Epic’s Rainmakers

We Are the Dreamers of Dreams
Tim Sweeney’s office is austere. He has two unassuming desks housing two computers each. There’s a well-traveled couch on one wall and a painting leaning against an eight-foot window. The only decoration to speak of is a Ferrari standard, pinned up opposite Sweeney’s main workspace.

“You’re a Ferrari guy?” I ask. “I didn’t see you car outside.” Epic’s parking lot resembles a luxury car dealership. I saw two Lamborghinis, two Hummer H2s, one H3, a vintage VW bug, three Porsches, a Corvette and a number of sport sedans that made me vaguely ashamed of the financed four-banger I bought at CarMax. I had to look for a car I wouldn’t drive. It took me five minutes to find one.

“Oh, I didn’t bring the Ferrari today,” Sweeney says. “I brought the Lamborghini.” He’s so matter-of-fact I don’t immediately register that the man spent half a million dollars on 1,000 horsepower and Italian metal, and that every day he must choose which machine to drive to work. Looking at him, you wouldn’t peg him as a guy who hits 160 mph on North Carolina’s back roads, but here he is anyway, telling me about speed traps and horsepower-to-weight ratios and figuring in his head if my motorcycle could take his car off the line.

Sweeney is Epic’s founder and CEO, and in many ways he is the company’s Willy Wonka. There are stories of brilliant people locking their doors for months on end and emerging with never-before-seen wonders. Sweeney is one of those people. Much of the Unreal Engine’s technology can be traced back directly to him, tinkering away in his office.

“Epic began as my scam to avoid having to get a real job out of college,” Sweeney says. He attended the University of Maryland with a degree in mechanical engineering, but lost interest in the field. He’d always had a knack for programming, and in 1991, his third year at school, he decided to dedicate six months to building his first commercially released game, a text adventure with a built-in level editor called ZZT.

He released ZZT as shareware, an “experiment to see if I could make some money from that. It was making about $100 a day in shareware registrations.” His next game, a Mario Bros.-style platformer called Jill of the Jungle, took him nine months to create and brought in $30,000 a month at the game’s peak. From there, a business was born.

He kept the company loose originally, structuring it in Apogee’s shareware publishing image. He invited developers to submit their games to Epic. He’d give them the official stamp and would use his experience with ZZT and Jill to sell them nationally. “A lot of the early, core Epic folks contacted us,” including Cliff Bleszinski and James Schmalz, Unreal‘s lead designers. “My role transitioned from programmer to cheerleader at that point, helping people along their projects … and running the business.”

At the time, Epic employed about 10 people, many of whom were tasked with taking shareware orders over the phone and shipping games to customers. In 1992, Mark Rein joined the company to lend more of a business perspective to the operation, at which point they began establishing retail channels to sell the games they were publishing. And by 1994, thanks to the success of Bleszinski’s Jazz Jackrabbit and Schmalz’s Epic Pinball, Epic was a multimillion-dollar publisher. Unfortunately, their core business, snailmail software sales, was about to be rendered obsolete by the internet.

“It was clear that the shareware business model wasn’t going to hold up and sustain growth much longer,” Sweeney says. After the internet grew popular, “it wasn’t just a couple of shareware companies putting out huge games every once in a while; now there were a whole lot of companies getting into it. And also every retail developer was putting out demos of their games, which directly competed with shareware, too.” It was then Sweeney decided the company needed to develop a more robust game, one that could compete in the new environment, and he set to work with Bleszinski and Schmalz to create Unreal.

Though the three had no previous 3-D development experience, they opted to make a shooter using their own technology, which Sweeney personally developed. He says he employed a lot of what he learned from working on his first game, ZZT. Part of what drove ZZT‘s early sales was its built-in level editor, which allowed players to continually add content to the game. Sweeney actually designed the level editor and added the game’s content like a modder would. “A lot of the concepts that really were key to Unreal were actually developed back then in that little game,” he says. “The game with a built-in level editor with complete moddability, a built-in scripting language … all those trends went into the architecture of Unreal.”

It ultimately took Sweeney and crew three years to release the first Unreal, built on the company’s Unreal Engine. The game itself was a hit, and they enjoyed modest success licensing the engine, but it wasn’t until the company integrated a traditional management structure into their operations that they became rainmakers in the game industry.

The Jazz Band Meets The Military
Contrary to Sweeney’s office, the place where Dr. Mike Capps does business is actually befitting of a president of a multi-million dollar corporation. He has a large desk that faces the door, huge windows overlooking the company’s outdoor basketball court and a sitting area adorned by leather furniture, all directed at an HDTV. It’s tasteful but modern, understated but elegant, like a psychiatrist’s office mixed with a war room.

Capps’ story begins at Duke University, where he was studying to become a flight surgeon in the Air Force. His scholarships dried up during the Gulf War, so he left Duke to pursue a math/creative writing double major at the University of North Carolina at Chapel Hill, “which was cheaper, and the girls were nicer,” he says. He then obtained a master’s degree in computer science, and eventually found his way to MIT to earn another master’s in electrical engineering. He found his way back to the military by way of the Naval Postgraduate School, where he was approached to create America’s Army, the military’s free-to-play multiplayer combat game/recruiting tool.

Capps jumped at the chance to create the game, but the enormity of the project meant he’d need to use licensed technology to build most of the title.

“We licensed the Epic tech,” he says. “I was the designer, producer and lead programmer of that game, which was a really fun two years, as you might imagine, and I was still teaching at the time.”

Once America’s Army launched and gamers devoured it – the game was downloaded over 30 million times in the first year of its launch – Epic approached Capps about forming a new subsidiary company called Scion.

“They wanted to do the Gears of War thing, and they knew they couldn’t do two things at once, since they were sorta disorganized,” he says, “so they wanted to shop Unreal out, but their experience had been not so great. With UT 2003 [sales] went OK but not great, and Unreal 2 was dragging along, and so they said, ‘Maybe we could have a studio make Unreal right down the hall from us.’

“So I showed up right in the middle of crunch for Unreal Tournament 2003 and helped ship that. Then I built up a team and made Unreal Championship 2, and as we were making it, Epic was moving to this new building, and they said, ‘Gosh, it kinda sucks we’re going to be moving away from you guys,’ because there was a lot of synergy back and forth between our two companies. So we decided to merge the two shops. My management team kinda took over the combined company. We were the organized, military-hierarchy approach to game development, and they were the jazz band.”

As a result of the merger, Capps was named president of Epic, and currently runs the company’s corporate day-to-day. “Some days, I’m just a switchboard operator,” he says. “Mails come in, and I point them to the right people. So I get a mail from Midway saying, ‘Oh, God, we’ve got this marketing thing we need to figure out right away!’ and I’ll say, ‘Ah! You’re wanting to talk to Jeff and so and so,’ then it comes back and I kind of put the stamp on it.

“I don’t contribute much to the development anymore. I still write. I wrote Unreal Tournament. I wrote all the scripts for that. And I do a lot of our licensing work. I’m working with the comics, and the novels, and the action figure folks, and stuff like that. But really my job is to make sure everything’s moving smoothly. … I’m sort of the final arbiter of ‘Yes, we’re going to do it,’ or ‘No.'”

Capps feels he plays the role of bad cop. It’s not a role he likes, but he feels it’s necessary to the company’s success. “Tim Sweeney … is the sweetest guy on the planet. He does not want to be the one who tells someone, ‘Hey, that’s three days in a row that you’ve shown up at 3:00 in the afternoon drunk.'”

“I’ve never been a great manager,” Sweeney confirms.

“That’s something that was needed here,” Capps continues. “There was not a lot of structure at Epic, and I’m pretty proud of adding structure.”

That structure has certainly helped the company’s bottom line. Though they adapted to their role as a AAA developer rather well, Capps’ leadership has allowed their engine business take off. Since he joined the team in 2002, the company has licensed the Unreal Engine 3 platform for “hundreds” of games and is the de facto middleware solution on both the Xbox 360 and the PS3.

Capps’ bottom line for his time at Epic: “I didn’t screw it up, and I made them able to do more.”

Church and State
One thing I’ve always wondered about the big engine shops, but Epic in particular, is what role their games play as engine test-drives for other developers. Naturally, the first game using the latest Unreal Engine is a game made by Epic, but does the team consider their title something more than a 40-hour tech demo? Is Epic a game developer or a middleware developer? How much does church and state intermingle?

“The last thing I’m gonna do is ship crappy foliage,” Capps says, “because then one of our other competitors has a hallmark of, ‘Look, ours is better than theirs.’ … So we think of that sometimes. We should make sure this feature is better than what id Tech just showed, or what Valve just showed. I feel like we do pretty well for that. But that’s the engine team trying to make features better, not the game team being told, ‘OK, guys, there’s not enough ragdoll physics in this level. We need more in order to be competitive with Crytek.’ That’s not how we do it.

“We are probably more competitive with games than we are with engines. I’m more worried about Brothers in Arms coming out with a cool feature that looks great and makes our game look crappy. That’s more concerning to me than what id Tech or what Crytek does. We’re game developers first.”

Tim Sweeney Plays Dice with the Universe
The game engine business is a harrowing one. In order to remain competitive, you have to begin designing an engine around technology that doesn’t yet exist. It’s a form of gambling at which Sweeney excels. He dedicates much of his time to developing new versions of Unreal, and right now is trying to read the minds of Microsoft and Sony to figure out how their consoles will work in 2012.

He says determining what to update within the engine reduces risk. A lot of what’s in the Unreal Engine 3, like the network code, likely won’t change much in Unreal 4. “Everybody’s not gonna have multi-gigabit broadband connections,” he says. But “things like graphics engines and things like support for multi-threading will be completely replaced.”

From an outside perspective, that makes sense. My cable internet connection has been getting progressively better, but even if the rate at which my transfer speed were to double every year for four years, I’d have a 96-megabit transfer rate. That’s a lot more data, but from a layman’s perspective, the fact data can move faster doesn’t change the way you move it. But in the past five years, I’ve mastered, lost, re-mastered and re-lost the secrets of graphics card technology. It’s a safe bet there will be another revolution in time for Unreal 4.

Sweeney also says the prognostication isn’t entirely original. “You look at all these external references and see what you can ‘steal’ from them. What other game developers are doing is one thing, but more importantly there’s this wealth of information of what people were doing with offline editors 10 and 15 years ago. And all those techniques are now possible to do in real-time. … Often when you’re developing a rendering technique, you’ll find, ‘Oh, this was actually developed in the late 1980s!’ … Now you can do it at 60 frames per second.”

But you still can’t eliminate all risk. Sweeney details one of Epic’s early gambles with the Unreal Engine 3. The original design specifications for the PS3 indicated Sony planned on just adding more PS2 graphics processors into an array, effectively speeding up the PS2 but not taking advantage of multi-core processing like the Xbox 360 would. Sweeney didn’t think Sony would follow through with the plan when Microsoft intended to set the processing bar so much higher, and guessed (rightly) that Sony would change their design to something more similar to the 360’s. Had Sweeney guessed wrong, they’d have gone back to formula on the PS3 and would’ve missed its launch window by years.

Controlling the Future by Mastering the Present
When I ask Capps why Epic has been so successful – they currently have a multi-game deal with Midway, appear to be beating the tar out of other engine shops on current-gen consoles and are about to release the sequel to one of the best-selling games on the Xbox 360 – his answer is wry.

“We have three people here who practice the dark arts,” he says, laughing, “and we’ve got a number of folks who are devout Christians. So we figure we’re playing both red and black. I don’t know. That is … something we constantly fear fucking up. Independent studios have had a really hard time lately. … One bad game is basically all it takes to kill most independent companies. We’re lucky at this point that the engine is sort of that thing that keeps us going. … We could have a really bad-selling game, and it would hurt, but we’d make it. We couldn’t have two; that would kill us. But we could have one, which is nice to be in that position.”

Sweeney believes it harkens back to the company’s original, customer-centric days. “We started out in the early shareware business model. When you release your game and only make money if people like it enough to buy more, which is really very different from retail. With retail, you can make a pretty crappy game, and if you have it in a pretty enough box and you do enough marketing, you can actually make a profit. It wasn’t like that for us at all. We’d only make money by building great games.”

As for what’s to come, Capps believes Epic’s future is bright. “World domination has been Tim’s plan for a long time. This seems like a weird path to take, I know,” he says.”We love where we are; we love wagging the dog. We’re a little, tiny company, and we help architect consoles and what’s going to be in them. We’ve been driving high-end graphics and processors for a long time. … Intel calls us to see what we need three years from now.

“I wanna make Gears 2 and 3, 4, 5 if that’s what we wanna do and we think it’s gonna be fun, and keep making cutting-edge, mature, really cool games that our guys love to make, because that passion comes through.”

Sweeney feels the same way. He believes their dominance in the current-gen console market means they’ll be able to influence how the next version of DirectX will handle gaming-related applications, given how many developers make their games using Epic’s technology. “I guess the best way to control the future is to have a say in how it will play out.”

As I leave his office, I look back and Sweeney’s hunched over a tiny laptop, peering into line after line of code and typing furiously. He’s back in his factory and preparing his latest masterpiece for public consumption. Maybe this one will buy him an Aston Martin.

Joe Blancato is an Associate Editor at The Escapist. He remains unconvinced his motorcycle could take Tim Sweeney’s Lamborghini off the line, but he’s more than willing to try.