Researching The Next Wave Of Innovation

[What purpose does R&D serve in the game industry? Gamasutra speaks to Rod Humble and Sony's Richard Marks, in charge of the team that developed PlayStation Move, to find out what pure research can do for games.]

First you come up with a lot of great ideas. Then you talk about them. Then you implement a lot of art and gameplay. And then you go "shit, that kind of sucks" ... and you toss everything out and start all over again.

That's Blizzard Entertainment's version of research and development, according to a talk game design senior VP Rob Pardo gave a few years ago about the developer's keys to success.

Nintendo and Valve are two other companies merciless about cutting games that aren't working, polishing and fixing games that are, and developing titles for a long time in secret until they are ready to be announced -- rather than announcing a game just because it exists.

Indeed, Nintendo president and CEO Satoru Iwata has said that trial and error is the key to creating fun, that sometimes the prototype phase lasts more than two years -- and sometimes the prototypes are scrapped altogether.

Given the three companies' successes, it is surprising how few other game companies do anything like R&D -- let alone pure research into game mechanics. What's dissuading them?

Ron Carmel, co-founder of World of Goo indie developer 2D Boy, believes there are potential benefits to launching just such a program in major studios. In his keynote speech at the Montreal International Games Summit last year, Carmel urged large publishers to use their resources to create small internal teams that would work on groundbreaking games from within.

World of Goo

"We need a medium-sized design studio... something that is larger than a typical indie but has the same propensity for talent density, focus, and risk-taking," said Carmel, formerly an employee of Electronic Arts prior to going independent.

Carmel recommends that a focus on profit be eliminated from the equation.

"Creating this within a major developer doesn't present a problem," he said. "With a budget of $1-2 million, 10 staffers could be hired to work on creatively ambitious and forward-thinking projects."

He likened the idea to the car industry which, alongside its mainstream consumer products, works on concept cars, few of which enter production as regular models.

"The concept car is a marketing expense to build your brand and say, 'Look at all the amazing things we're creating.'"

Rod Humble is another EA vet, now CEO of San Francisco-based Linden Lab, the internet company best known for creating Second Life.

He agrees with Carmel, recalling that the "good old days of lots of 'blue ocean' design was back before the console era, when platforms were almost all open platforms, and one or two programmers were capable of creating an entire game."

When consoles became popular, he recalls, "games grew larger, teams got bigger, and expenses inflated. By nature, that restricts your ability to really do blue ocean kind of innovation because it's so expensive.

"I don't think it's any accident that you are currently seeing a lot of the really wild innovative stuff -- including in-game mechanics -- appearing on platforms where it requires just one person to not only program but also self-publish."

That's because not only are the games smaller, but the open platforms require no expensive licenses, says Humble. And the tool sets -- "from Unity to Blitz BASIC to DarkBASIC to Game Creator to MultiMedia Fusion -- do so much of the heavy lifting for a developer that it's possible for even one person to make a fairly sophisticated 3D game," he adds.

He cites what he says are two of his current favorite blue ocean designs: Jonathan Mak's Everyday Shooter, which is "one guy, a guitar track, and a great shooter that I love" and Keita Takahashi's Noby Noby Boy "which epitomizes the fact that the digital market now enables one creator to complete a vision and do it cost-effectively."

Noby Noby Boy

Humble recalls that during his tenure at EA's Maxis subsidiary, when he was helping to develop The Sims, EA CEO John Riccitiello was a strong believer in "coming up with new, cool stuff. And he encouraged us to do R&D. He knew that some of it wouldn't work -- that a lot of it wouldn't work -- but some of it did, including an awful lot of prototyping that we did on The Sims 3. John would always say 'Hey, your job is to surprise me,' and I think a lot of what we did paid off."

In his current position, Humble says he likes to "put the 'lab' in Linden Lab and really do some innovative stuff," which he tries to accomplish with a team of two other designer/programmers.

"What happens is we'll code up a fast prototype or just have an idea, kick it around, and then quickly see if we can stand it up and prove it out," he explains. "Then we decide whether to throw it away or keep going on it."

So far this year he says he's thrown away four ideas and kept two that his team is continuing to develop.

"It requires a lot of discipline to be able to recognize that a wacky blue ocean idea isn't working out and that it needs to be killed," he admits. "But that's the nature of R&D."

And even when projects don't work out, Humble believes there are benefits: "Even with the worst failures, you can learn something or come away with a little mental tool that will make a future project better."

He also sees R&D as the perfect remedy for programmers who have downtime between projects and don't know what to do with themselves.

"Usually they spend their time bug fixing," he says. "Instead, I tell them they've got a little blue ocean time. At worst, the R&D might improve their skill sets a bit. At best, the project succeeds and suddenly you've got a brand new unexpected product to sell."

While R&D may be part-time work for Humble, for Richard Marks it's his lifeblood. Marks is in charge of Sony's R&D department and the man credited with developing the PlayStation Move and the EyeToy.

Marks says he joined Sony R&D mainly to leverage his academic research background -- in avionics and robotics -- in an industry he loves. Much of the same technology used in those fields is now of great relevance to video games, he says, namely in real-time video processing, sensor fusion, and digital filtering.

At Sony, Marks says his teams follow two major avenues: improving existing experiences, and creating new compelling experiences that might appeal to new audiences. Clearly, he adds, success with either avenue is beneficial.

He likes to break down his research into three categories -- "pull," "push," and "stockpile."

"Pull" research is when a product group requests something that "isn't possible," and the researchers are responsible to figure out how to make it possible. In that case, the research is clearly useful, since it was requested, and it is nearly always utilized. For instance, Sony's speech recognition research that resulted from the company's game teams requesting the ability to recognize voice input.

"Push" research occurs when researchers have a new idea that they develop into something that seems valuable -- at least to them. In that case, there is no product group requesting it, so it is often more difficult to get it into a product. Often, the researchers must become technology evangelists and "push" their research into the product groups.

The EyeToy is a good example, since no game team was requesting a camera that would enable gamers to interact with games using their bodies, and the teams really didn't have any way to use it for the games they were already developing.

Lastly, "stockpile" research occurs when a piece of technology gets developed, but it isn't evident that it has any immediate use, and so it is stockpiled with the hope that it may be used in the future.

The camera tracking of PlayStation Move is a good example. During the development of EyeToy, Marks' team created a method for 3D sphere tracking, but it wasn't very useful until they could combine it with the improved PlayStation Eye camera and the high-performance, low-cost inertial sensors that are in PlayStation Move.

When asked how far out his team looks and how much freedom he has to research what he wants, Marks says his "stock answer is two to five years, but it really depends on what part of the cycle we're in. Sometimes we are looking five years out, other times we are looking much closer in.

"While we tend to leave the blue sky research to the universities, Dominic Mallinson (my boss) has always given me quite a bit of freedom to pursue the areas I think are interesting."

And that can range from the "really crazy stuff to the very mundane," Marks explains. "It is a cycle we try to follow: try a bunch of crazy things, then pick a few that actually might have some promise and take a more serious look at those. I think a good researcher is someone who can generate lots of innovative ideas, but then also follows up with the ones that might actually have some impact."

While most developers may not be doing as much R&D as does Sony, Marks believes they are "actually doing a bit of R&D all the time out of necessity, but aren't necessarily able to formalize or generalize it because they are too busy making their game.

"Game technology is always changing, so it is unreasonable for a small game company to invest much resource on R&D that may be obsolete soon," he explains. "The inherent real-time nature of games makes this technology turnover a bigger issue than many other industries; the 'best' method for doing something this year might be completely different next year."

Still, he says, Sony has a strong commitment to R&D.

"We tend to focus on creating new experiences -- and 'new' is generally a quality that has value to our customers. But I don't think that is the only path to success in our industry ... and our approach may not be right for all companies. There is also value in the refinement of good ideas and the creation of products of high quality."