expected people to help me with it,” Guasti says. “All of a sudden, it felt like an actual game.”

Guasti admits that the lure of the gold rush in the burgeoning indie scene in the early 2010s caused him to consider abandoning the fan game entirely at several points in development — especially once he left full-time game development to pursue a junior programming job at a software company after the birth of his first child left him strapped for cash. But when he considered the amount of effort that his collaborators were putting in, he decided that it was worth seeing it through, even given the possibility that Nintendo might threaten legal action against the game. “It felt like a possible success when every other thing I had worked on up to that point hadn’t worked out,” he says. “I felt like I was teaching myself how to make a game, and learning how to manage a team better because of my programming job. I was meeting a lot of awesome people, so I was hoping it would work out once the game was released. That was my thought process.”

When AM2R was finally released in 2016, it got rave reviews from a variety of gaming outlets, which raised its profile. Unfortunately, Nintendo cracked down on the game later that week, sending a cease-and-desist order that forced Guasti to make the tough decision to end development prematurely. Though he says it devastated him at the time, he adds that in some twisted way, it was a blessing in disguise. Free of the weight of his exceedingly ambitious fan game, he was able to pursue employment at notable gaming studios, landing a job at Moon Studios, the all-remote team behind indie hit Ori and the Blind Forest. “It’s like a tragedy that you’re waiting on for so long, the best thing is for it to just happen,” he says. “We were able to fix pretty much all of the major bugs, and some people are still working on a fork of the project. I wish them luck. It was 10 years of my life, and it’s out there in some form, and that’s what matters to me.”

Though Guasti says that the era of fan games is over, and that he’d probably never work on a project for that long again, he says he doesn’t regret the decade-plus that he spent on AM2R. That said, he wants to make one thing clear: It takes true dedication to put yourself through something like that. Like the minds behind Owlboy, he views it as a sort of crash course in game development that he foisted upon himself, and he’s not sure what his place in the industry would be without it.

“I wanted to become a game developer, and I didn’t know how to do it,” he says. “Well, as it turns out, when you aren’t working on a project, it’s almost impossible to teach yourself the skills. I was becoming a good manager, a good designer, a good programmer, and I didn’t even know it. It hurt, and it was hard, but it was worth it for me. Is it worth it for everybody else? I don’t think so. […] For those people who work on commercial projects for that long, I almost think it’s a form of torture. When you do it to yourself, that’s one thing. But when they do it to you, it seems unbearable.”

Source: Polygon.com

Leave a Reply

Your email address will not be published.