League of Legends' design director talks champion evolution

Greg Street talks balance, new ideas, diversity and more.

League of Legends' champion pool constantly grows as Riot continues to add new personalities to the Rift. Each champion brings something fresh to the game, offering unique abilities, backstories, voice acting and skins for players to choose from. So far there are 133 champions, with Ivern being the latest to hit the servers. But what actually goes into making a champion? I put this question to Greg Street, Riot's design director.

PC Gamer: How does a champion go from the drawing board to our screens? 

Greg Street: Through a process that involves a lot of blood, sweat, and tears. You can think of champion creation as occurring in three phases. The first phase is ideation. The team spends a lot of time locked in meeting rooms rapidly brainstorming. They’re talking not only about cool new ideas, but how a new champion would fit into our current roster. How long has it been since we have made a Marksman for instance? Or does he or she feel too similar to a champion we recently produced? The failure rate for ideas is high here, and only concepts that really get the team excited have a chance of moving forward.

Other abilities have to be thrown out entirely, which can really set back the art work being done on the champion.

The second phase is implementation. The designer on the team will use a stand-in model (typically another champion) and whatever spell effects he or she can cobble together to test out the passive and four active abilities. Meanwhile, the art team will start concepting, then modelling, texturing, animating and ultimately producing the FX for the new champion. 

The third phase is a whole lot of playtesting with a rapid iteration loop. We frequently go through a lot of ability churn at this phase. Some ideas can be salvaged by changing the rules or just the numbers. Other abilities have to be thrown out entirely, which can really set back the art work being done on the champion. As the champion gets closer to completion, we start to work on the reveal to get players excited about the new champion.

What inspires you and your team when you come up with a new champion?

We consider the DNA of the champion team to be the designer, narrative writer, and artist. (Get it? DNA. No?) The idea can come from any of those different disciplines or sometimes even from Rioters outside of the core team. An artist might be brainstorming different silhouettes or animations that we’ve never seen before. The writer may come up with a cool hook in Runeterra for something we’ve never explored. The designer might come up with a cool mechanic that they always wanted to try. 

How do you design champions that can be universally understood and accepted by a global audience?

On the one hand, we’ve had a lot of experience designing champions that speak to League players regardless of where they live. The ability to make flashy plays, bringing something cool to your team, or sometimes a really troll-y voiceover can have some amount of universal appeal. At the same time, we don’t want to pretend that cultural differences just don’t exist. You have to be careful not to boil down an entire culture too much, but we do know that many Korean players have an affinity for champions of high mechanical complexity. Chinese players tend to gravitate towards big, heavily-armed bruisers.

At the highest level, we’d love for every player out there to have a champ that they really identify with. It doesn’t seem like an unreasonable goal to have our champions be at least as diverse as our player base.

Skins are one way we can call out different regions while hopefully still having some resonance in other regions. The High Noon line for example is tied to an archetype that is pretty North American, while the luchador skins probably have broader appeal in Latin America, and the Lunar Revel line has connections to the Lunar New Year, a big deal in Asia. But you never know.


Do topics such as sexuality, ethnicity and gender play an important role when you draft new design concepts? 

Having a game with space dragons, robots, scarecrows, mummies, and giant bears, actually creates plenty of room for diversity. We do talk a lot about sexuality, ethnicity and gender, and what are the right decisions for individual champions or League as a whole. At the highest level, we’d love for every player out there to have a champ that they really identify with. It doesn’t seem like an unreasonable goal to have our champions be at least as diverse as our player base.

How do you nerf a champion without crippling their overall performance in game?

Balance of League of Legends is equal parts art and science. We collect an enormous amount of data from the live game about who is doing too well and who is falling behind. We know that very subtle changes can have a huge impact. For example, just changing the recommended items for a champion can shift win rates by 5% or more. So we generally have to be judicious in the changes we make, sometimes changing cooldowns by mere seconds or attack damage by just a few units.

We try to support niche builds, because that kind of experimentation is a lot of fun for players, but we can’t support them at the expense of game health overall.

The risk there is that you might not reduce say win rate by an appropriate amount, and end up having to make an additional small change next patch and the patch after that. Seeing nerf after nerf can really be crushing to players of those champions. The team tries to really focus on goals - if we think early game poke is too strong, we try to target that specifically, not just change some numbers and see what happens.

In League of Legends we often see certain builds come to the forefront of the game. How do you nerf or buff specific champions without hindering and neglecting the less popular niche builds?

We try to support niche builds, because that kind of experimentation is a lot of fun for players, but we can’t support them at the expense of game health overall. If the niche build is getting played because it’s really hard to counter, it won’t stay niche for long and may soon be the only viable way to play that champion. Sometimes we have to move quickly. When we detect a balance problem in a patch, we may have only a few days to react before our content is locked for the next patch. In those cases we may not be able to come up with a solution that can be implemented safely that solves the problem while also preserving the niche builds.

Is the idea of ‘balance’ where every champion is viable and unique even possible in a game like League of Legends? How do you handle this balance when there is such a huge champion roster?

We try to operate as if there is a Platonic concept of balance that can be achieved, even if it’s a ridiculously gigantic task. The reason this strategy is useful as a mental construct is that it helps us focus on long-term changes that bring us incrementally closer to perfection. Not taking this big picture view would make it easy to justify change for the sake of change instead of change that makes the game better.

League is an extremely complicated game and individual player skill is a huge factor. We don’t obsess too much about achieving 50% win rates for every champion at every skill level. Instead, we focus on balance problems that keep players from playing the champion they want to play, because they feel like the tuning of the game forces them to play someone else.

Have there ever been cases where you ran into a bug and decided it was fitting for the champion and kept it? If so how common is this?

It depends on the nature of the bug, but it is pretty common. We’re not crazy about supporting them if they violate our design values, such as a bug that made a champion a lot harder to counter than intended, or a bug that was so obscure that players really wouldn’t be able to understand what had happened. Two fairly famous examples are animation cancelling for Riven and ward hopping for Lee Sin. We didn’t intend for either of those champions to be played that way, but you could argue now that many of those players stick with those champions because of those features.

We definitely don’t want to be in a situation where Riot is dictating to players who or what a champion is and how they should be played.

Who do you believe defines a champion’s purpose?

At the end of the day, nobody solely owns a live champion. Before a champion is released, it absolutely belongs to the team that is developing that champ. If you think of a pre-release champion as a plant seed, they are the ones deciding where to plant the seed, how much water to give it, and so on. But once the seed is in the dirt, you have a lot less control about whether it blooms or withers, how tall it gets, whether other plants crowd it out, and so on. Champions definitely take on a life of their own “in the wild.” 

Players have a lot of ownership over the champions, because they are the ones who spend so much time with them. But we also need to step in when a champ needs a little redirection, whether that’s from a balance perspective or a frustration perspective or whether their art just feels out of date. We definitely don’t want to be in a situation where Riot is dictating to players who or what a champion is and how they should be played. We know we don’t always get this right; it’s particularly dicey for champion updates, where we sometimes have to say “We know you really love playing this ball of stats that stomps on lanes, but we really believe the game will be better for everyone if we steer the ball of stats into something else.” 

What happens to the ideas that don’t make the cut? Do they get locked away in Riot’s basement or do they get reused elsewhere?

It can go either way. A lot of the ideas get sealed into the vault, and occasionally someone brave on the champ team will venture into that vault and try to take a stab at an idea that nobody has been able to make work yet. Sometimes the champ design reaches a dead end because it’s a pretty terrible idea, but even in that situation, I’d love to see someone with a great idea make a pitch for how to fix those problems.

We recommend