ANDREW SITES WAS freaking out. He was in his late teens, enrolled in college, and had stumbled into the job of a lifetime: An entry-level position at a video game studio. Determined to make it work, he cut his class schedule to part-time. One semester later, he dropped out.

His father was less than thrilled. “He was immediately like, ‘Wait a second. You're dropping classes to do what? What are you going to be, a Game Boy repairman or something?’ We joke about that to this day, because obviously it worked out well.”

Sites was excited, but levelheaded enough to formulate a plan. He would spend one semester working full-time at his job, a QA (quality assurance) tester paid to put PlayStation games through their paces. Then he’d go back to school and continue working part-time.

One semester passed. Sites didn’t sign up for classes.

By 1997, he’d advanced to lead QA tester when he learned of a job opening at 989 Studios working on EverQuest. He applied and joined the team that year as an associate producer. “It was an associate producer position, but really it was a ‘do everything you can to help the team’ because there were only 20 people on the project. I came over and was doing design work, content population, managing the build processes and servers, stuff like that.”

EverQuest entered beta, a state of production when the development team is ready for feedback from players, in November 1997. Sites jumped in with both feet, ordering dinner for the programmers and artists, working on pipelines, making sure zones were exporting from the computers of artists and designers into the game where they could be played.

“I wouldn't say I was living at work because I didn't sleep there at night, but I did spend every waking moment there. From the moment I got up, I'd go right in, and I'd get home really late at night. I did that for a year and a half straight up until launch.”

To Sites, every task was a task worth doing. At his previous job, he’d tested games others made. On EQ, he was part of the crew creating something from nothing. The work was never-ending, like the game itself, but he found himself among friends.

“I was the young guy. I turned 21, and of course they had to help me celebrate that. We got up to all kinds of shenanigans. It was a great experience.”

One of his most important tasks was making sure anyone who wanted to play the beta could access it. “I ran the beta program, which early on was me burning discs. We worked for 989 Studios, Sony Interactive, so we had a lot of disc-replication machines for all the PlayStation One testing. I would make copies of our beta, hand-write Beta: Phase 1, and another producer and I would write out all these envelopes and mail them out.”

When developers weren’t building content or burning and shipping out CD-ROMs, they were playing the beta alongside players. Brandan McDonald, one of the game’s artists, had fun dropping breadcrumbs to help players find their way through areas and quests. “I'd never give everything away, but I'd point them in the right direction.”

His hints were invaluable. At that stage, parts of EverQuest played similarly to text-driven adventure games like Sierra’s early King’s Quest titles: NPCs would talk in a stream of text, and players would have to pick out certain portions and reply to get more information. “I thought that was super-hard, but it was also cool because you had to pay more attention to the game, which made it more immersive and realistic. I really liked giving people direction, and just walking around and listening—so to speak—to what groups were talking about.”

McDonald also reveled in playing superhero. One zone consisted of a chessboard-like surface where skeletons and ogres dwelled. Players would stray too deep into it and end up fleeing from packs of monsters. At that point, McDonald would swoop in like Batman to save the day. “I was figuring out the game and having a good time playing it. It really came alive. It was awesome.”

All the developers were finding that EverQuest played quite differently in the beta than it had in smaller-scale, internal-only tests. They had built it, but its massive size and influx of players in the test made it unpredictable, like a living game of Dungeons & Dragons. On one occasion, McDonald observed players stocking up on rare supplies and posting up at the entrance to tough zones. When other adventurers wandered by, the players-turned-merchants would step out and offer to sell their goods.

“People were just so helpful. It was really cool. They made their own marketplaces. It was totally player-guided: everyone would gather and sell stuff. I was dumbfounded that the [community] had its own economy.”

Right then, McDonald knew that EverQuest had the chance to be a success based solely on the power of a community who seemed willing to buy into its fantasy. “All these gamers are in this fantastic fantasy game they've always wanted to play, or never knew they wanted to play, and everyone's having a great time. I think that's what made the game stick.”

“I suppose, looking back, you'd think, Jeez, what were we thinking?” added Kevin Lydy, fellow artist. “That happened a lot. It was the case when I got into the arcade business: It was the Wild West. There were no rules. Nobody knew how things worked or what you should do. You just made some games, and hopefully they worked out. Working without a net was constant.”

Player feedback was only one source of suggestions. The launch of Ultima Online was another. A long-running series of RPGs created by Richard Garriott, Ultima evolved into a massively multiplayer online RPG (MMORPG) that swept players into its world. It was graphical, like EverQuest, but 2D, and displayed from an isometric, overhead view.

“I fell in love with that game so hard,” said John Smedley of UO. “It's one of the best MMOs ever made because of the freedom it gave you.”

Smed exercised UO’s freedom by going to the dark side and PK-ing, or player-killing, other adventurers. PK-ing metastasized, and while some of EQ’s developers dabbled in it, most agreed that their game should do the opposite. Where UO was becoming an outlaw territory unsafe for newbies, EQ would promote cooperation and camaraderie.

Ultima Online also showed them another path to retail success. Originally, Smed had planned to ship EverQuest as a traditional boxed product. UO was a subscription-based game, where players forked over a monthly fee to log in and access their characters.

“We realized, ‘Wait a minute—this could be a service. It could be this thing that goes on and on and on,’” said Smed.

“Nowadays we think nothing of subscription games, but back then, it was pretty new,” added Linda Carlson, who started out as a player before joining the EQ dev team later on as global director of community.

Before launch, the developers landed on a subscription of $9.89 per month. “The weird part to me was the fact that the cost of living has gone up,” Carlson continued. “Nine dollars and eighty-nine cents was named after 989 Studios, which was very clever. But even though it's since raised to around $15 a month, even though everything else has changed over time, you still get a month's worth of entertainment for that 15 dollars.”

“It was funny because it validated the business model of building our game in the first place, showing us a new way to go with subscriptions,” Smed went on. “I still love the subscription model. I think I prefer free-to-play now because it's the purest form of capitalism I've ever seen: You made this thing, it cost us fifty million dollars—hope you like it! That's a lot different than selling boxed products, which has nearly gone away. But if you get this game for free and don't like it, you just stop playing it.”

BRAD McQUAID FELT like a gnome who’d taken a wrong turn and ended up in Grobb, home city of trolls.

It was 1998, and McQuaid and a few other developers were attending the annual Game Developers Conference, a place where developers gave talks to peers about the technologies they were developing and reflected on past successes, as well as showed off upcoming games to attendees. The team occupied a booth equipped with 10 computers arranged in two rows. All ran EverQuest.

As soon as the show floor opened, attendees rushed the booth, queueing up to wait for their turn to play while 10 lucky players explored the zone of Qeynos—“Sony EQ” spelled backwards—and Qeynos Hills, the only two zones available in the demo. The crowds never thinned.

“We had to ask some people, ‘Hey, you've been here 30 minutes. Can you please let someone else play?’” McQuaid recalled.

Players had to be watched. Whether they quit playing of their own volition or had to be firmly but politely asked to let the next guest in line have a turn, attendees had a habit of walking away only to creep over to another terminal.

At first, McQuaid was annoyed at having to police the game. He didn’t want anyone waiting in line to get upset. “Then it occurred to me: This is actually amazing. People are ready to fight to get in 20 minutes of EverQuest. They loved it. I would consider that a turning point in our confidence, and that we were really on to something.”

For Andy Sites, the experience of developing a game grew more surreal by the day. Some of the players he was befriending in EQ’s beta were developers themselves, and hailed from popular studios such as WarCraft developer Blizzard Entertainment and Diablo creator Blizzard North. “They'd play the game and email me, saying, ‘Man, this is amazing.’ It was funny because all I could see were the rough edges, all the holes we were trying to plug. But other developers were very complimentary about what they were playing.”

Smed was engrossed in the game as well. He’d been playing since alpha, an early phase of development when few zones were online, and even those that were playable had placeholder artwork until artists could finish them. Witnessing the manifestation of the potential he had seen in online gaming since before he’d convinced the higher-ups at Sony to let the “ghouls and goblins guys” develop their fantasy game was a dream come true.

“Seeing others in the game was almost this spiritual awakening—and then playing with people who weren't even in my office? That was pretty cool. And other people having a great time? This was all so awesome.”

Brandan McDonald spent as much time browsing newsgroups and Web forums as he did playing the beta. “When the beta came out, I remember one of the developers talking to some guy in Ohio. He was a beta player, and he was like, ‘Hold on, let me shut off my tractor.’ I was like, ‘What the hell? There's some farmer in Ohio playing EverQuest.’ I would help people in the game, too.”

Despite rampant excitement within the dev team and without, Smed hedged his bets. At meetings, he’d write figures on a whiteboard to illustrate how well the game needed to fare at retail. If they could sell 45,000 units—one game box per customer plus each customer’s monthly sub of $9.89—they’d all still have jobs. If they could sell, say, 130,000, EverQuest could be considered a smash hit.

As the beta wore on, the only people who weren’t over the moon about EverQuest were Smed’s superiors at Sony. The “ghouls and goblins” team were still the black sheep at 989, which remained focused on sports games for the PlayStation console. “We were not only a non-console game, we were online-only, subscription-only, and 3D card-required,” Sites explained. “Strike one, two, three, and four. If this didn't do well, we were done. Everyone was super-passionate, but we also had one hand on our resume, making sure there was an opportunity somewhere else just in case this didn't work out.”

Smed did his best to shield the developers from management talk. Not even Brad McQuaid, Bill Trost, and Steve Clover, the game’s leads, knew how hard Smed was working to keep Sony from interfering in production. “When I was producing the game, I didn't need to know how many times it potentially got cancelled,” McQuaid said. “He did a great job running interference. He was basically my mentor. I learned a whole lot about, not just the games industry, but production, leadership, dealing with larger teams, things like that.”

For Smed, keeping EverQuest alive was a daily battle. “We would do these product reviews, and they were weird.”

At regular intervals, leads from 989 would show off games in development. NFL Game Day, ESPN Extreme, and other titles wowed attendees from other departments. Then Smed would take the floor.

“I'd show what we had, and up came this early version of EverQuest, and there were snickers—literally snickers—across the room.”

If Smed was the wall between his team and Sony management, Kelly Flock was the shield between management and Smed. He still believed in the potential of an online-only RPG, but as part of Sony’s brass, he had to prioritize the bottom line. “Kelly almost cancelled a couple of times. I had to, very passionately, ask him not to. To his credit, he let us keep going up until around three months before launch, when he spun us out,” Smed remembered.

Flock met with Smed in December of 1998, 13 months into EverQuest’s beta. “Look, I don’t think this thing’s going to work here. I’m going to give you a month to find another home for it.”

“What if I can’t?” Smed asked.

Flock held his gaze. “Then we’re going to have to fire the whole team.”

Smed pounded pavement. “I also talked to Microsoft. EverQuest almost became a Microsoft game. If they had moved a little faster, it would have.”

He also met with managers at another group within Sony. Smed was relieved: Like Flock, these executives saw something in EverQuest. The team was spun out, becoming an indie studio called Verant Interactive in early 1999 only to be merged back into Sony Computer Entertainment America a year later, and is known as Daybreak Game Company today. 989 Studios remained concentrated on sports games for PlayStation.

“I always thought that one of the reasons our team got spun out was because Sony didn't want to PC games,” Smed explained. “This is the first time I've ever talked about this, because this is information I got a few months ago. Apparently, what happened was that Kelly said yes on his own. He had full budgetary controls over what was then 989 Studios. They preferred we focus on PlayStation One stuff. I'd always thought Kelly spun us out because he didn't have full authority with Japan, and they were upset about it. But he said no, he did, and he spun us out because he wasn't sure if EverQuest would be a major success.”

ANDY SITES ANSWERED his phone and listened to the excited voice on the other end of the line. “Hey, I’m at an Electronics Boutique in Dallas. I’m in a line of about 30 people, and they’re all buying your game.”

Sites smiled. It was March 16, 1999, and the caller was one of the developer friends he’d made during EverQuest’s beta. Apparently the game was destined to become a hit in one major city, at least.

Then Sites got a call from a friend in Los Angeles. Then from a dev buddy in Minnesota.

“Holy shit,” one said, “I’m standing in line, and all these people are buying your game.”

Nearby, designer Bill Coyle had his fingers and toes crossed. “The talk among all the developers around the time of launch was, ‘If we can just get a hundred thousand people and keep them for six months, we're good. We've made it. That's all we're asking for. If only we can do that.’ On a wing and a prayer. Then we opened the servers.”

All throughout Verant Interactive, the reality of EverQuest’s popularity rolled over developers in waves. After installing EQ, players typed in their login credentials and were dumped into a big chatroom with a long list of servers. Time and again, hundreds of people would drop into the chatroom at once. Time and again, the server crashed under the weight of logins.

“We had a couple of customer service people ushering them into servers, because you'd come into this big text chatroom with a long list of servers,” Sites remembered, “and our people were saying, ‘Hey, choose a server. Don't hang out in here.’ We're in there watching all these people come through.”

EQ’s developers watched the festivities by typing in a special command that showed how many new players registered the game.

One minute: 5,000.

A minute later: 5200.

A minute after that: 5800.

All the while, the login server kept crashing.

One developer hustled over to the engineers. “Hey, there’s too many people in here.”

“No,” the engineer said, “You guys keep running this command and it’s querying the entire database.”

The developer went back to his desk and sheepishly told his coworkers to lay off the command—or at least only use it once or twice a minute.

“I can't remember the numbers for the first day,” Sites admitted, “but for days after launch, we gained between 10,000 and 20,000 people every day. I was thinking back on those whiteboard meetings where Smed was talking about 70,000 players over the game's lifetime. We immediately surpassed that, and it just kept going.”

For Coyle, launch day felt like being in the control room watching a space shuttle prepare to blast off to the moon. “I don't say that lightly, because watching a shuttle go up is a huge thing for all of mankind. But for me, this really was like that. We'd put so much work into this thing. All our hopes and dreams were involved in it.”

John Smedley watched the servers fill up with immense satisfaction. “Sony Online, as soon as it launched, saw it take off like a rocket, and ended up buying the rest of the company,” he said, referring to Sony’s decision to bring Verant back into its fold in 2000. Even if Verant had remained independent, EQ would have been considered a Sony property since the game had been made at 989, a Sony-owned studio.

“To this day, the decision Kelly made in funding us is a big part of why online gaming is what it is today,” Smed continued.

EverQuest’s developers remember March 16, 1999, as one of the most exciting and fulfilling in their careers. It was also one of the most stressful. The login server continued to crash, and so did the servers containing the game’s population as players explored, battled, and quested.

As much as Coyle believed in the game, he questioned whether players would stick with it through problems that brought EQ down for hours or days at a time. “We basically sucked all the bandwidth out of our ISP. We weren't prepared, and they weren't prepared. There was not enough bandwidth to meet the player demand, and that caused problems and crashes.”

The bulk of EQ’s launch issues boiled down to demand. “What happened was, this huge influx of people came in, and remember, this game was not digital downloadable; you could only get this by going into a store and buying a disc in a box,” Smed explained. “We sold out of the stock we had immediately and were putting way more people into the game than we thought we would, even over the game's life.”

Verant’s crew scrambled to add servers. “We didn't have data centers with racks of server blades,” Andy Sites remembered. “We had desktop PCs that were the world servers and zone servers in a data center. We bought racks where we could stack up 16 to 20 of these boxes. Our idea of being clever was figuring out that if pulled all the rubber feet off them, we could put one more row of boxes into one of the racks to avoid taking up more space.”

Time and again, the team’s estimations fell short. They bulked up to support 10,000 players, only for the servers to collapse. They bulked up to support 15,000, and the servers crashed again. Smedley recalled in the ballpark of 22,000 players registering on the game’s first day. “The servers would go down, people would slowly come back in, then they'd go down again. We had to start throttling the number of people who could log in at once. That's when we realized we had a huge hit on our hands.”

The problems were eventually straightened out through around-the-clock maintenance. “I still wonder, though,” Coyle admitted. “Are there people who tried playing the game on the first night, or the first week, and couldn't get in, and just went to the store and got their money back? It makes me very sad if those people exist. I would like to apologize to them. They missed out on a great experience, but we just didn't expect that kind of reaction.”

SHORTLY AFTER LAUNCH, EverQuest’s developers posted on their forums: They’d be hitting up a local watering hole after work to relax, and any players in the area were welcome to come hang out. They arrived to find over 100 avid fans waiting.

“It was pretty amazing that we could say, ‘Hey, we're meeting at this place’—which just happened to be right next to our office—’at six o'clock tonight. We'll buy the beer. Come if you want to.’ And 100 people show up, from our home area,” Smed remembered. “That was amazing to me, to be able to enjoy that moment. It had the whole team enjoying a feeling of celebrity status.”

Despite myriad hiccups, most developers remember launch day fondly. “We'd never done an online game like this before,” said Andy Sites. “We felt like as we were going, we were paving the way and figuring things out as we went.”

“For a while I dabbled in mobile games, and I enjoy playing those games, but I don't enjoy making them. I enjoy making MMOs,” Brad McQuaid said. “I guess it's just in my genetics. I could not imagine myself doing anything else.”

“When you're in that moment, and servers are going down, all you're thinking about is things that are going wrong,” Smed explained. “But when you see this insane amount of people coming in, more and more every day, it really makes you feel good. The team, too: Even though we were scrambling to solve problems, we were happy because we could say that we'd made something really good that people wanted. Today, stuff gets shared very quickly thanks to social media. Back then, social media didn't exist. EverQuest was very much a word-of-mouth thing. A few major websites covered it, but not many. All the people coming in were doing that because friends were playing. I would get emails from friends within different groups inside of Sony saying, ‘Hey, can you spare me an account?’ It's amazing the people who come out of the woodwork when you launch a successful game. It was an amazing time, and I'll remember it for the rest of my life.”