search slide
search slide
pages bottom

Riot talks Taliyah and balancing new champions

Daniel ‘ZenonTheStoic’ Klein started life at Riot as a community manager, but has since moved into champion design, where his efforts have brought us mechanically complex members of the League of Legends cast such as Azir, Tahm Kench, and Lucian. His latest creation, Taliyah, landed on the Rift around a month ago, and much like Aurelion Sol before her, she’s an unusual mage with a brand new toolkit for midlaners to toy with.

Taliyah’s first few days outside of the Public Beta Environment, where players test new features pre-release, were unusually tough for a new champion. Stagnating at a 37 percent win rate, The Stoneweaver was severely lagging behind her peers. Since then she’s been on a journey — not a magical one, she’s not Bard after all — but a very real transition as the community began to master her rock-slinging skill set. We picked Klein’s brains about this bumpy launch, and how champions are kept in rude health post-release. No designer can predict every result or interaction, and as Klein tells us, ensuring that a champion has a bright future in competitive and solo queue play is more complex than simply staring at a screen of metrics.

Both Lucian and Azir are top-tier picks in competitive play, but their dominance is irksome to Klein, and something he wanted to avoid with Taliyah. “To us as designers it’s actually not a mark of honor to see our champion dominating in terms of pick rate in pro play. It’s actually something where you go ‘oh, I’m going to have to look at this guy and see what we can do to fix that.”

He admits, however, that in her case, the team may have initially swung too far in the opposite direction. “We messed up on [Taliyah’s] release; we thought she would be stronger than she actually is. We do miss the mark quite often and honestly that’s fine — it means we’re taking risks and making interesting new champions. I could make you a champion right now and tell you within two percent where their win rate was going to be, but the chances are that they would be pretty boring and we like to make champions that keep people engaged.”

Each champion goes through rigorous testing before they even make it to the PBE, but Klein has an inkling that it’s here the mistake in balancing was made. “My personal theory is that we trusted the playtesters who had three months or so to get used to her, too much, and we basically balanced her for someone who had been playing for ninety days. Obviously that made her too weak. Even though she’s in a decent spot now there is a chance that eventually down the line we’re going to have to nerf her, as the player base as a whole learns how to play her.” Such anomalies are hard to detect after the fact, much less predict before they happen.

The team hasn’t yet identified exactly what was causing Taliyah’s woes, but when discussing the mage’s weaknesses, it becomes clear that one dynamic that was coded into her kit may be to blame. “One of her pronounced weaknesses is that she needs to keep moving to fresh ground. If she is in a prolonged siege, whether offensive or defensive, she will run out of fresh ground to use and her usefulness will just fall off a cliff. That is what we call a non-class based weakness because other disruptors don’t share it. Anivia can siege you for hours: she can stall the game till it runs to 50 minutes, as I’ve seen many times in the past in professional play. Taliyah cannot do this, she will run out of fresh ground and when she does she doesn’t bring a lot to the team.” The worked ground dynamic has proven itself divisive within the community, with many players finding its limit on Taliyah’s damage output too restrictive.

Undoubtedly some players would have preferred more consistent Q, as it’s Taliyah’s primary damaging ability, but Klein explains that weaknesses are as important as strengths in creating a compelling champion. “We made the prediction that people would hate worked ground. As a matter of fact, if people hadn’t I would have been very very certain that I screwed up. This is a concept in game design I call ‘sharpness’. When you define a good and a bad outcome, if they’re not sharply contrasted against one another the whole thing becomes muddy. The example that I like to use is Poppy’s Heroic Charge which only stuns you if it carries you into a wall. If we made it so that Heroic Charge always stuns you for 0.75 seconds, but if you hit a wall it stuns you 1.5 seconds, the use case for Poppy would be much less clear. Should you still charge at them even if there’s no wall behind them? You don’t get the same stun duration, but maybe I can use it more often and that makes up for it? That’s what I mean when I say ‘muddy’. It becomes much less clear how to use an ability optimally.”

It was vital, according to Klein, to apply the same principle to Taliyah’s Q, Threaded Volley. “If it was sort of OK to use [her Q] from worked ground and, let’s say it cost half mana and you didn’t have to keep moving to fresh ground all the time, all of a sudden as a Taliyah player I’d be asking myself whether I even need to put myself at risk. Should I really hold the ability until I reach fresh ground? Or do I just spam it from worked ground because it’s cheaper? Right now it is very very clear: if you can, you will find fresh ground and use the ability from there because it’s ridiculously more powerful, three times more.” No other champion in the game is forced to remain so mobile in order to retain their damage output. Midlaners are used to roaming to pick up kills, but in Taliyah’s case, such behavior seems non-negotiable.

Clearly, players in Diamond will immediately be more comfortable with such enforced lane drifting than those in Bronze. Riot must always take into account the wild differences in player skill when assessing a champion’s general health on the Rift. The problem is, where relatively unskilled players might be failing miserably, those at the top of their game may be thriving. Klein doesn’t necessarily see that as a conflict. “It’s fine that champions have different win rates between high and low ELO; we can actually learn something about the champion by examining that difference.”

“Generally a champion that has a high win-rate in low ELO and a low win-rate in high ELO either does not have a lot of unique ways to play or has universal counterplay available to them that the enemy just becomes better at engaging the higher you go in ELO. Whereas the other way around where you have low win-rate in low ELO and high win-rate in high ELO, we sort of like that as long as it’s not too extreme. If it’s within three or four percent then that means that there is something unique about the champion to master that people in high ELO are much more likely to have grasped.”

Taliyah’s rocky start, pun absolutely intended, has somewhat smoothed out now that she has had time to settle in, but that doesn’t mean that Riot isn’t watching her closely. “I personally feel like the numbers that she’s at right now is dubious, if not just outright too strong. She’s currently sitting in our internal data at something like a 48 percent win rate which I think is great. More importantly, her ‘first played’ win rate is at 45 percent, it was 37 percent when she released, which is abysmal. The experienced win-rate, so the win rate if you have played about fifty games [as Taliyah] is possibly already too high. But it basically means that if you invest those fifty games into Taliyah you’re going to be a force to be reckoned with.”

It sounds initially like numbers and metrics alone rule the roost at Riot, but Klein is keen to stress that the team’s approach is far more holistic than simply staring at screens of statistics. “The buzz word is that we’re ‘data informed’, not data driven. We don’t just look at win rates and try to get them to a number that makes us happy, that’s not how it works. Win rates have a strong power to indicate what may or may not be wrong about the champion, but so do anecdotal stories on Reddit. For instance if people on Reddit are saying that X is the problem then that’s very interesting to us. What that means however is that we don’t take their word for it, we don’t just say, ‘you said X is the problem, therefore we must fix X,’ we say ‘I can understand why that might be the problem, let me go and watch fifty games of Taliyah and see if it really is.’”

Listening to the reactions of community with so many different voices, often with completely contradictory opinions, and stratospheric expectations of every champion, is a hard job. It seems like every passive ability on new champions always gets slammed by the Riot Police, but Klein doesn’t let this get to him. “I love that people get upset about passives or that people get upset about anything at all. The worst thing that could happen is no one giving a shit about what we do. However, and this is going to sound a little self-congratulatory: not everyone is a game designer. Not everybody has the understanding of exactly what things do. Taliyah’s passive isn’t really relevant in the laning phase because you’ll constantly be taking and dealing damage so you’re constantly in combat, but then, if you roam well, which means moving through the jungle, you get a whole bunch of movement speed for free, so we’re enforcing a gameplay pattern we know is great.”

Klein speaks with great confidence on the topic, and it’s clear that even though Taliyah’s launch didn’t necessarily go to plan, the design team are taking care of lovingly her behind the scenes. However the real test of a champion’s impact, and success doesn’t come in the weeks following a release, but years. Whether The Stoneweaver will prove a popular pick, make an impact on pro-play, or shove the meta in an interesting direction remains to be seen.

Leave a Reply

Captcha image