Torigaa

For some reason you cannot comment their blog posts?

Are they afraid of what people think about this champ?

Nah, we're actually working on bringing back comments (hopefully this year). It's something we lost we launched the new website, and it's taking a bit longer than expected to re-add them. But we wanna see your comments. BRING ON THE FLAME :')

ContentDetective

This is riot's equivalent of LS saying "nemesis don't have a champ pool issues, in fact he has a champion ocean" and proceeds to lose because of champ pool issues. She's braindead in the sense that after people play enough games on her they never make punishable mistakes. It's a mistake to have put her in the game.

Hi, we respond to this in the post. Yuumi should indeed be more punishable, which is a counterplay issue and makes her non-interactive. A champion can be both hard and non-interactive—that means opponents are likely to think she's "braindead" when she's actually not. That's not an excuse, but instead a more accurate statement of the problem, which should still be resolved as much as possible. If a champion has to play 34-dimensional solitaire to win but is permanently invisible and untargetable and never leaves the fountain so opponents have no input into their outcomes, that's no good.

Kadexe

They weren't as popular before they were nerfed either.

This is correct. Power is definitely impactful to both breadth and depth, but nowhere near that impactful.

DestructiveParkour

Quick question: if two champs have the exact same pick rate, but one is picked only by mains and the other is picked only by first-time players, will they have a different depth score?

Also, really appreciate the point about being hard but having no counterplay- that's why I (and a lot of other people) hated the akali-irelia meta.

Yes, they'll have a different depth score. That's what depth is meant to measure.

To your second point, absolutely. It's easier to identify this lack of counterplay when the champion looks easy or the result is consistently achievable across all skill levels, but Akali and Irelia at high levels of play during that period often completely shut out their opponents' options.

JohnnyTruant_

I have a question about the champion mastery curve brought on by the mention of people twisting stats to fit their agenda, if you don't mind me asking.

It made me wonder, is it even hypothetically possible for a difficult champion that for whatever reason in this scenario is only skilled players to have the same mastery curve as an easier champion that for whatever reason in this scenario is only played by not skilled players?

I know in practice skilled players will use the easier champ if it's good enough, and lesser skilled players want to try out complex champs too, I'm just wondering if it's even theoretically possible for similar graphs to come from different situations like that, or am I misunderstanding that particular graph entirely?

I've never considered this because it's definitely not realistically possible. It is theoretically possible for a high-skill champion to have a flat mastery curve if every person who played that champ is already skill-capped on it from their first game. So, your answer probably depends on how you define "skilled player." If it's "high MMR player," then no, for a number of reasons. If it's "someone who has mastered the champ already (somehow, even though they just came out)," then...maybe.

Aotoi

I hope you see this comment: have you folks looked at a very similar character from dawngate called Mina(i think)? She also became untargetable by attaching to her allies, but she took a percent of damage they took, and was encouraged to detach with cc on detatching. Obviously yuumi can't be the exact same as she currently exists, but mina never felt as bad to play against because you knew you were punishing her when you hit her host.

Yes, of course we considered and even tested a number of mechanics like this, both in original develompent and when I was making the large set of balance changes to Yuumi. Mina came up in conversation more than once, as well. They generally felt absolutely terrible to play, and generally betrayed the core conceit of the champion. They also often led to a lot of resentment for the person you were on, because they were actively killing you with everything they failed to dodge, and that was a pretty huge miss on the feel we were looking for. Ultimately, my takeaway was: was a fine path to look at, glad we tried it, it really was not even close to working in any incarnation we tested.

I think a totally different champion could probably exist with a mechanic like this, but Yuumi is quite far off from being that champ and (more speculatively) that champ also likely doesn't hit the same audiences that we aimed for with Yuumi. That's a good question though, thank you.

betweenskill

I think a big portion of complaints I’ve seen in game, from friends etc., is that Yuumi has to choose to be interactable with by the enemy team.

An enemy has no tools to force Yuumi to be interactive except solely by killing the champ she is attached to. If Yuumi wants to sit on a juggernaut or tank and spam E late game she can and there is absolutely nothing the enemy can do about it.

Yes, it may not be the most optimal play, but considering the vast majority of players are Gold and below, it becomes extremely hard to feel like you have options against her.

If there was some way to force her off of a champ, through aoe spells or hitting enough hard cc on her attached champ or something to allow champs to detach her, even if its difficult to do, would give some tangible agency back to opponents.

Her problems tend to stem a lot from perception rather than balance, which seems to be a running theme with a lot of newer champs and reworks lately that people tend to dislike to play against. Yes they may be balanced, but they don’t feel like the opponent has choices or agency when fighting against them compared to most other champs. A lot like your Akali-Irelia example.

Yeah, I generally agree with that. The tack we're trying right now is making her significantly suboptimal when she refuses to give opponents a window of interaction. It's a bit tough when Presence of Mind exists, but we're giving it a go regardless.

To use an extreme example of why I don't think this is a qualitative problem and is instead an issue that can be solved with proper incentives: Xerath could play by only using his Q from out of range, and he would be borderline uninteractable. However, when we gate a bunch of his power behind actually getting closer and therefore becoming interactable, he becomes tolerable to play against. Right now, too much of Yuumi's output is accessible without detaching. In an ideal world, which hopefully we are moving towards, you would look at a Yuumi that never detached similarly to a Xerath who only uses his Q and R: "Why are they playing so incorrectly?" rather than "Why can't I do anything about this?"

I understand that we may be a ways away from this world, and we will probably need more changes to get there, but that's the direction we hope to move. Thanks for the thoughtful comment.

betweenskill

Thanks for the reply! I do agree with a lot of your guys' ideas.

I think it's just a psychological sticking point, especially among those not in the top 5%+ of players, when it feels like you can't do anything to force a decision on her part.

It's an interesting thing how the skill of the players involved doesn't just change the power of the champ, but also fairly changes how the champ feels to play and play against. It doesn't seem to be that way with many champs to the same extent, but I feel a lot of champs that get complained about (warranted or not) tend to be champs that the majority of lesser skilled players feel bad playing because they feel that they don't have options.

That's particularly why I was just throwing out ideas for some hope of forcing interaction directly for players that are not constantly thinking about high-level champ mechanics like "weaving" Yuumi's passive for mana sustain and more thinking "Why can't I ever hit her! She's so unfun!". It wouldn't even have to be especially easy to do or consistent, as that would then make Yuumi feel powerless to stop enemies from knocking her off. Just something that is a "goal" an enemy could work towards to try to force that interaction in some way, simple and clear cut for both the Yuumi and the enemy to see as something to either go for or to avoid.

To tie in the incentives you mentioned so that some might enemies feel like they have some small amount of agency, just some ideas:

Maybe she gets some sort of bonus for the champ she's attached to the longer they are in combat, but 2/3 hard cc spells in quick succession knock her off.

Or perhaps she takes reduced AOE damage when attached, but if she takes a certain threshold of damage she is knocked off, making it riskier for her to sit on a tanky juggernaut but still able to pull it off if she times it properly. Enemy champs would have to choose between focusing down the big bad motherfudger before she jumps on, or waiting to bait her in and knock her off in the middle of their team.

Or perhaps she loses healing potency on the same target multiple times in a row but gets a bonus when she jumps to a new target, making the "I sit AFK on a melee carry and press e" playstyle.

Or maybe she has a slight delay when jumping from or jumping to an ally champ where she is still targetable, allowing quick reflexes or predictions from the enemy team to pull her off or at least put a bit of poke on, forcing her to either stay on a particular champ or jump off to use her passive and be able to heal herself back up. She wouldn't feel any different to play with mechanical timings, but jumping onto a champ in the middle of the enemy team would be much riskier, but still possible if timed well. Again, the enemy team would have pretty clear decisions to make around this small change.

Anyways, ignore my washed-up former plat redditer ramblings as they probably aren't worth much. But thanks for taking the time to read and reply, and I hope you guys can find a way to reach your goals with her :)

P.S. If you use any of my ideas I reserve the right to feel privately smug and add it to my fake internet point bank vault. And if you don't well... that's probably a better choice for you haha.

I tried exactly that damage threshold -> knock-off mechanic briefly. It's possible that it could have worked out alright with some new UI/UX for both the Yuumi player and her ally, but overall it was really jarring and confusing. Yuumi felt like she was randomly falling off even with some pretty loud prototype visuals/signaling, and her ally was just not interested in playing the "keep Yuumi from falling off" game because they're trying to play their own champ at the same time. It's a pretty complex space, because none of our characters demand so much coordination in a tense competitive multiplayer game. Yuumi really stresses the social contract folks sign when they queue up.

Generally, most of the things you're suggesting might work, but they're pretty laser-focused on solving this specific problem, and as a result don't respect the reasons Yuumi is successful and worthwhile. We have to consider the fact that Yuumi is, as the article mentions, a highly resonant champion that people love, and a lot of that is due to some of the same things that some players really dislike about her (like the fact that her in-combat pattern is much more strategic and much less mechanical than most champs). We must reckon with that in our solutions as well. For instance, I often see suggestions that would tell Yuumi she has to hop from person to person really frequently to get anything done. This would definitely make her harder, and would make opponents feel like she was working for her output. I specifically aimed to make Yuumi easier overall, though (at least to hit her reasonable output floor), because there's a serious audience mismatch at the product level between "League's only cute cat champ" and "hardest champ in the game." Hopping from person to person constantly also creates a heavy high-MMR and especially pro skew, because un-coordinated allies are very likely to randomly strand Yuumi on an undesirable target with nobody else in range, unable to achieve reasonable output. Yuumi is so dependent on her allies' movements that pushing this property any further leaves her almost entirely at her team's mercy. That's a recipe for a champ who is 100% p/b in pro at 46% winrate, or balanced in pro at 42%. Also, Yuumi is designed (in both gameplay and theme, including VO etc.) to foster a feeling of "hey, you're my buddy and we're going to work together to both keep us alive and take down our enemies." She's a bit mercenary and will hop off to save her hide, because she's a cat, but it's an additional cohesion loss if her optimal play isn't usually picking a person and trying to win alongside them, and sharing in the ownership of their victories in fights. When Yuumi and the person she's on carry a fight, both of them feel really good about their partnership, and that's important.

This wall of text isn't even to say the ideas are bad, because we have all the same ideas in dev at some point and usually try them out. These are just the learnings we come to, and some of the many things we have to consider (and why this sort of balance is never quite as cut-and-dried as Reddit thinks it is). Of course we still have to actually make progress towards making every champ a healthy addition to League, but we can't throw the baby out with the bathwater. So, that's why it sometimes takes a while and we'll often wait a bit between changes because overall sentiment can change pretty dramatically with smaller changes and time (look at Zoe—if I recall correctly, she was supposed to be deleted too?).

Anyway, thanks for the interest. I'm glad we were able to get this article out, it's nice to pull back the veil and also open up opportunities to talk shop like this occasionally.

JohnnyTruant_

Hey, thanks for the response.

For skilled players, I was more referring to the physical ability to pilot the champion to its fullest potential. So in this hypothetical somebody with a physical disability that impacts their ability to do certain combos or interactions might not be "skilled" in that sense, or somebody that gained knowledge through pro play/streams but their PC gaming experience is non-existent and they just can't operate M+KB very well.

Whereas there could be a lower MMR player that doesn't know when to push, how comps work against each other etc but has great reaction time and dexterity and whatever else.

I wouldn't know if it's possible to differentiate those types of players when it comes to a champion mastery curve, now that I think more about it. I might just be asking the impossible even as a random shower thought type hypothetical. Thanks for indulging my curiosity.

Maybe if I ask about it in a more general way you could more easily apply (if it's possible to do so)

Thinking something along the lines of, do grade 10 students learn grade 10 math at the same rate grade 4 students learn grade 4 math? Substituting the age gap for "ability to pilot a champion optimally", if that's possible.

Again, thanks for helping satisfy my curiousity, even if the result still comes up as maybe I appreciate the communication on the subject in general, not just towards me.

It sounds like what you're describing is somewhat approximated by MMR? So, high-MMR mastery curves include players with greater mastery over all of the game's systems. Of note, I don't believe high-MMR mastery curves look different in any consistent way, though it's been a while since I looked at that in particular. A big part of this is that the skill and knowledge of your opponents also scales up commensurately.

That is, a Challenger picking up Akali for the first time on release would still have a much steeper learning curve than a Silver player picking up Garen for the first time on release, because, while their skill level is significantly higher, there is still more room to misplay on Akali than on Garen and more champion-specific mechanics and knowledge that must be leveraged to yield winrate. Challenger opponents have equally mastered their champions and will exploit errors to the fullest.

I hope this helps at least somewhat, although I can't shake the feeling that I'm still not fully grasping the question.

vooffle

How about balance changes? As a top laner that made attempts to pick up both Taliyah and Aurelion Sol, my biggest frustration with both was having to read patchnotes every time I felt like playing them. Furthermore, those balance changes impacted how the characters played more than with other champions. I picked the new Irelia up while she was getting nerfed too, but all the gameplay that changed with her was the number of passive stacks changing and not being able to 1-shot quite as many people at level 6.

Don't believe patch fatigue has much of a measurable impact on breadth/depth, but we still keep it in mind (even more so now, with a more rigorous framework) as a less tangible outcome to try and avoid. It's always a push-and-pull, with factors for and against. Sometimes a champion needs changes more than we need to avoid patch fatigue, and some players will eat that cost. Unfortunately, it's almost impossible to do any real change at all to a game of this size without causing some segment of players some level of inconvenience or worse, which is a reality that I'm still not really happy with but have had to learn how to stomach over time...

SuperMrBlob

As a bit of anecdote, I 'clicked' with new Akali instantly, had maybe 50? games on her very early on, wanted to put in the time to really master her. But patch fatigue, constantly expecting her to be nerfed AGAIN, and the priorities re: balancing her for pro really put me off her so much that I don't play her at all anymore. At one point I just went 'enough is enough', I don't want to waste time trying to get good at a champion when she's just too volatile. Sucks :( Really think your (I assume not your personally? Riot's) approach with her was flawed.

Well, it's not my personal approach but I definitely had a hand in it, and it's definitely possible that it was flawed. The other possibility being "it was just hard, and we got it wrong," which doesn't necessarily indicate a philosophical flaw in approach but rather execution error. Or, maybe it was some of each. Either way, I think it's pretty clear that Akali's post-release balance was not a success.

We absolutely must balance her for pro unless we discard the notion of pro balancing, so there's not much to be done there re: priorities. In retrospect, we could have acted faster on bigger mechanics changes to avoid patch fatigue from constant changes. I think that's been a problem in the last couple years with new champions in particular, and we have recently somewhat revamped our approach to avoid, at least, the "nerf them every patch to keep them balanced as players learn them" outcome which should save some volatility on this sort of steep mastery curve champ especially. It's always a tough balance, when you don't truly know how far off a champion is from being balanced in pro (pros sometimes drop things without major mechanics changes) and you want to preserve her unique and incredibly cool mechanics. There's a lot of tension, and it's easy to fall into the trap of hoping that a couple rounds of high-vs.-low-MMR skewed numbers changes will do the trick, because sometimes it actually does and you've had your cake and eaten it, too. I think the era Akali/Irelia/Aatrox era has taught us

Anyway, the moral of the story is: Akali has been a really tough case and I think that, while we made mostly reasonable decisions at most junctures with the information we had, I wish we could have done her more justice. I'm sorry it caused you to bounce off the champion, because that's definitely a failure. We'll keep trying to get better, and I hope that our decision to release champions at what we believe is long-term-balanced will pay off for players like you who want a more stable base to begin learning from. Thanks for the measured feedback.

...I say all this, but I'm not actually balancing the game actively anymore—however, I'm confident in the team that is and work with them closely, and I believe they're aligned with this viewpoint and have picked up similar learnings.

betweenskill

Being able to spitball stuff like this is what I miss from the old, old boards with you guys. Sometimes it’s just fun to be able to talk cool stuff like this with the devs, and if you guys manage to get a shred of a good new idea, or just some non-Riot talking about your efforts at work, all the better.

It’s what made me want to get into game design in the first place back when I was a teen, though after seeing the sheer talent on display at Riot it made me rethink my plans into an indie direction for any chance at success facing that job competition. Reddit memes aside, I couldn’t compete with what you guys do . Thanks for taking the time, all 200 years of it!

Yeah, I personally preferred the old boards as well—I think the upvote structure of both Reddit and the new (and now defunct) boards leads to the spotlight being hogged by inflammatory and extreme opinions which are really difficult and often not that productive to engage with, and I often stumble on topics I do want to engage with too late to even make a post that will be noticed. Good luck in your career, maybe I'll see you at Riot in the future :)

SeizeTheKills

So how about giving her the (old) Soraka treatment, gate her W with her own health pool instead of just mana?

That way her health pool can be affected while she's attached, since damage taken by host requiring heals will hurt her health pool. Which will also mean detaching is significantly riskier since she'll be lower health and if the host dies despite the heals it would mean Yuumi is very likely to also die (almost instantly) making it a higher payoff to the other team to focus down Yuumi's host.

Could possibly work. The flipside to consider is that it it becomes too risky (and too close to certain death) to jump out, Yuumi will refuse to do so and will instead just content herself with having almost no output, which both makes her less interactive for opponents and less fun to play. I think being at low health is pretty different on feel than being at low mana and having other mechanics that make being un-attached risky, like the CC trigger. Right now, I'd say the goal is not to make jumping out riskier but to make her actually do it with much greater frequency.