StarCraft: Ghost:
What went wrong

Robert Huebner was on top of the world.

The Tokyo Game Show. 2002. Less than two years prior, Blizzard gave Huebner and his team at Nihilistic Software the green light to start development on a new game.

The deal was, in many ways, an experiment. In working with Nihilistic, Blizzard gave a good amount of control over one of its major franchises to an external development studio, a rare move for the company.

Blizzard wanted to make a dent in the growing console market, and Nihilistic provided an opportunity to do that. The original Xbox was less than a year old, and the PlayStation 2 was dominating console sales. Nihilistic wanted to grab the players who preferred a couch to a PC. It was bold. It had to be.

Sonic artOver the course of years in development, StarCraft: Ghost had many ideas thrown out and replaced, which proved to be one of the biggest challenges behind the scenes. As new ideas came along, the development teams had less time to see through the original ones.

In 1998, Blizzard made a mark on the industry with StarCraft, a real-time strategy title that topped PC sales charts and became a cornerstone for the professional game circuit. Nihilistic wanted to take that franchise and give players a fresh look at the game's world through the eyes of a popular character.

All that added up to StarCraft: Ghost.

Two years in, Huebner and his team were enjoying the fruits of their hard work. The entire staff — about a dozen at the time — had flown to the Tokyo Game Show, where playable versions of the game were available on the show floor. With StarCraft hot on the competitive gaming scene, Warcraft 3 just released a few months prior and World of Warcraft still years away, a console game featuring a new take on an existing universe allowed Blizzard to maintain its momentum.

"Tokyo was the high point," says Huebner.

Less than two years later, frustrations that had boiled beneath the surface came to a head. A conversation with Huebner and Blizzard in a conference room at E3 2004, mere feet away from crowds eagerly playing new levels, more or less marked the end of Nihilistic's work on the game.

The story of StarCraft: Ghost is a complicated one that spans two development studios, a buyout by Blizzard and declarations that, even though no work was being done on the game, it was never technically canceled. Polygon recently spoke with nine developers involved to look back at the project.

"StarCraft: Ghost was in the wrong place at the wrong time," says James Goddard, an action game veteran who worked on Ghost for years.

"The game had an identity crisis," he says.

Lock and load

Before StarCraft, there was Star Wars.

Working at LucasArts during the 1990s was a dream for many developers. But for Robert Huebner that dream quickly grew tiresome.

After working at the company for a couple of years, Huebner began preliminary work on an expansion pack for Jedi Knight: Dark Forces 2, the sequel to the successful 1995 shooter Star Wars: Dark Forces. It wasn't exactly what he had in mind.

"I jumped ship," says Huebner. "It was a great place to get started. We used to call it Lucas University, because the pay wasn't great but it was a great starting ground."

Huebner had experience. His history included working on Descent, the 1994 sci-fi flight shooter that helped popularize the "six degrees of freedom" method of player movement at Parallax Software (now Volition, known for the Saints Row series). But after two years at LucasArts, he was ready for a change.

At the time, Blizzard was ramping up its resources. It had yet to release StarCraft and was hot from its success on Warcraft 2 and the original Diablo. It wasn't yet a behemoth, but it was growing quickly.

"They had their lion's share of the heavy hitters [from LucasArts]"

Huebner ended up landing a gig there. Hired to work on a new role-playing game, he eventually worked on the original StarCraft before its release in 1998.

But in 1997 he received a call from some friends, including Ray Gresko, who also worked on and helped lead the Jedi Knight titles, and Steve Tietze, who had worked at Rogue Entertainment on expansions for the original Quake. The pair pitched Huebner on starting a studio.

Despite being terrified of leaving Blizzard so quickly, Huebner took a chance and left for the new venture: Nihilistic. This parting of ways would eventually help create the path for StarCraft: Ghost.

Nihilistic's first title, role-playing game Vampire: The Masquerade, was a hit. Chris Millar, a key member of the Blizzard team on StarCraft and Diablo titles during the 1990s who now runs Fat Princess developer Fun Bits, says Blizzard was more than impressed.

"Nihilistic was top of the line with what they had done with Vampire," Millar says.

Chris McGee, another LucasArts veteran Huebner brought to Nihilistic, says the small company was one of the best teams he had ever worked with.

"They had their lion's share of the heavy hitters [from LucasArts]," McGee says.

With Vampire under its belt, the team had to decide what to do next. Huebner wanted to make an impact on the industry. Teaming up with Blizzard would provide Nihilistic a chance to benefit from the rising giant's success.

"The first direction was about that question, 'What if you were in a StarCraft battle?'"

No one we spoke to for this story remembers who came up with the initial idea for Ghost, although Huebner says the team was greatly influenced by the rise of stealth games.

"Metal Gear Solid was out. Splinter Cell wasn't too far out. So we felt it was a good fit," he says.

Huebner and his team, including co-founder and creative director Ray Gresko — who left Nihilistic in 2004 and eventually joined Blizzard itself — presented their idea to Blizzard, suggesting it could use Nihilistic as a way to enter the console market.

Millar was in the room when Nihilistic pitched the idea.

"Ray Gresko showed that he knew StarCraft better than anyone else in the room," says Millar.

The premise was simple. Players would take control of one of StarCraft's most powerful characters — but not from a god's-eye perspective. Ghost would put the player on the battlefield, giving them a view of the universe they had never seen before.

Bill Roper, an industry veteran who joined Blizzard in 1994 and went on to become the creative head of Disney Interactive, says Nihilistic took a known character to new places.

"The first direction was about that question, 'What if you were in a StarCraft battle?' You were actually in the battle. And it was exciting.

Sonic artAs time went on, team members at Nihilistic say the game went through enough development revisions that at times it felt like they were chasing their tail.

"In fact, one of the first milestone builds of the game [had] a character in a trench with Zerglings coming overhead. ... It really felt as though you were in the StarCraft universe."

The pitch demo featured impressive tech, according to people in the room, with battlefield views of staple StarCraft units like Siege Tanks and an intense air strike.

Combined with Nihilistic's passion for the project, that demo showed Blizzard what it needed to see. It also came at the right time — Blizzard was in the midst of working on a range of projects to keep its momentum going.

"One thing about having a large team is you have no shortage of ideas," says Millar. But Blizzard only had so many resources, especially in the early 2000s. Much of the studio was still working on finishing Diablo 2.

Roper says Nihilistic adored the world Blizzard had created.

"That was one of the reasons we wanted to work with them," Roper says. "We had a lot of respect and enjoyment for their design sensibilities. I remember some of those meetings because their enthusiasm for the project was infectious."

The pair struck a development deal whereby Nihilistic would contribute the main work on the game and Blizzard would guide development with constant feedback. Blizzard would also provide cinematics.

According to Huebner, the terms of the deal were relaxed in that there was no specific deadline. The teams would simply work on the game until it was good enough. It felt almost too good to be true, he says.

"[Blizzard] knew it was an iterative team, and the whole success there had been on taking their time," says Huebner. "It was ... like, 'every month we'll pay the team and you'll give us what we ask for.'"

"We thought it was the best thing ever," says Huebner. "It provided us security for the team, and gave us time to ramp up. But the dark side was that if it kept going and going ... it can be trying to work on something for a long time and not see something shipped."

With the idea, the team and an agreement between the two studios, Huebner and his team started development from Nihilistic's office in San Francisco.

It was time to get to work.

Jacked up and good to go

After signing the deal with Blizzard, the Nihilistic team was eager to begin work on Ghost, and according to many accounts, early development was a fun ride.

Huebner says the team put a lot of effort into making the character, and by extension, the player, feel as though they were in control of a ghost. Abilities like cloaking and swift movement would make the player feel like a combat veteran — and the power to call down nukes, taken from the original StarCraft game, would provide a feeling of battlefield domination.

While multiple team members speaking for this story couldn't describe a consistent plot, as it went through several iterations, the general consensus was that players would take main character Nova through a journey in which she would come up against enemies in all three of the StarCraft factions.

"Our goal was to make it so compelling that people would practice moving before doing anything," says Chris McGee. "Nova was just fun to move. People made fun of me while I was just moving her around. I was cruising and rolling, calling down Valkyrie strikes on big Zerg monsters — it was great."

"We were sure it was going to sell a lot," says McGee.

According to McGee, Nihilistic was a positive place to work, mostly because of Huebner's emphasis on collaboration.

"He really wanted everybody's input on decisions," says McGee. "Almost to a degree I had not seen in any other place."

The team was also in constant communication with Blizzard, which appointed an internal employee as a producer. For a time, Chris Millar held this position.

"I was excited because ... hey, it's Blizzard. 'We're going to get a game under Blizzard!'"

"I had a good relationship with [Nihilistic]," says Millar. I would visit them two times a month, and I would see how hard they were working. It felt like a family."

"Just imagine a [small] team trying to work with a company the size of Blizzard."

David Ryan Paul, an artist who joined the team toward the final stages of development, was yet another LucasArts veteran. He says that while Ghost wasn't in continuous crunch, the small group was serious about the work.

"We weren't a rowdy bunch of people," he says. "I was excited because ... hey, it's Blizzard. 'We're going to get a game under Blizzard!'"

McGee recalls the mood being overwhelmingly positive.

"We thought we were making a kick-ass game. That was the vibe," he says.

"It seriously looked cool," Ryan Paul says. "I wasn't even the biggest stealth fan. But it played extremely solid, and for someone not having any experience in that genre, I could pick it up quickly."

Ryan Paul goes so far as to say the game played as a shippable title, even though by 2004 there was still a significant amount of content missing. This was normal for development of a game of that size, many on the team say, but this is also where gaps started appearing in the development process. According to several members of the team from both Nihilistic and Blizzard, Ghost's open-ended development timeline allowed for lengthy debates and significant iteration to occur without the team making progress toward a shippable title.

Never knew what hit 'em

In interviews for this story, no team member could cite a specific point when they believe Ghost started to fall apart. Huebner and other developers say it started seeing trouble when regular meetings with Blizzard became more about adding new features and experimenting and less about perfecting an established idea.

While Blizzard was used to doing work this way — the company had scrapped multiple games it believed were not up to scratch — for Nihilistic, it was a new way of working.

"I've been on projects where you chase the tail of what's hot, and it was a little bit like that," says Huebner. "It wasn't focused."

"We would riff on ideas, but after a while, it got to be trying because it would be like ... whatever the new game was that month, they'd want to add those features," Huebner says.

With Splinter Cell hitting shelves in 2002, Huebner noticed some of the feedback from Blizzard started to mirror the features contained in the new title.

"It was like, 'In Splinter Cell you can use your legs to hold yourself on the wall; I'm sure they're going to want to add that feature,'" he says.

Sonic artAs games like Splinter Cell and Metal Gear Solid grew in popularity, Nihilistic team members heard requests to add features from those games into Ghost, which frustrated some who felt their game was being pulled in multiple directions.

An added problem was inconsistent feedback. Huebner, McGee, Roper and Millar all agree Ghost suffered from a number of producer changes, which meant that no one person at Blizzard was responsible for the entire life of the project. This also made the Nihilistic developers feel as though Blizzard wasn't making Ghost a priority.

"There would be large gaps in feedback for Nihilistic," says Millar, whom Huebner names as Nihilistic's most positive and productive producer on the project.

"When you're not on-site every day, and it takes a week to review something, that 10-15 person team has already moved on, and then there's a huge chunk of changes and feedback requests," says Millar.

"I think that's where things started to get off track," he says.

Millar echoes Huebner's criticisms that Blizzard never gave Nihilistic a clear direction.

"It started as stealthy, and we loved what the team had done, making levels based on sneaking and using the laser spot to drop a nuke and so on," he says.

"But as we played the demos, there was a large contingency of people who believed it should be more action-based ... a certain amount of the team felt like an action game would be cooler."

"Blizzard wanted more action," says Millar.

"It started as stealthy ... but as we played the demos, there was a large contingency of people who believed it should be more action-based"

"They wanted different moves. I remember Nova had a stealth jump on top of Marines. She could take on a lot of hand-to-hand combat, and they upgraded her rifle.

"But then it came full circle. It was stealthy, then it felt like it had too much action. Then they needed more stealth. And then what about multiplayer? It became so rough. The Nihilistic team was killing themselves working."

Much of the confusion was contained within what Blizzard called "Strike Teams," where people from across the organization — including those not involved in the project being discussed — met to give feedback.

Ghost felt "true" to the universe, but the wild iterations between genres were evident in the Strike Team presentations, Roper says.

"You'd get a Metal Gear Solid coming out, then a Halo," says Roper, who says Nihilistic was simply responding to feedback on what Blizzard thought was the right direction at the time.

Part of the problem, Roper says, is that Nihilistic was good enough at delivering new demos that Blizzard felt confident in being able to request slightly new takes on existing ideas and ask for them be developed in relatively small time periods.

Roper recalls a build in which players would run through trenches and then call in a nuclear strike on a very large StarCraft-like battlefield. Another build "that felt much more like Splinter Cell" had players sneaking through hallways and taking down individual enemies.

"We even had hover bikes at some point. There was just a process of trying to figure out what the mechanics were to tell a ghost story, and that kept lengthening out the process," says Roper.

StarCraft: GhostNihilistic and Blizzard had extensive discussions over the look of a boot in the game, one of many aspects that the two sides couldn't easily agree upon.

Huebner says the team at Nihilistic became frustrated with Blizzard's inability to come up with a direction. Each time the senior crew flew down to Blizzard to present a new build, the team joked about what game it would be given as inspiration next.

Huebner also became frustrated over discussions regarding small details that, while not as important to Nihilistic, were crucial to Blizzard.

"I remember we had a lot of fights about the design of a Marine's boot," Huebner says.

According to Huebner, one of the first signs that fatal cracks were starting to appear in the development process came when Blizzard hired an external producer, James Goddard.

"He's not on my Christmas card list," says Huebner.

Goddard had just finished working on another project and was setting up a consultancy. He had mentioned this to Blizzard chief creative officer Rob Pardo and offered his expertise in character combat design. Pardo brought Goddard on board as a consultant, and his participation grew into more of a production role over the course of the project.

Given Goddard's background, Huebner felt he was pushing the action aesthetic too much.

"Blizzard started to use me for more messaging ... there might have been some moments causing tension," says Goddard. "There was probably some tension."

Goddard's view on Ghost is not that he wanted to change the game's genre, but rather that the game needed action elements in order to keep the player moving.

"This was the identity problem," says Goddard.

"I pushed the idea at the end of the day that it was a stealth game, but it's not a hardcore stealth game. I think people really struggled with that."

That conflict between action and stealth would continue throughout the game's development, says Goddard.

Goddard's appointment, among other decisions, contributed to the cumulative feeling at Nihilistic that no one at Blizzard took the project under their wing. Among team members, there was a collective feeling that Rob Pardo provided much-needed guidance and specific feedback before Goddard came on board, but that didn't last long.

StarCraft: GhostAs time went on, many at Nihilistic and Blizzard felt communication began to break down between the two companies.

"When Rob Pardo worked on the project, he really helped, but once he left, the project was in big trouble," says one Nihilistic senior team member who requested to remain anonymous to avoid burning bridges.

As development continued, Huebner says Blizzard pushed to add more features that he believed were inconsequential or unrealistic to include with so few resources at Nihilistic.

"After the Tokyo Game Show, where we had some good levels and we felt the game was on the right track, we brought in a multiplayer mode," says Huebner. "The art style changed. We were chasing team co-op multiplayer that was not even part of the original design."

Whether it was due to too many decision-makers involved, or there weren't enough resources on either side, the game just wasn't moving forward and in a place where Blizzard felt it could be shippable.

One senior developer, who also requested anonymity for fear of harming future business relationships, says the story also suffered from a number of rewrites, and that exacerbated the development process.

Chris Millar says the sheer amount of work meant there was "no way to get consistent consensus from the Strike Team."

In late 2003 and 2004, the tension was coming to a head. The game's development had continued for so long, and no progress was being made. Huebner even signed another job for Nihilistic in anticipation that the contract between it and Blizzard would fall apart.

"We had booked work to start with EA, in anticipation of Ghost either finishing, or us leaving it," says Huebner. "The writing was on the wall."

"We thought that if we can't buckle down and finish by a certain date, then [development would cease] ... I don't know if we ever said that explicitly, but I think both sides knew that," says Huebner.

"It was not a shippable, super polished game at the Blizzard standard, but there were glimmers that were fun," says Huebner. "There were ... levels that were playable. A really good stealth level, a Zerg level, a stealth-heavy Protoss level in a Protoss ship. They all played like different games.

"The writing was on the wall."

I'm not reading you clearly

There weren't many people at the meeting. Just Huebner, a few people from Nihilistic and some Blizzard representatives. Huebner says he could feel the project was coming to a head.

It was 2004, more than three years after development had started. Eager players were trying out Ghost on the E3 show floor mere feet away. But in a nearby conference room, Huebner and Blizzard met to discuss how the game was coming along.

"It was mostly about where the game was relative to shipping — how many aspects of the game we considered 'at alpha' and how much was still left to do to complete the game," says Huebner.

It didn't go well, Huebner says. While neither party explicitly stated development was about to end, Huebner says it was clear from the things being said that both sides of the project felt production needed to wrap up.

"That's when we were thinking, 'We should line up our next thing,'" says Huebner.

Sonic artAfter Ghost had spent years in development and many on both sides of the fence had become frustrated with the situation, Blizzard met with Nihilistic and indicated that its contract might be winding down.

Goddard remembers the meeting taking place, but says he wasn't there.

"It made sense. But it was heartbreaking," says Goddard.

Huebner says the team was fatigued by the development. It had worked for more than three years without seeing a shipped game.

"I don't [think Blizzard] ever quite knew what they wanted," says Huebner.

For David Ryan Paul, it was a normal day until Huebner gathered the staff members and told them they would no longer be working on the game.

"It sunk in, but it didn't have the weight [for me] it did for people who had been working on it for three-plus years. It was shocking and disappointing," says Ryan Paul.

"It was heartbreaking, particularly for artists. They've got a ton of work they can't show," says Huebner. "That was definitely a short day."

Huebner says that he was glad no one was laid off or missed a paycheck, but that many people took it hard. "I remember one guy cried and went home," he says. "It was traumatic for some people."

"You try not to think about it and move on. I still have a StarCraft Ghost shirt from E3"

The game's composer, Kevin Manthei, says he had no idea it was about to happen.

"The feedback I was getting was really good," he says. "Everyone is happy. Then I got a phone call thanking me for everything, and I thought the project was dead.

"You try not to think about it and move on. I still have a StarCraft Ghost shirt from E3," says Manthei.

Nihilistic had the ability to move on by using the Ghost tech it had developed on a new game. But Huebner says the resources didn't make up for an unshipped game with three years of work.

"I would have liked to have seen the original Metal Gear-, Splinter Cell-type game. That felt true to the Ghost aesthetic. That's the part I enjoyed the most," says Huebner.

Heubner thought he wouldn't see Ghost again.

He was wrong.

Wanna turn up the heat?

Swingin' Ape Studios had been on Blizzard's radar for a while. Its 2003 debut, Metal Arms: Glitch in the System, was a third-person action adventure game that received favorable reviews — much like Nihilistic's Vampire did years earlier.

So when Nihilistic stopped work on Ghost, Goddard says the feeling internally at Blizzard was that Ghost was in good enough shape for Blizzard to give it to another studio.

"Swingin' Ape seemed like a good fit," he says. "We talked about it, and [Blizzard] wanted to reboot it."

Blizzard hired Swingin' Ape to work on Ghost in July of 2004, immediately after Nihilistic stopped working on the game. The news broke shortly afterward. Nihilistic had started working on its next title, Marvel Nemesis: Rise of the Imperfects, with Electronic Arts. And Huebner was surprised work was continuing.

StarCraft: GhostAfter Blizzard handed Ghost to Swingin' Ape, the new team moved quickly to make the game its own, repairing much of what it saw as broken in Nihilistic's version of the game.

"That was weird," he says.

"I don't remember how we found out, exactly. It was a short window between us and Blizzard getting [Swingin' Ape]," says Huebner.

"I don't have a lot of insight into that or how they found them," he says. "It seemed opportunistic, like maybe this developer was next door and they were able to buy them cheaply? Something like that, maybe."

Sources say a shift started as soon as Swingin' Ape began work. Development became less stressful. Feedback became more streamlined and consistent. After several months of Swingin' Ape working on the game, Blizzard purchased the studio in May of 2005.

Multiple sources point to multiplayer being the primary reason reason Blizzard gave Swingin' Ape the responsibility of shepherding the title. While Nihilistic introduced some multiplayer elements, according to Huebner and Goddard, there was pushback because Nihilistic team members felt the game was moving too far away from the original vision. By bringing the game and the Swingin' Ape team internal, Goddard says, Blizzard could develop a multiplayer system properly and with more attention.

"Swingin' Ape did a really good job, and the game was really going somewhere and really impressive," Goddard says. "The gameplay was good. The shooting was good."

"When they got [multiplayer] working and you're playing multi-console and jumping on speeders ... that was just, holy shit," says Goddard.

Matthew Bell joined Swingin' Ape before work on Ghost began. "I have fond memories of working there and seeing the banks of CPUs that were designed to be the farm for World of Warcraft," he says.

At the beginning, Swingin' Ape worked apart from the main Blizzard crew in its own facilities, but Bell says the company made them feel welcome. And unlike with the Nihilistic team, Bell says he felt as though Blizzard then made Ghost a priority.

"It was a very positive time," he says. "I would say they definitely made us feel as though we were part of their family."

Sonic artAs Swingin' Ape's version of the game progressed, many at Blizzard were happy with how smoothly the game played.

"They were so invested, they brought us to headquarters so the feedback could be hands on. They didn't just pay us to develop it. They bought the team so they could make it into a Blizzard title and gave us all the support we needed," says Bell.

Much like Nihilistic had its high point at the 2002 Tokyo Game Show, Bell says BlizzCon 2005 was the high-water mark for the Swingin' Ape team.

"That's when I was the most positive about the game," he says.

Blizzard showed a cinematic at BlizzCon 2005 to confirm that work on Ghost was still underway. It had been four years since development started on the game, and there was still no ship date. But at that moment, Bell says the team was more positive than ever.

"It went over really well with the hardcore fans, and everyone on the show floor was having an awesome time with the multiplayer. It was just a lot of fun," he says.

Some Nihilistic developers had a different reaction. After working on a stealth game for so long, seeing an action-oriented game with heavy multiplayer capabilities confused them.

"I remember seeing a playable character of a Firebat and then seeing a multiplayer deathmatch ... I didn't understand it," says Chris McGee.

"It was frustrating. But this whole industry is full of frustration," McGee says.

Chris Millar left Blizzard well before the game moved to Swingin' Ape — but once he saw the footage, he says, it was more than a surprise.

"I was shocked by how different it was," he says.

Nuclear launch detected

Positive sentiment grew during 2005 among Swingin' Ape team members, especially after that year's BlizzCon. But those same developers say that pooling Swingin' Ape talent with the rest of Blizzard staff meant they couldn't expect to have their own project cordoned off from the rest of the company's pressures.

And in 2005, Blizzard was under a lot of pressure.

With World of Warcraft released in 2004, all hands were on deck fixing problems, working on the game, making sure it was running smoothly. Sustaining the momentum required a significant amount of attention. By early 2005, less than a year after its launch, WoW had 1 million subscribers, and that number was continuing to grow.

With so much attention on WoW, other projects and games naturally lost some attention, team members say. Including Ghost.

"It was good by most people's standards, but the technology was aging out"

"To say that WoW was an all-encompassing distraction is an understatement," says Roper. "We had to make WoW as big as possible ... and so there was certainly an amount of letting Ghost go in that course of development."

And there was another problem.

"The industry was turning a corner," says Matthew Bell.

Ghost had taken nearly an entire console generation to develop. A game designed for the PlayStation 2 and original Xbox, it was now scheduled to be released in 2005 — right when the Xbox 360 was hitting shelves.

Blizzard faced a problem. The company could push Ghost into a new phase of development, which would mean adding more resources to the game to make it ready for a new console generation. Or it could simply shelve the game and possibly come back to it at a later date.

James Goddard says the choice was obvious.

"The trajectory was a problem," says Goddard, who left six months before Blizzard made the final call. "It was good by most people's standards, but the technology was aging out.

Sonic artFollowing many ups and downs over the years, eventually Blizzard called it quits on Ghost and moved on to other projects.

"If you look at it that way, the decision makes a lot of sense. It's heartbreaking, but it makes a lot of sense."

Bell says he wasn't at any particular meeting that ended development on Ghost. But the whirlwind of World of Warcraft and other projects took hold, and slowly, from his perspective, Ghost began to lose influence due to having both feet in the previous console generation. Additionally, with Blizzard seeing such success on PC, getting into the console market seemed like less of an essential proposition, from Bell's perspective.

One anonymous source familiar with senior decisions on the project describes a meeting in which a major retailer told Blizzard that Ghost would not receive premium floor space as long as it remained in the previous console generation.

Blizzard did not respond to questions about this meeting, or to any other questions for this story, and Goddard also says he never heard about it, but he agrees that Ghost was in "the wrong place at the wrong time" and that Blizzard viewed the turn of the console generation as a critical problem for the game.

"I think WoW changed the culture of Blizzard, and all the focus was on that for a while," he says.

Bell believes that if the Ghost developers had made a choice to put Ghost on the Xbox 360, Blizzard's executives would have supported the team. He describes Blizzard decision-makers giving that choice:

"The generosity of the executive team was that they said, 'You guys should do what you think is best.'"

But Bell says that with so many resources in the company competing for talent, including work on StarCraft 2 and a new iteration of Diablo, the Ghost team would have struggled to get things done.

"We just weren't competing for talent. So they dissolved the team," says Bell.

Developers moved on to new projects inside the company. After more than five years of work, Blizzard put StarCraft: Ghost aside.

Ready to roll out

For some developers who worked on StarCraft: Ghost, another canceled game was simply a part of working in the industry. But for many of them, Ghost was something special.

Goddard says there have been two or three canceled games he's worked on that were more upsetting for him than "several" others.

"StarCraft: Ghost is one of those," he says. "I think if it had come out, it would have been awesome. Just awesome ... If they wanted to do it now on current systems, and if they could get past the identity crisis ... I'd love to see it happen," Goddard says.

For McGee, the time spent developing the game was worth the effort.

"Shipping a game is not the ultimate pleasure as a game designer — it's making good gameplay ... That's more important to me than going out the door ... You can ship a lot of things but it doesn't make them good," McGee says.

Looking back, Huebner says, the project had its problems from the start. And he harbors no bad feelings towards Blizzard for why the project failed.

"Blizzard wasn't set up as a publishing company," he says. "But they have their own culture. And they had to kind of break through that cultural barrier."

"I'm no fan of working with external developers either!" says Huebner.

Roper says that working on the project and seeing it canceled was obviously a disappointment — but that the lessons and experience gained by both teams during development don't disappear just because the title didn't ship.

"I hope people realize that it's no more of an easy thing for a company and developer team to cancel a game than it is for the fans to hear about it," he says.

"It's an extremely difficult thing to do. This whole process ... it's as much of an art as it is a science." Babykayak

Illustrations: Daniel Purvis