A look at the LoL champion design process

"It's so much better to hear someone complain that you're overpowered in person," Game Designer Colt “Ezreal” Hallam says, prepping me from over my shoulder for a round of League of Legends. It's 5 p.m. at Riot's Los Angeles headquarters: time for the daily developer playtest. All unfinished work is pushed aside and the trash talk starts flying. I know better than to make any bold predictions of victory, because I'm going to get my butt kicked. I'll be playing a champion I'm unfamiliar with—because it's never been shown to anyone outside of Riot.

You see, this isn't the LoL you and I play at home. This is the development build, loaded up with four future champions (in states ranging from ready-for-release to barely implemented) and plenty of behind-the-scenes stat tweaks currently being tested for balance. Right now, I hope it's imbalanced in my favor.

The battlefield is anything but certain when I join the match as the living-tree champion, Maokai (pictured above). In fact, I don't even look like a tree—Maokai's stats and abilities inhabit the model of an existing champion, because the modelers haven't built one for him yet. The "acorns" that I lay like spider mines around the map currently use tiny versions of yet another champion's model. The whole 5v5 match is hilarious and crazy fun, mostly due to that the fact that no one knows for sure what these works-in-progress are capable of in the current build.

Afterwards—yes, I'm skipping the part where my team lost—everyone crowds into a conference room. One by one, they describe their experience and offer suggestions for how to improve what they just tested. A few of my opponents mention that my acorn spell did negligible damage to them, but that my teleport rooting spell was very effective—perhaps too effective. Another chimes in that he couldn't believe it when I accidentally (or heroically, as I tell it) intercepted Caitlyn's Snipe and saved our healer's life. Suddenly, a debate erupts: should it be possible to bodyguard-dive in front of the sniper's shot? Should it pass through everyone it hits? Should it do diminishing damage as it progresses?

The team of artists, quality-assurance grunts and development leads don't reach a definitive answer but they do reach an agreement on what to test next. This is development by playing, and if something doesn't work, they'll scrap it and try something different. The Moakai I played today will almost definitely be an entirely different beast from the Moakai you see when he hits live servers in February—that one will be much better.