GDC 06: Dissecting a zombie

SAN JOSE, Calif.--With the 2006 Game Developers Conference winding down Friday afternoon, Wideload Games founder Alex Seropian discussed the creation of Stubbs the Zombie, both what worked and what was an abomination.

This wasn't just any old postmortem, however. For one, Seropian delivered his findings with a style and at times off-kilter humor that should be familiar to those who have played Stubbs. Beyond that, Seropian's presentation was a look back on an experimental method of game development that he conceived upon leaving Bungie (another studio he founded) after the Halo developer was acquired by Microsoft.

"I was trying to think what I wanted to do with my life," Seropian said of how he spent his time after leaving Bungie. "What I wanted to do was make games. I wanted to be independent, I wanted to work on original games, but I was kind of terrified of that prospect after watching how Bungie evolved for about 10 years and watching a lot of indie developers, what they were faced with."

With game development budgets more commonly soaring over the $10 million mark, and team sizes swelling into the hundreds of people, Seropian decided to buck the trend and came up with the Wideload model of game development. He created a core team of developers, a staff of 12 full-timers. That staff would handle all pre- and postproduction work on games. For the trench work in between, they would hire outside contractors.

To guide the choices made in setting up the model, he created the Wideload Commandments.

Thou Shalt Establish Thine Own Creative Direction--"This industry makes more money than the US box office, but we can't be a creative ghetto."

Own Thine Own (Intellectual) Property--"We want to determine whether we're successful or we fail. We don't want to be wholly dependent on somebody else for an IP, or on somebody else to market our games, to fund our games, or to control our brand."

Be No One's Beeotch--"This really speaks to why we're independent," Seropian said, calling Wideload's situation with Stubbs publisher Aspyr Media a "partnership of peers."

Keep Thine Overhead Low--"You have a small team, then you have a low overhead. A low overhead saves you money, obviously. And since Wideload has a small team, Wideload equals money."

With the basics laid down, Seropian delved into what worked and what didn't, starting with an unexpected bonus.

"We didn't really plan this, but probably the coolest thing is that we had a really creative environment," Seropian said, noting that brainstorming was more productive when handled in a small, casual group. "It's kind of free of design politics."

Seropian also said the company's low overhead gave him leverage to walk away from a proposed bad publishing deal and the ability to survive a four-month delay that would have been disastrous if he'd had a larger team to pay.

As for working with contractors, Seropian invoked what he called the "Cold, Mean, Bastard Rule." It's a lot easier to fire a contractor if things aren't working out than it is to toss a full-time employee who was brought after a lengthy search of numerous applicants, and Seropian said that rule came in handy on a couple of occasions during the development of Stubbs.

Using contractors for the production phase also helped the company quickly scale the number of people working on the project to meet needs, Seropian said. And Internet tools like instant messaging and giving contractors remote access to the game's assets library were invaluable in working with people from Austin, Texas, to Bangalore, India.

Then it was time to look at what went wrong. First and foremost, Stubbs was built using the Halo engine that Bungie built from the ground up. This was less than ideal for contractors who hadn't worked with it before, as Bungie had no documentation for the engine, nothing to use to train people how to work with the engine or troubleshoot problems that arose. As a result, Seropian said he spent an inordinate amount of time not just training people how to use the engine, but deciding which contractors needed training, and how much.

Seropian also said he ran into problems by not doing enough due diligence when hiring contractors (hence the need to invoke the "Cold, Mean, Bastard Rule") and that Wideload needed to manage them better.

"Going from 12 people to 60 in a few weeks or a month is a very different experience, and you need to have your s*** together to pull that off," Seropian said. "Some of our s*** was loose."

One problem Seopian said the team ran into is that it didn't have enough producers.

"I'll be honest," Seropian said, "we had no producers on our project. Which is insane if you think about it."

Finally, Seropian's "Crunch Avoidance System" didn't quite work as planned. He had anticipated that the Wideload model would place the brunt of crunch work on the shoulders of contractors while the core team would be spared. It didn't work out quite like that.

"I'll be honest. Our crunch was actually pretty friendly. We had maybe three months of long hours, and I don't know that we worked much in the way of weekends."

All in all, Seropian declared the Stubbs experiment a success.

"It had its problems, but we did ship a game," Seropian said to much applause. "We are still in business and working on a new game, which is pretty cool."