Information Security & US & Life & Love & Fashion & Hacking & Passion

What if, like Dr. Frankenstein, you could stitch together the best bits of your favorite games to create a guaranteed hit? What if you knew adding a few key components to a title would boost its sales exponentially or extend its shelf life? What if there was a recipe for designing a perfect game?

Those were the answers I sought when I asked Electronic Entertainment Design and Research ( to run an algorithmic regression on a mock game called "FutureNot" that I whipped up. But such a recipe does not exist, says EEDAR President Geoffrey Zatkin. Games can't be assembled like a Mr. Potato Head doll--swapping out eyes and ears until you find the most agreeable feature set. But if you look at the whole game, it is possible to project sales and determine whether it is worth including specific ingredients.

This is the sort of stuff the Carlsbad, Calif.-based research company does all the time for clients like Electronic Arts (nasdaq: ERTS - news - people ), Activision (nasdaq: ATVI - news - people ) and Ubisoft. Since 2006, the 20-person firm has been busy compiling sales data, classifying games and cataloging key components--everything from the art style to the gender of playable characters.

With its database of 6,000 games, EEDAR analysts can use historical and competitive data sets to project sales for upcoming releases. It can also determine how certain developers and publishers, marketing budgets and release dates align to affect game sales.

"Every game I have ever worked on, we've gone in blind as to which features would sell the game better," says Zatkin, who designed games for 11 years before co-founding EEDAR. Not knowing whether it would be worth an extra $500,000 to design a multiplayer mode "would scare the crap out of me," he says.

Only 4% of games that make it to market actually make a profit, he says. About 60% of a game's budget is spent reworking or redesigning a game. Armed with all this data, companies can make those tough calls early in the development process.

As for "FutureNot's" potential success, Zatkin analyzed it and says it's "a surprisingly decent game" that could sell 216,000 copies in six months. He notes that most developers won't make a blockbuster like "Halo 3," which has sold more than 3 million copies so far.