Congratulations, good people of Valve Software. You make terrific video games. Tell us how it's done!

"I'm stammering because I've tried to explain to my wife how it works, and I haven't been able to," Erik Wolpaw, co-writer of Valve's Portal 2 said last week, answering a question from a moderator about Valve's game-making process. "She doesn't believe that it should work."

What could this certain-to-fail-yet-somehow-it-doesn't process be?

"It's just, like, everybody kind of pitches in and order sort of emerges from that chaos," he said.

Absurd! Borderline Communist! Or... genius?

Wolpaw, who doesn't just write at Valve but does a tiny bit of programming and a smidge here or there of other stuff, said the secret is in the willingness to be a proven failure. That's where all that "order" comes from. "A lot of that is the playtesting," he explained. "You know, it's forcing people to put their ideas to the test and watch a lot of people go through it. It's hard to keep defending some idea when you see it fail repeatedly. It would take this huge act of hubris to cling to something. It's kind of a humbling experience. It's sort of liberating in a way, because you expect all sorts of things to fail. People get used to just not feeling like 'I've really got to cling to this idea because this is my idea and I worked on it.'... Most of the ideas don't pan out."


Subject yourselves to the scrutiny of the average person, game developers, and you too can work the Valve way.

Wolpaw's full NYU presentation—covering Valve's methodology, Portal 2 heroine Chell's reluctance to speak, and plenty more—was recorded by the University and is scheduled to be posted online soon.