Taliyah's designer gives us a behind-the-scenes look at her creation

Riot's latest League of Legends champion, Taliyah, was released just a few weeks ago in patch 6.10. While many players have spent the time since becoming acquainted with the new rock mage, champion designer Daniel Klein has been spending time with Taliyah in various iterations for many months now.

We spoke with Klein last week about the process of creating Taliyah, some of the designs and powers he thought up but didn't use and where we should expect Taliyah to go from here.

Rift Herald: How does the champion creation process begin? I’m sure it’s a little bit different from champ to champ, but let’s use Taliyah as an example. What was your initial goal that you had as you were going into creating this champion?

Daniel Klein: The first thing we have is something that we call fantasy slotting. A number of leads sit down together and go over the prospective next four or five champions that we will release until whichever champion you’re about to work on comes up. Then we say, "Hm, in the next year, we will not have X." Or "It looks like we haven’t done a human being in a long time." Or maybe we haven’t done a monster in a long time. Or maybe it’s been forever since we last had an assassin. So we look at that and identify what we think would be most valuable.

RH: So to clarify, you’re not always just looking at what roles you need in the game? It’s not always looking for someone in a certain position or role, but sometimes it’s also their look or type of character?

DK: Absolutely. Ninety percent of the time it tends to be role and function, but very often we’ll see that we’ve just released, I dunno, four mythical creatures last year, so we try to make something more human.

RH: When you started that process with Taliyah, what was the role you were looking to fulfill?

DK: It was very clearly a mage, with a secondary goal of if possible not making it a creature or something mythical. We internally call 2015 the year of the myth, because everything was mythical in some kind of way. We wanted to move away from that a little bit and do something more grounded in reality. It was really mage. We had a big white board that had the word "mage" written on it, and then we started throwing stuff on there.

RH: Okay, so you know that you want to make a mage. What is your next step at that point? I’m sure you’re bouncing a lot of ideas around.

DK: Three people sit in a room. We call this – this is one of my favorite abbreviations – we call this the DNA of a champion. It’s the design, narrative and art behind the champion. We sit in a room and decide, okay, we want to make a mage. What possibilities do we have for mages? In the beginning, it’s very open brainstorming. Everybody throws out ideas. But we discovered very early on that we wanted to a.) make a traditional mage, and b.) there is a great opportunity space that is a little underused in traditional mages. That’s the elementalist slot. We have a number of good fire and ice mages. We have one semi-decent air mages and a couple of good okay-to-good lightning mages. We didn’t have an earth mage. We have Malphite, but he’s a shard of the earth itself, so he isn’t really a mage.

It was very clear to us in the beginning that we wanted to make an elemental mage. I pitched a lightning mage for a while, just because I had a kit idea with lightning conductors that you would place. But people just kept challenging me how that champion would be different from Kennen. I didn’t have a great answer to be honest. [laughs] They were probably right.

RH: At that point, when you have an idea that you’re really into but it isn’t the right one for this champion, is that something that you put in the back of your head to come back to later?

DK: Absolutely, particularly if you’re a little bit further along than just the initial brainstorming. You may have a piece of concept art already. If it just doesn’t work out for this slot, it gets put into our backlog of concept art. The next time we come around to make a new champion where this may be a good candidate, we throw all this old concept art together into a big overview piece that people can pick from.

RH: About how long does the DNA process you’re talking about go on? Is it a couple of days? Weeks? What’s the total length of initially coming up with the concept?

DK: We plan eight weeks for it, but it often takes longer than that. Because we have three groups that make champions, there’s also an availability of artists. If you finish your ideation super early, but the artists are still working on the last champion, then, well, may as well take a little bit longer. While I was ideating on Taliyah, the same group was still working on finishing Illaoi. That’s the last champion they had worked on. So we break down into different parts of ideation, but that goes down a rabbit hole. Basically, there are exit criteria at which point we know this is the champion we’re making. It’s generally two months until we hit that criteria, sometimes a little longer.

RH: What is your next step in the process then?

DK: Generally, I will have at this point already written a paper kit, which is just a Google Drive document where I write down what this champion could do. I ship that around a lot to other designers. We have a highly collaborative design discipline here at Riot. There’s a good dozen or designers that will beat up my paper kits and tell me why I’m stupid, and mostly they’re right. So I’ll try to do it again and do it better.

I demoed [my idea] to a Rioter, and he just walked off in disgust

Ideally – this is not always the case – I will also have an early prototype in the game at this phase already. Generally these very early prototypes are just aimed at proving out a very high-level pattern. In Taliyah’s case, at that point, I was trying to prove out a bunch of patterns. The only one that ended up sticking was the exhausted ground – or, we’re calling it worked ground now, because we wanted to move away from the purely negative connotations, but it was always called exhausted ground during development.

So that’s actually a story I like it a lot. This was an ability that I actually had zero faith in when I first put it in. It was meant to be a placeholder. My goal implementation was to have her rip up the walls around her and use that as ammunition for her spells, but I knew that would be pretty hard to implement. There were some really tough questions to answer around that – particularly, where does she find walls in her lane. So I was like, "Okay, cool, eventually this thing is going to use walls, but for now, just so we have something we can playtest, I’m just going to use ground." Everybody loved the abilities. "Oh, it was so cool, I was always looking for new ground to stand on." Hmmm, maybe that’s what I should be making instead.

RH: What are some of the other abilities that you played around with for Taliyah that didn’t end up making the cut?

DK: We had a very absurd one that I demoed to another Rioter when I was working on it, and he just walked off in disgust. [Laughs] What it would do is take one continuous piece of terrain on Summoner’s Rift and just destroy that piece of terrain for 15 seconds. The other team is doing Baron? Cool, destroy Baron wall for 15 seconds and attack them from behind.

RH: So you could actually destroy a piece of the wall?

DK: Correct! You could take out the base wall. You could take the jungle full of bottlenecks and turn it into a wide open battlefield to fight in. It was absurd. It was really cool. It was extremely hard to do well. We ended up actually deciding that we liked the ability, but we did not like it on a mage kit. It felt extremely support. It opened up the ground for your team to move around in. Taliyah actually wants people to funnel in through bottlenecks. If anything, she wants more bottlenecks. So a.) it didn’t work on her kit and b.) there were both art and tech questions that we didn’t get around to fully solving that would have jeopardized the release slot. We came together and decided that we’re not going to do this on this champion, but this may very well be something we revisit in the future, whether it’s on a champion or on an item or a summoner spell.

RH: It’s such a cool idea. I’m glad to hear that you might revisit it.

DK: I mean, any designer who takes on that as a project is a very brave designer. That stuff is hard to do right.

RH: How long was the process for designing Taliyah, then, from beginning to end?

DK: I looked this up recently. The first concept art that is recognizably Taliyah came in July 7 last year. Kit lock, where we have all abilities locked, was early February this year. That’s about the period. After that kit lock, we still did a couple of changes to the R skill. The fact that you can ride the ultimate came in very late. In the beginning, instead of riding it, you would stand at the origin point, but then you could input another command to curve it. That was pretty cool too, but riding is just so much more fun. Movement in our game is just baseline fun. We went with the riding there.

So, yeah, it takes about that long. Eight months, nine months, something like that.

RH: You had a hotfix for Taliyah that went out last week, and there’s been a lot of talk around the win rate for her being very low at launch. As her creator, do you think the low win rate is due to her being a little undertuned at the moment, or do you think it’s due to players just not knowing how to use her yet because she’s so new to the game?

DK: I think it’s both. We try to make predictions before we put champions live on where their win rate will fall based on our playtesting. The prediction for Taliyah was 45 percent on day one up to 50 percent within a week. We would have been happy with that. She ended up releasing at 37 percent on day one. She’s almost been out for a week, and she’s now at 40 or 41 percent. We have a very hard time predicting what’s going to happen once a champion is played millions of times a day all over the world.

It’s also very hard to judge how hard a champion is going to be. Like we knew Aurelion Sol would be super hard to play because the stars is such a new mechanic. You have to constantly look at where three points that are orbiting around you are relative to everything else in the game. We thought he would be much weaker than he released. In his case, it turned out, well, the enemy needs to dance around the stars as well, so that sort of evens itself out.

I think for Taliyah, particularly, she’s harder than she looks at first. Just keeping the stream of Q missiles on the enemy looks pretty easy but is actually a very mechanically demanding skill check that you’re asked to perform every five seconds or so. That’s a big one. A lot of Taliyah’s power is, I think correctly, tuned to be in her strategic, map-wide impact. She can get her sidelanes ahead better than most other midlaners, in the game. Maybe other than Twisted Fate. Because of that, she cannot be quite as powerful as somebody else would be in midlane who doesn’t have that strategic level impact. When you’re just trying to learn a champion and you’re struggling in midlane, the last thing on your mind is, "Well, I may not be doing well in midlane, but I should just roam bot and get us a double kill." Maybe you could, but you just don’t think about that.

RH: As a designer, when your champion comes out and isn’t performing well – if their win rate is low or even if their play rate goes down – is that something that you take personally?

DK: There are two answers here. There is the intellectual designer answer where I tell you that it’s perfectly fine for certain champions like Tahm Kench to be niche picks. We need to make visually unappealing champions, again like Tahm Kench, just for the game to have variety, and all of that is fine.

Then there’s the emotional answer: Of course, they’re all my kids! My hearts bleeds when they’re not being picked. Low win rates are sort of okay. Azir has been sitting at sub-50 percent for almost his entire existence as a champion in the game. But low pick rates do make me pretty sad. Certain champs like Azir that have this incredibly steep mastery curve – and by mastery curve we mean how much better you get with previous games played – they need to have, on average, 47 or 46 percent win rate. They’re actually balanced at 46 percent. There I can convince my heart that it’s fine for them to lose most of their games.

But when I see champions just not get picked, of course, it hurts a little bit. However, I should really not be complaining. With Lucian, I have probably the most played champion in the game right now.

RH: I haven’t played a game without a Lucian in quite some time.

DK: I’m gonna be the first to admit that he’s bullshit. He just – why do you pick Lucian? Well, because he does everything. Oh. Hm. Let’s see. Who else does that? No one.

So, yeah, it goes both ways. But I do want to go back and figure out how to make Tahm Kench more acceptable to the enemy team. He’s very frustrating when he’s on a team. That’s the last champion that I worked on before Taliyah, and seeing his abysmally low pick rate makes me a little sad. That’s a bigger project.

RH: That leads to another question I have. Once a champion that you’ve worked on is out, how much are you still involved with the day to day as far as patches and tweaks on the stats and any other changes that are going to come to that champion?

DK: The rule of thumb is that six weeks after release, ownership of the champion passes over to the live gameplay team. Those are the guys that work on the day-to-day balance of the game. Ideally, though, if you’re not an awful person, as a champion designer, live gameplay should want to come over to you and consult you as a subject matter expert. They’ll say, "Hey, we’re thinking of doing X. How do you feel about that?"

Very recently I worked with live gameplay – a couple of months now – on Tahm Kench changes. They wanted to make the ult easier for allies to opt in to but weren’t sure how to do it. I submitted something for them.

At some point, though, you just have to move on and trust that your colleagues have the best interest of the game and the players at heart, even if you disagree with something that they’re doing to a champion that you worked on.

RH: At this point, then, Taliyah’s brand new. So I assume you’re still working very much on tweaks with her?

DK: Absolutely, yeah.

RH: At one point will you be moving on to what’s next?

DK: Six weeks is a pretty decent rule of thumb. Normally, we say two patches after release she should have stabilized to the point that we don’t need to do anything patch to patch any more. I’ve already started writing paper kits for the next champion, and there’s a couple of little side projects that I’m starting to work on as well. It’s very fluid. It depends on what the champion needs, right? Like Azir needed extensive aftercare for months after release, and Tahm Kench was basically good after a couple of patches, and I could just move on.

RH: Obviously you guys haven’t even settled on what you’re working on next at this point, but can you give us any hints about what you’d like to do next?

DK: We are starting with fantasy slotting again. This time, we are definitely looking at a specific role. I won’t give away which role that is yet, but it’s something that we haven’t seen in a while. We want to try making a good version of it. But remember that there are three parts that work on champions, and each part works on two champions at the same time. They have one champion, generally, that’s in ideation and one champion that’s in production. My part was working on production for Illaoi, and then ideation on Taliyah started. When Taliyah went into production, ideation on the next champion that this group will make started. As soon as this next champion goes into production, which should be soonish, we can start working on what I will be doing next. So I expect that to be spring next year or so when that releases.

RH: You’ve said on Twitter that Taliyah was designed as a mid-lane champ. Now that she’s live, she’s being used in a lot of different positions – basically everywhere but ADC – and her highest winrate so far is at support.

DK: So that’s not entirely true. I’m going to pull up some data here myself.

RH: Sure, I guess we’re going off of the champion.gg data, but I’m sure you guys have more direct data.

DK: Yeah, let me see. The most recent breakdown we have is up to Tuesday, May 24. I’m seeing a 57.06 percent pickrate in mid lane, so by far the most often played in mid lane, and also the highest win rate in mid lane. The average win rate is 40.4 percent in mid and 39.5 percent in support, so close-ish but definitely highest in mid lane.

RH: Do you get frustrated at all when a character you designed to be played in one role is played in other roles, or is that just part of the game?

DH: I think as a designer, you should have a good idea where your champion will be played. You should have a good idea of the levers that allow you to push or pull a champion into other roles. It hasn’t happened to me, actually, that a champion is played in a role I didn’t want them to be played in. We started out making Tahm Kench as a top lane tank, but we pivoted in mid-development and decided he should actually be a support who sometimes goes top lane, and that’s exactly how it worked out. With Taliyah, we said, "This should be a mid laner who sometimes plays support." Mid and support are our two highest pick rates. Those have worked out.

I’m happy for players to experiment. I think that’s great, and I think it’s cool when the game throws curveballs at you. What I’m a little sad at is when players come to me and say, "Hey, I tried to play Taliyah in jungle. Didn’t work. Please buff." If you want to try that, that’s cool. Godspeed to you. But we’re not going to support that. We’re not going to specifically go in and make it so you can totally rock the jungle with her.

Here’s the thing that players don’t understand. This is really hard to explain. This is sort of like explaining to your kids why eating too much chocolate isn’t good for them. Options are great, but leaving all the options open for you is never how this works out. If I specifically buffed Taliyah for the jungle, let’s say – that’s kind of a ridiculous example, but let’s just run with it. If I specifically buffed Taliyah for the jungle, eventually consensus to the playerbase would swing over to saying, "Well, obviously jungle is her optimal role. Why are you playing her in mid lane, you troll?" It would actually cramp on her mid lane playrate, which would be bad, because we consider her most healthy there. So just giving options is not a thing you can do as a designer. There’s always a cost.

RH: You’ve been pretty open on Twitter about talking to people about Taliyah and asking what people want changed or how they’re feeling so far. What kind of general feedback have you been getting? Is there stuff that you’ve heard a lot?

DK: Yes, there is, particularly two things. People want the worked ground Qs to be stronger, and they want the W to either be changed to vector casting or they want me to allow Taliyah players to input other spells before the W erupts. I’m not going to do either of those things, but I started a thread on the forums where I’m inviting more feedback, and I’m hoping that somebody posts one of those two suggestions, because I actually have a large block of text prepared.

There’s a concept we call sharpness in design. The Q, when you cast it off of worked ground, feels really wimpy. It barely does anything. That’s absolutely necessary. If I allowed the Q of worked ground to still be a meaningful damage tool, then all of a sudden players would have to do the math in their head of "Oh, the Q costs less mana from worked ground; maybe I want to cast it from worked ground." At the same time, the enemy would be a lot less certain in their counterplay. Currently, they can say, "Well, Taliyah is standing on worked ground, so she is not threatening to me until she runs off of it." If that were less sharp, then that whole game would become much less interesting. It doesn’t really matter if she stands on work ground, she’s still a threat when she is on worked ground, so the enemy would just always have to be ready to dodge her. That’s one of those things where players want all of the options, but actually giving them the options would make the game less interesting.

The W is much more contentious, I would say. We have an input paradigm in the game that’s called vector casting. Viktor’s laser and Rumble’s ultimate use it. It’s basically where you move your mouse somewhere, you hold down a button on your keyboard, you move your mouse somewhere else, and then you let go of your button. It casts from point one to point two. This is a very, very hard input paradigm. I personally find it impossible to learn. I’ve tried to learn Viktor for the longest time. I just can’t do it. I have a lot of theories around why that is, but the TL;DR in my opinion is that it’s both a very novel input type and something where you are forced to make your input as quickly as possible, because the delay only starts after your last input. As a Viktor player, you’re asked to do something completely unlike anything else in the game, and you’re asked to do it as quickly as you can. Learning that ability is incredibly hard.

What I tried to do with Taliyah’s W is to make a version of that that is actually learnable. You can take as much time as you want to input the second key press and not actually lose out on any delays. The delay starts when you press the first key. Baseline, it’s already a new input that people have to learn. I think it’s much more learnable than vector casting.

The other thing I’m being asked to do now – some players get it, but they want to be able to cast immediately after the press the second W. That’s another one of those really subtle design things. I could give them the option to do that. Like, sure, if you’re fast enough, press W twice and after you press it a second time, you don’t have to wait for it to erupt, you just cast your other spells. The problem is, if I ever give you optional power, eventually, as the game state becomes solved, that power will become mandatory. It will be clear to the playerbase that this is the optimal way to play her, and if you don’t, then you’re missing out on power. By making a restriction that our high-end players find annoying, I’m hoping to keep the character more accessible to people who don’t have, like, Faker’s reflexes.

For more on Taliyah, check out our fanpost guide to the new champ.