Travis "Dovagedys" Boese, Senior Producer on Legends of Runeterra, went on Swim's Twitch stream for an interview to discuss balance philosophies, patch cadence, card development, and more. Below is our summary of that interview, lightly edited for formatting and clarity.
Philosophies Concerning Balance
On buffing and nerfing specific champions:
- Goal is to make sure the game is fun. Metrics are just a tool for that goal, not the end goal itself.
- Getting every card to exactly 50% winrate would not be fun.
- Some metrics are easy to quantify; others aren’t.
- Every champion should have a deck that it belongs in.
- Use decks’ aggregate winrate instead of a specific card's winrate.
- E.g., TF is in some decks with extremely high winrate (TF Fizz, TF Aphelios), but TF overall is actually below 50% winrate.
- Region play rate. Ionia is falling below their target play rate.
- Think in terms of decks and not champions.
- E.g., TF has been in a lot of top-tier decks (Go Hard, TF Swain), but that’s just because of his flexibility. Previous balance patches addressed the problematic cards from those decks (Make it Rain, Go Hard), not just TF as the common factor.
- “All work together to paint a picture”.
Timing
- No specific metric for when a thing has been too powerful for too long.
- Part of it is the question of how to handle cards that devs know are powerful but haven’t caught onto the community yet.
- For example, Draven was initially a 3 mana 4|3 Quick Attack. The devs knew he was the strongest champion, but he had a low play rate. They waited until it caught on before nerfing him to 3|3.
- Would like to see every champion in the game have "its time in the sun."
- E.g., Vladimir is a very competitive champion right now for the first time.
- He would like to see a natural “rotation” where champions rise and fall out of the meta, but it's unclear what the devs should do to make that happen.
Meta
- Their definition of the “perfect meta” is when the community says the meta is perfect.
- Meta diversity is valued.
- A goal is that 10 decks are competitive.
- HS and MtG usually have 2-5 decks.
- They don't know what's acceptable with regards to how long a deck should be allowed to be good.
- They also don't know how proactive the devs should be in forcing meta shakeups versus allowing players to discover their own solutions.
“Sharpness”
- Do think about the “sharpness” of the matches (i.e., how bad it feels to play against).
- They'll use surveys that ask questions like “Was it frustrating to play against that deck you just played against?”
- Pre-nerf Ezreal’s winrate was never too high, but he was changed just because it felt so bad to go 20 -> 0 in a single turn.
- Fiora wasn’t previously addressed as her play rate use to be low even though her winrate was high. This has recently changed, so now she’s on the radar.
Quote From Dovagedys We feel like we are making this game with the community.
TF Fizz
- TF Fizz has a high play rate (10-11%), but not as high as social media implies.
- Other decks have had higher play rates (18-19%).
- It’s an example of a skill-expressive deck, so its over-represented in the upper ladder and tournaments.
- Agrees that TF has been too strong for too long, although he wasn't as consistent prior to the addition of Stress Testing.
Balance Patches
On why we've only seen a single card change since Patch 2.1 (February 2nd):
- Shurima release didn’t have a balance due to many things:
- Didn’t want old content to be too weak to play.
- Ideally, both new content and old content would be equally viable on Day 1 of the new release.
- It also would have been contrary to the planned schedule.
- The standing plan is/was for balance patches to only come once a month excluding new content releases, but because they've made so many exceptions that's not clear.
Quote From Dovagedys The actual plan was for balance patches to happen once a month…This is what the normal schedule should look like [under that plan]:
- 2.3 [March 3rd] was Shurima.
- 2.5 [March 31st], which is coming up soon: balance patch.
- 2.7 [May 5th]: next Shurima release.
- 2.9 [June 2nd]: balance patch.
That has been our schedule for a while. Unfortunately, when we find a problem that can't wait, we make an exception which makes it hard for players to understand the schedule. So when players call out "hey, you use to do balance patches every 2 weeks!", in theory, they are wrong but practically they are not wrong because we've made exceptions. That makes it really hard to tell the difference between a [non-dedicated balance] patch and a balance patch.
Live Balance Patch Cadence
- Internally working on how balances and tournaments overlap.
- Very proud that from the 4 Monuments of Power seasonal tournament winners, 10 out of their 12 decks were unique.
- Very disappointed that from the Cosmic Creation seasonal tournament winners, almost everyone had some version of TF.
- Having uniformity like that will happen sometimes, but if that happens all the time he would consider that a failure.
- Agrees that the current plan for balance patch cadence isn't working.
- Conversations about revising the cadence plan have just started, so it might be some time before it’s put in place and shared.
Quote From Dovagedys The things that players have seen recently in the last few months are not part of some plan. They are the exceptions to the current plan because the current plan wasn’t working. When we come up with a new plan, which we haven’t yet, we will share that with everyone. Hopefully it will be a good plan that meets all the needs and then we will execute that plan accordingly and be more consistent so players have a better idea of ‘This is the plan and we’re following it’ rather than making these exceptions all the time.
The Logistics of Making a Patch
- Everything has to be completed 2 weeks prior to going live.
- Mobile submission & localization (translate into different languages) takes time.
- They don’t want to commit a balance change unless it is clear it will be necessary.
- I.e., they don't want to start a change and then the meta has already changed to make it irrelevant by the time it goes live 2 weeks later.
Balance Patch Goals
- Balance patches have specific goals -- e.g., making a specific champion stronger.
- Sometimes that means changing followers/support cards in that champion’s decks rather than the champion themselves. That doesn’t always come off in the patch notes though.
- They are scared of disrupting the meta too much.
- E.g., if they promote a particular deck and it becomes very popular, then the decks that naturally prey on that deck might start to dominate and its very hard to anticipate what those predators might be.
- They look for “inflection points” which usually means nerfs and not buffs.
- If you strategically nerf a specific dominate card, then multiple other cards which were being oppressed by those cards become stronger.
- If you buff a card, it usually only makes that single card stronger.
- They try to package changes so that when a region gets a card nerf, it also receives a buff on a different card from that same region to compensate.
Hotfixes
- Unscheduled / immediate live patches aka hotfixes are reserved for when things don’t work and/or are breaking the game.
- Examples include inconsistencies with Sejuani's level up and the Zephyr Sage infinite exploit.
- They are prepared to do an emergency balance hotfix if a card/deck was above a certain threshold, but no card has reached that emergency threshold yet.
The Watcher
"Why does Watcher cost 17 mana? And why is it an 11|17?"
- Dovagedys doesn’t know, but he says weird numbers are often associated with lore.
- There are 2 more releases of this set that could tell us more.
Card Releases
Mini-Expansion Approach
- They changed to the mini-expansion approach (new cards every 2 months) because one of the major feedback from players was that they didn't like the original plan of 4 months between new card releases.
- The major reason why the Targon and Shurima releases felt awkward is because they were converted to this approach after they were already designed.
- It takes a long time to release cards; they were working on Set 4 (Shurima / Empires of the Ascended) before the game went live.
- Sets 3 (Targon) and Set 4 (Shurima) were already designed when they made the decision to release mini-expansions, so they had divide up the cards as best they could.
- Content for Set 5 (August release) has been designed for this 3-expansion approach from its start, so it should feel more "whole".
The 10th Region
- Confirmed that there will only be 10 regions (i.e., only one more in addition to the current 9).
- Dovagedys won’t say what it is, but cautions against setting expectations.
- He’s not saying whether The Void is or isn’t the next region; he’s just saying if they chose that path it will be a difficult path and perhaps won’t be like what players expect.
- The Void only has 8 champions -- Vel'Koz, Cho'gath, Kassadin, Kha'zix, Kai'sa, Kog'maw, Malzahar, and Rek'sai -- which would leave it short-staffed relative to other regions.
- Riot has put hints out about what the 10th region is, and Dovagedys hasn’t seen anyone post or tweet about the biggest hint.
Quote From Dovagedys If we do the Void [as the final region], [fans] should set themselves for the expectations that there’s probably going to be a lot of [champions assigned to that region who aren’t Void in LoL] or we might not do [the Void as a region] at all because there aren’t enough.
Editor's note: Out of Cards made our own prediction about how the Void would be incorporated into LoR based on datamining already in February 2020. We're not saying Dovagedys' comments prove it, but it certainly makes it sound very convincing!
Power Creep v Complexity Creep
- Power creep is a loaded term. LoR intentionally has complexity creep but tries to avoid power creep (i.e., no "strictly better").
- A goal for the initial release was to be new-player-friendly in the initial release so that players – especially LoL players – could come to LoR without prior CCG experience.
- Complexity rewards more skillful play and players sometimes misinterpret that as power creep.
- We don’t want players to think new cards are strictly better than older cards, but we’re also still learning.
Rotation
Quote From Dovagedys We talk about rotation a lot internally. Just because other CCGs do rotation doesn’t mean our game will do it… Our goal is to make the best game we can make and give players the best satisfying experience they can have. If rotation ends up being the solution, then we will go down that road. But we don’t know if that’s the only solution. There are downsides to rotation that players that don’t talk about a lot. One of the biggest is if you love playing Yasuo or Garen or whoever from Set 1 and suddenly you can’t play that, that sucks. One of our goals is that players should be able to play the cards they own whenever they want to play them without being punished or feel stupid for playing them. It’s a tough problem we don’t necessary have an answer for yet.
Champion Design
Bringing LoL Champions to LoR
- Goal is to bring all LoL champions to LoR eventually.
- One of the challenges in bringing LoL champions into LoR is that they aren’t evenly distributed across regions.
- ~13 of the most popular 20 champions in LoL are Ionia, which is why Jhin, Master Yi, Ahri, Xayah, Akali, and Rengar aren’t in LoR yet.
- ~3 years ago, they actually considered making up a “Ionia civil war” to split Ionia into 2 different factions just to divide up the Ionian champions.
- “Whenever we make a champion in our game, we ask a bunch of LoL players what makes this champion special in LoL?”
- Proactively make the champion feel like the LoL champion, but that sometimes relies on flavor rather than mechanics.
- E.g., Zed strikes quickly and can kill quickly, but doesn’t have the mechanical complexity from LoL.
- They want players to feel like they can build an at least OK deck around any champion. I.e., even if its “just a worse version of Deck X”, that’s OK for less skilled players or fans of that particular champion so long as it doesn’t feel bad to play and isn’t a “carbon copy” of the other champion.
Backrow Champions
- “Backrow champions” (Aphelios, Twisted Fate) are champions who provide value without attacking.
- Riot wants LoR to be a combat-centric game, so champions that either discourage your opponent from attacking or overly reward passive behavior aren't good.
- Kindred is an example of a good back row champion: you can keep them in the back row, but it’s usually better to also attack plus they don't outright deny your opponent their attack.
- TF and Aphelios are examples of bad backrow champions that reward stalling and can deny your opponent's attacks (Gold Card, Gravitum, Severum, etc).
- Azir is a neat exception – as L1 he’s primarily a back row champion, but at L2 he’s often better attacking since he gains so much attack.
- TF and Aphelios give too much replacement value – even if they get removed, they usually have already generated enough value to create an advantage.
- They want future back row champions to attack on a different axis than just value generation.
- Riot will be releasing more back row champions in the future because players love them.
LeBlanc
- Loves LeBlanc.
- Important to introduce at least some “straight-forward” champions in every expansion.
- They want to give content for players who don’t like complexity and/or are new to the game.
- Strongly disagrees that they failed to capture LB’s flavor.
- Even if LB herself is straight forward, her follower cards (Mimic, Black Rose Spy, etc) are complex.
- Her deception is also reflected in her card art: in her L1 full art she's sitting at a table with others, and in her L2 full art they are revealed to be clones of her.
- Many players associate her with damage, and that’s what the devs focused on in LoR.
- In LoL, LB shows up out of nowhere, blinks on top of you, does a ton of damage usually killing you, and then blinks out.
Card Design Process
Gameplay Analysis Team
- In the beginning, Riot released new cards without consideration of the meta.
- Now, there’s a “gameplay analysis team” whose major responsibility is thinking about upcoming content in the context of the current game.
- E.g., they have been playtesting cards from the upcoming August expansion to see if Azir + [Upcoming Undisclosed Champion] to see if its too powerful and see if it needs to be changed before release.
- Concurrent Timelines + Commander Ledros combo is an example of something that slipped through.
- Each team that is working on new content now has a member from the Gameplay Analysis Team in all their meetings, etc.
Pre-release Versions of Cards
- In development, Twisted Fate’s Blue Card gave 2 mana instead of just 1.
- The Gameplay Analysis Team identified that this would be insanely broken.
- Funsmith was originally a 3 mana 3|3 with the passive damage ability.
- Trueshot Barrage was 5 mana.
- Legion Drummer was a 3 mana 3|3 that gave all units +1|+0 and overwhelm whenever you cast a spell.
- Aphelios’s Severum moon weapon permanently granted +2|+2 and lifesteal.
Developing New Cards
- It was hard to evaluate how powerful spells are due to the banked spell mana mechanic.
- Designers want champion cards to look exciting and potentially game-breaking.
- Also want to inspire lapsed players to come back with new cards.
Internal Tournaments
- Riot sometimes has internal tournaments with unreleased content to find broken cards.
- Originally everyone built their own deck, but recently the Gameplay Analysis Team provides the decks as they playtest and already know the "good stuff".
- Someone got to compete playing the version of TF with that 2-mana Blue Card and just dominated.
General
- Biggest lessons: content more frequently and gameplay analysis team involved in making content. Also trying to make game modes that aren’t just ladder – e.g., Lab of Legends.
- Riot is investing more into LoR's future, not less.
- LoR’s budget grew this year over last year.
- The team has gotten bigger.
- Less communications recently (no developer videos, fewer dev blogs) hasn’t been intentional.
- Would like to see more rioters be interviewed by streamers in the future.
Quote From Dovagedys My favorite thing about working on Legends of Runeterra is the positivity our community has. It’s nuts; I’ve never seen this kind of positivity. It’s really powerful, and it helps us when we make a mistake.
You can watch the full interview on Swim's Twitch VoD or in the embedded video below -- the call starts right at 1:01:30.
A big thanks to Dovagedys for spending his Saturday morning to be interviewed and to Swim for hosting it and doing such a great job. You can follow Swim online on his personal website or with the below social media links:
What did you think of Dovagedys' comments? Did they make you optimistic for the future of Legends of Runeterra? Tell us below!
Comments
"In LoL, LB shows up out of nowhere, blinks on top of you, does a ton of damage usually killing you, and then blinks out."
Based on that, something like Jarvan iv's summon from hand ability could have been interesting. And more useable with a 3-mana card.
This seems like a very bad definition of a 'perfect meta'.
It does seem like, with a wealth of information about what decks are seeing play, etc., they should have a better definition than just "the community says it's perfect." But there's something to be said for saying that the meta is only perfect if the fans are happy with it.
One of the things I think Hearthstone was very smart to do was to add new characters to the WoW lore when it helped create interesting legendary figures for the game. Riot being concerned about adding The Void as the 10th region due to a lack of champions suggests they aren't willing to see LoR and LoL diverge in terms of their "main characters."
That seems like a very bad idea long term. The "core fantasy" of any given champion may not be easily realized in CCG format (or vice verse) because of how fundamentally different MOBAs are from CCGs - they could just be setting themselves up for massive, and totally unnecessary design hurdles in the future.
I can’t find the quotes right now, but I’ve heard the designers say that they do plan to add their own champions eventually. I think they just want to start each region with established champions. That makes sense to me - Hearthstone wasn’t making up much lore until it was older than Runeterra is now.
I have been thinking the same thing, LOR having it's own exclusive champions could be a cool thing.
I’m very concerned about the how the game will work with (or without) rotations. I agree that it’s a tricky issue, and any option will have pros and cons. Specific regions and champions mean so much to players that it’s hard to take them out, and because the regions are not evenly distributed over expansions, a typical rotation would feel very unbalanced.
On the other hand, the number of powerful potential combinations keeps rising, and I don’t think it’s sustainable to keep growing the pool of cards without taking some out.
I’m open to creative solutions for this, but if the team doesn’t even have a plan yet then I’m worried that the game is heading toward a bad place in a year or two.
I wouldn't say that they don't have any plan; it's more like they have multiple plans that they are internally assessing and they don't want to say anything publicly until they know which one they're going to use.
For what its worth, there is a major CCG that doesn't use rotation at all: Yu-Gi-Oh TCG. They instead rely on a fairly extensive ban/restriction list to keep things in check. Fans of that game can quickly point out flaws with this implementation, but many of those issues arise from the fact it's a physical game and can't do live balance patches like digital CCGs can.
Thanks. I’ve been assuming they had some plan in place - when they announced the system of new “sets” every 6 months and “expansions” every 2 months, they even said they this terminology would be important for future changes in the game. That sounded like they had something concrete in mind at the time. I can’t tell where they actually are now in these plans.
Thank you for putting this article together. It’s great!
Saw the whole interview on stream, very insightful and I would recommend you guys to listen to it if you have the time.
They seem to imply the that the void won't necessary be the 10th region due to only having 8 champions but the other regions have even less, Ixtal only have 6 and the Bandle City only have 7 two of them are already in the game (Teemo and Lulu), maybe their planing to mix the latter two regions or maybe this just some misdirection.
Also Rengar is from Ixtal not Ionia.
Great catch! I actually re-listened just to confirm that I got it correct, and Dovegedys did say "Rengar" (around 2:10:00).
I'm actually terrible with lore, but I do know Rengar's lore has been retconned a lot. Currently his official home is Kiilash Village, which is technically Ixtal but its right at the border of Shurima and therefore commonly associated with it. On Riot's official LoL pages, Rengar actually currently appears under both "Ixtal Champions" and "Shurima Champions". I don't know much about earlier iterations of his lore, but since Rengar was released way back in 2012 while both regions were only conceived in recent years, it's possible that Dovagedys is thinking of some obsolete lore that associated him with Ionia. It's also possible this was a Freudian slip, spoiling that Rengar will be "relocated" to Ionia similarly to how Nocturne was relocated to Shadow Isles and that Ixtal won't be a region. Or its possible he just had a brain fart when he was trying to list Ionia champions on-the-spot. :-P
Like you said, I think that he just made a mistake when trying to remember Ionian champions from atop of his head.
As for Rengar his latest lore says that his original tribe hail from Shurima but that he currently lives in Ixtal were he is hunting Kha'Zix.
Thanks for the massive write-up. This must have taken you a VERY long time. We appreciate the effort.
As a response to one of the devs actual points, regarding TF/Fizz, I have to say that they are completely missing the mark with their evaluation of that deck, and of TF in general. "Only a 10-11% playrate" ... yeah, because people are fed up with it and many of them are refusing to play it, on principle, not because they don't think it's good. It has a super negative stigma, so a huge percentage of players are just flat refusing to touch it. The playrate metric doesn't account for this.
About TF specifically - they insist he wasn't a problem and was just "flexible" - so how do they explain the fact that probably about half of the nerfs in the game since Rising Tides have been targeted at TF's decks? If he has been OK all this time, how does he keep finding his way into the decks that end up needing to be nerfed?
*holds up his hands* "Not my fault - it was those other guys ..." *sneaks off to find another deck to force into being nerfed a couple months later*
Edit: And just to call out the dev on his stats, I just checked mobalytics - yes, some people are meme-ing with TF, which is pulling down his overall numbers. I saw one list that has Concurrent Timelines in a TF/Fizz deck - ok, yeah. Let's include those stats in an argument that TF isn't that bad. But anyways, back to the point: Mobalytics has data for 267,000+ games right now with TF. His overall winrate is 51.8%, which is not less than 50%, obviously.
Edit #2 - I overstated his playrate, because I did not read the fine print of playrate on mobalytics. He's 48% of all bilgewater decks, not of all decks. That's still noteworthy, but not insane.
I think you and Riot actually don't disagree as much as you think you do. Riot has already confirmed that TF is a problem and will be nerfed, so they understand that TF is the real issue. The only reason he's bringing up TF's playrate is to show that the issue is less with his playrate and more with how long he's been top tier (if this were the first time TF was viable, I think fewer people would be annoyed). Also, to be fair I don't think most people knew that TF was fundamentally broken before the TF/Fizz and TF/Aphelios meta, previously most people thought TF/Swain and TF/Go Hard might just be flukes (and lots of people resisted the idea of a TF nerf, such as Mogwai). Because of that, I think that it made sense for them to wait a bit longer to nerf TF, although I think it could have been nerfed at the launch of Shurima along with the Aphelios nerf.
But you know that Riot has access to much broader data than Mobalytics can hold? Mobalytics only holds data for people who use the deck tracker.
This may be true (though I've seen people saying they can get data for all the matches of any user who has signed up on mobalytics with their Riot ID), but as I mentioned in my comment, the TF data was for 267,000 games. That seems like a large enough sample size to make very meaningful conclusions.
But yes, I agree that Riot has access to much broader data than Mobalytics. For example, they have Expedition data, casual data, and even AI data. But think about this - according to the same guy who was talking in this interview, TF/Fizz is about 10-11 percent playrate with a really good winrate. But TF "overall" is less than 50% winrate. And he doesn't give a playrate for TF/Aphelios, but let's just be generous and call it an even 15% total playrate for TF/Fizz and TF/Aphelios combined, the only "strong" TF decks, according to the dev.
Now, in order for TF "overall" to be "less than 50% winrate," you would need another 15% of the playerbase to be playing abysmally bad TF decks in order to pull the winrate below 50%. That's at least a 15% playrate of really bad TF decks, not just mediocre ones. And you know there are people playing mediocre ones, but that just makes it even harder to pull the winrate below 50%, because even most mediocre TF decks are still above 50% winrate. So, you'd need the 15% to be losing BAD in order to pull the winrate down so far. (And incidentally, this little math scenario already pushes the overall playrate of TF to more than 30% of all decks in the game, just to make the dev's math work.)
How likely do you guys think it is that 15%+ of ranked players are just hard losing with TF for the fun of it?
Conclusion - that dev is doing some funky tricks with his data to make TF sound less problematic, for some bizarre reason. And the data from mobalytics backs up that conclusion - 51.8% winrate for TF across all decks that anyone on mobalytics is playing. And yes, mobalytics players definitely are playing some bad TF decks - just nowhere near enough of them to pull the winrate below 50%.
Edit - On a final note, Mobalytics does actually show Aphelios being just a tiny shade under 50% winrate currently. So maybe the dev got mixed up and accidentally quoted the overall winrate for Aphelios as being TF's overall winrate? This is definitely possible.