League of Legends dev details missteps

Who was there: Riot Games design director Tom Cadwell delivered a talk titled "Designers Are Humans Too--Causes of Poor Design Decisions."

What they talked about: Cadwell started off by explaining that poor design decisions can come from really good designers as much as, if not more than, from poor ones. He identified a few reasons talented designers make serious mistakes. For one, the processes a developer has in place can impede good decision making. Simplifying the point, Cadwell said it doesn't matter how good designers are; if they are told to make a AAA game in a day, it's not going to happen.

Second, designers can be influenced by social or emotional motivations in their lives, and that impedes their abilities as a designer. Finally, designers accept unnecessary constraints that can convince them a compromise is necessary when, in fact, a different approach would get around the supposedly inflexible constraint.

Getting specific, Cadwell told the story of a character named Omen, which was planned but never included in the game. Omen was a quadruped demonlike creature that wasn't great, but few people on the team thought it was actually a bad design. There were some reasons it didn't fit. It wasn't clear from the design if it was a ranged character or a melee character. It also had some overlapping design with other League of Legends heroes, and about halfway through development, the team realized it wasn't getting any better. Eventually Omen was scrapped in favor of Riven.

The real problem with Omen was that the management didn't communicate to the team just how important it was for everything to be exciting. They hadn't made it clear what the standard for "good enough" was, and they didn't actually lay down the law and say "no." The problem is it was difficult for people on the team to say Omen was bad, so Riot needed to make it easy for people to say it wasn't good enough. What they should have been doing was asking the decision makers on the team if they would think this character would be a 10, Cadwell said. And nobody thought it would be.

The next problem Cadwell talked about was "The Grand Unveil," which touches on a problem with iteration. The Grand Unveil happens when the developers ask an artist or designer to see what they're working on and get a response of, "Oh, it's not quite ready yet." The problem is, early iteration is important, and that process eliminates valuable feedback that could keep things on track. While the situation was a little different, on League of Legends, Cadwell said the Kayle and Tryndamere remakes were delayed because there wasn't enough play test time for the team to get feedback on them.

The solution is to treat the first crack at anything as "a crappy first draft," Cadwell said. If the expectations are set low, there's very little downside for people seeing the output and giving valuable early feedback. On top of that, Cadwell said it's important not to let people hide their progress and to ensure that early peer review has no negative repercussions so there is a mistake-forgiving culture in the studio.

Cadwell's next trap was "Too Awesome to Cut," which is typically when the team members have an idea they love (like a shark with a laser beam on its head) and they try to bend the game over backward to make it fit. When developers get excited about a thing like that, Cadwell says they invest emotion and lose objectivity regarding it. On top of that, when the entire team is excited about it (since excitement is contagious), nobody wants to be the buzzkill to tell everyone it's a bad idea.

For example, Cadwell pointed to League of Legends' Shaco. Early on, Shaco had a permanent stealth ability that was clearly not going to work after a few days of play testing. However, the team was so married to the idea that it wasted an entire month working on it before settling on a tweaked version of the ability.

The solution is to make sure decisions are reviewed by peers who weren't involved in the creation of the idea and so aren't "interested parties." The team should also structure processes to identify the opportunity costs of pursuing these ideas to put the choice of including something into a new perspective.

Cadwell then moved on to talk about the problem of forgotten goals, as evidenced by a Monkey King character in League of Legends, based on the mythological Chinese character. Riot wanted to make a character the global audience would enjoy and one that was true to the Monkey King lore. Speaking generally about League of Legends, Cadwell said that the team wants a variety of complexity in characters.

When the Monkey King was first devised, he had a clone following the player around and mirroring all of his moves, which they found incredibly complex to control. However, since the original goal with the Monkey King was to bring new players into the game with an accessible character, the team decided the original goal had been temporarily forgotten with the Monkey King.

To prevent those goals from being forgotten, Cadwell said it's important to check the work in progress versus its original goals from time to time. Beyond that, team members need to be crystal clear on what those goals are and trained to look for deviations from those as they arise.

Creative fatigue is also a major problem, Cadwell said. It usually manifests itself a couple of ways, starting with an inclination that when a lot of creative effort is required, people schedule really long meetings to get them done. However, psychology shows that creativity tapers off after a short time, so three-hour-long meetings are just going to wear people down and result in poor ideas.

Another way creative fatigue occurs is when the team is given a bunch of constraints (it can't be blue, has to be a human, can't be more than 400 polygons, and the like). Eventually, he said there's no wiggle room, and all that comes out of it is a bunch of mediocre solutions. He pointed to Poppy the Iron Ambassador in League of Legends as an example. When designing her, the team gave her three working abilities but was stuck on the fourth "ultimate" ability. Eventually she was given "diplomatic immunity," which Cadwell confessed is a "pretty problematic" ability. In retrospect, he said the team should have made one of the basic abilities optional and instead replaced it with an ultimate.

Finally, Cadwell talked about the problem of designers designing for themselves. It's a common thing for designers to do, but Cadwell said it often winds up too complicated or niche, and it just results in lower-valued features for the players to be prioritized over higher-value features. Cadwell did this himself in League of Legends with the "inverse power law of ninjas." For every ninja in the game, each ninja lost a hit point (the joke being that in movies, the more ninjas they have, the more fodderlike they are). Cadwell got a good chuckle out of it, but it didn't make the game better for players, and there were bugs that cropped up that made it not worth it at all.

Quote: "We all know what we like, and sometimes that's not what our audience likes. Even as you become experienced as a designer, you have to worry about this because the audience changes."--On the dangers of designers designing for themselves instead of the audience.

Takeaway: Mistakes happen, and people make bad decisions. Even experienced and talented designers fall into these traps, so it's crucial that a studio have a variety of processes in place and peer review to keep problematic design decisions from becoming time sinks and hurting the final product.

[UPDATE]: This story originally appeared under the headline, "League of Legends dev admits to Bevy of Blunders." It has since been changed as it may have misrepresented Riot Games' design director Tom Cadwell's comments during his session.