Mark of the Ninja devs explain 'the theory of failing quickly' at GDC

At a GDC panel today, Klei Entertainment founder Jamie Cheng discussed how the developers of Mark of the Ninja use their "theory of failing quickly" to create games, a sustainable studio and keep developers sane.

Jamie Cheng, founded Klei Entertainment in 2005, with a goal in mind.

"I wanted Klei to be a place where we're continually building amazing games over and over," Cheng said.

The only way to make that happen, he said, is to "make sure the process is sane" and "sustainable."

The studio's development process grew out of Shank's development, which was so stressful that Cheng promised himself he'd never do it again. He considered the process a failure. He wanted to do better. So that they wouldn't fail again, they started developing theories.

"There are theories that we can use that help us better understand how game development works," Cheng said.

In Cheng's estimation, the biggest waste of time during development is unwittingly building the wrong elements for the game. In the worst case scenario, that means developers have to cut what they've spent time working on, after they discover that they're are building the wrong components for the game.

To avoid the wasted time and "create processes that allow us to create art," Klei puts the theories into practice during development.

"The goal is to identify that quicker, and get the hell out of there," Cheng said.

Using Mark of the Ninja as an example, he revealed that the process began with a "massive failure" that ran from March-July 2011. At the time, Klei had several of the stealth game's basic elements, but those failed to coalesce into a unifying experience. The game felt "murky," he said.

"We were building a whole bunch of stuff with bad assumptions," Cheng said.

By July, Klei switched to action and deemphasized stealth, which he said made the game worse. So Klei returned to the game's roots.

"We went back and we said, I'm going to build testable hypothesis," he said. "This is what I actually believe. Let's test it out."

Klei began prototyping scenarios and testing them as the developers "tried to fail quickly."

Using that process, they were able to determine which of their assumptions about gameplay were right and which were wrong.

After building "about two dozen" rapid prototypes, the developers realized that the point of the game was to plan. That required a game that contained a "predictable cause and effect" that players could observe, gather information about, plan around and then execute.

After realizing that, Mark of the Ninja "kind of started designing itself," Cheng said.

The renewed focus allowed the studio to pare the game down to its most basic requirements.

"I see Mark of the Ninja as a finely sharpened blade because we just kept shaving things off," Cheng said.

Ultimately, it's about creating a "processes that allow us to create art," he said.

"Having an unsustainable process is going to suffocate your company — or yourself."