Hearthstone's Ben Brode: "If you knew how good a card is by looking at it, we wouldn't need a balance team"

No one loves Hearthstone more than Ben Brode. This is a commonly understood fact which is also possible to prove empirically: just hook him up to a microphone and count the decibels. The game director’s enthusiasm for what his team creates is unrivaled; when he laughs, the room shakes. Right now, he has good reason to be cheerful, with Hearthstone’s latest expansion, The Witchwood, due to drop tomorrow, bringing with it 135 cards—including some bonkers legendaries—and plenty of new mechanics and keywords. We’ll also see the spiritual successor to the hugely popular Dungeon Run mode from Kobolds & Catacombs in the form of Monster Hunt, which has new bespoke heroes and a spooktacular setting.

We spoke with the Brodester at PAX East last week, and as you’d expect had a lot of questions to ask about the new set. We were also keen to get his take on the recent ladder changes, which it's fair to say didn't go nearly as smoothly as they could have, and as a result caused no end of consternation amongst competitive players. Other topics covered include balance testing, the Deathstalker Rexxar update and, of course… finding the next Unicorn Priest.

PC Gamer: Lord Godfrey and Glinda Crowskin seem like powerful legendaries for Warlock. Given that the class is in such a good spot, what was the thinking behind giving that class such a power-boost?

Brode: Well, we're trying to give all classes a power boost here. The Standard rotation that's going to happen with this set will change the meta dramatically. So we want to give all the classes powerful cards to kick off the new year. I don't know that those cards are necessarily good in the decks that are already powerful for Warlock right now. We'll see. Generally, if things are unbalanced after the launch, and some classes are too powerful, then we don't mind stepping in and doing some nerfs. We've done it with almost all of our sets—we've come in a few weeks or months after the set launches and made some changes. So that's a thing we might have to do, but it's really too early to tell. Hopefully all the classes have powerful tools that make them exciting to play with.

But are you worried players will spend the first few weeks after the set launches wanting to experiment but instead finding themselves staring at a board of Voidlords?

I do think that's a concern. The Warlock decks that are good right now—specifically Cubelock—are very skill-testing. They are the kinds of decks that in our statistics, if you've played fewer than 50 games, your win-rate is pretty negative, but if you've played more your win-rate is pretty positive. It's not currently the best deck in the format, and when we nerf something you'd expect it to be the best deck. But we have [done that] in the past. One example is Quest Rogue, which was the 17th best deck and we think had a negative win-rate overall, but it was so emotionally negative to lose to that deck that we ended up nerfing it. There might be a similar case to that.

So I guess I'll say: yeah, we're worried about it, we would consider nerfing it, but it's too early to tell. The rotation's going to happen. A new set's coming out. That represents a significant amount of change. So it's too early to say whether or not it would need to be nerfed.

What can you say to reassure the competitive community that there won't be a repeat of the March problems with the way the Legend ladder works?

The biggest issue there was that this was the very first and only time that we transitioned from our old ladder model to a new model. And when we built this revamped system, we really built three different ranked systems. The first was when we launched the patch that had all this code in it but kept the old system until this reset. So that's Scenario A: new code live, old system in place. Scenario B was we're gonna roll from old system to new system. Scenario C is we're gonna roll from new system to new system.

So we've gone through phase A, which went fine. Phase B was somewhat of a disaster—we had to take ranked offline for a significant amount of time, and we issued packs to everybody as an apology for that. It did not work the way we wanted to make it work at all. And there were some big downsides to the way we had to repair data there, where some players ended up at rank one with five stars, and had to win one game to make it to legend. And there's a normal MMR process that happens on the way up to legend that kind of sets your MMR in the quote-unquote 'right' spot—that got all messed up. And then we had to do some backend things to make sure that some players didn't play literally no games of Hearthstone and clinch the number one [legend] spot. 

One of the biggest issues was communication. We did, I think, a terrible job of communication about some of these changes we made in backend. So we had both technical issues and communication issues. I don't expect to have either of those issues going forward. I think we've learned some good lessons about which things need to be communicated.

On the subject of good communication: for the sake of transparency, is there any reason not to completely lay out how MMR works?

MMR is very complicated. It's not a simple calculation. But it works the way people sometimes expect it to work, which is if you play someone who's a lot better than you and you win, you move up more significantly than if you play someone who's a lot worse than you in rank, because we just don't think that proves that you actually should be there. There's a minimum amount you can move up or down in MMR, but it can be bigger depending on what your matchup is like. It's a complex algorithm on the backend.

Do any of Witchwood's new mechanics have you especially anxious about whether or not they’re going to find their way into decks?

I'm really excited about the odd and even stuff, because it represents so many different unique new potential archetypes. So I'm sure there will be one good odd deck, at least. But which class will have the best odd deck? They are all going to play very differently, right? So I am very curious to see which class has the best odd decks or the best even decks—I'm curious to see how that plays out. I'm very excited about Shudderwock and all the combos you can get with that card, as well.

When you're testing, do you only have people sit down and make a ton of decks and try them out, or do you run simulated games with randomised decks?

We do a lot of automated testing. Usually it's like, play every card a hundred thousand times to make sure there's no issues. But we don't do a lot of machine learning balance testing. We have a team of very talented, super high level Hearthstone players to do that balance testing. And a lot of the time what they do is test the limits: they try and play the fastest rush deck, or the slowest, greediest control deck. And it's not just balance: it's also fun, which is a hard thing for a computer to output some kind of metric on.

Sometimes the word balanced is misused to just mean "good", when in fact unbalanced games can often be a lot more fun than balanced games.

Right! Quest rogue was an example where the deck was balanced—weak, even—and we nerfed it because it was not fun.

What kind of deck will you be crafting tomorrow?

There's a lot of cards I'm really chuffed about. I think I might go with Echo Rogue as one of my first decks. There are a lot of fun Echo cards in Rogue.

You infamously hinted that a secret 'unicorn' Priest deck might exist during the class's darkest days. Based on internal testing: which class is going to have the sleeper ‘unicorn’ deck surprise.

[Laughs] I wanna make this super clear because I am often HORRIBLY misquoted here on the unicorn Priest thing! What I said was: a lot of people make declarative statements about a class. "Priest is bad", "There are no good Priest decks." And what we've seen in the past is new decks pop up throughout the course of an expansion's life cycle. Like I think Water Paladin was one of the better decks in the Mean Streets of Gadgetzan era, and it did not appear until mid-cycle. There was a Pirate Elemental Warrior deck that was literally the best deck in the format for Un'Goro two weeks before the next set came out. So if that's the best deck and no one discovered it until two weeks before the next set, that means there's stuff that's out there that people are not seeing. 

That's just happened in every set: we see new decks crop up. I didn't know for sure that there was a really good priest deck, but I knew that a lot of people were really ignoring Priest, and we'd seen decks crop up in the past. All I was saying was that you can't definitively say [that all the good decks have been discovered] because it's very complicated to come up with a new deck to break the meta.

But one thing we've been trying to do is publish decks that have a positive win-rate, that are very good, but which are not being played very much. So we started this blog series that has high win-rate decks where players are getting 60, 70, 80 percent win-rates with off-meta stuff to show that there are really good decks out there, that no one's playing, that maybe you should keep your eye on.

So do you think there's any class that might surprise us in The Witchwood?

Almost always people are 100% off on card predictions, right? They see a card and they're like, "That's a bad card!" or "That's a great card!". Like, just historically, that's almost always been the opposite. There will always be surprises. If it was easy, if you could tell how good a card is by looking at it, then we wouldn't need a balance team. [Laughs] We would just look at the cards and work out if they're balanced or not. Part of the fun of the new expansion launch is finding out what cards are better than we thought they were and trying to break that meta.

What's the ETA on the Deathstalker Rexxarr fix? Will it be able to use Hunting Mastiff to create a Beast with Echo?

It will be when Witchwood launches that we'll be putting in the new beasts. I actually don't have the exact list of which ones we're gonna have but the plan is to put in as many beasts as we can from Kobolds and Catacombs and Witchwood. [Editor's note: Blizzard has clarified that Hunting Mastiff will be in the pool, and that old Beasts rotating out of Standard will be removed.]

Tim thinks he can guess your favourite new card: Blackhowl Gunspire, the new Warrior legendary which is also the first building in Hearthstone.

[Laughs] I actually helped come up with the concept for that card. I really liked the idea of a cannon tower. The design was that it can't attack, so I thought: maybe it's a building then! I like the effect of launching things, like maybe it's a cannon tower from Warcraft: 3, so I went and got some screenshots from Warcraft: 3. Then I thought we'd do the Hearthstone thing and turn this cannon tower into an insane-o cannon tower, so it's got the MAXIMUM number of cannons in it. I was really excited about the flavor and art for that card.

But I think my favourite card is actually Countess Ashmore. It's very similar in feel to The Curator. Her is "Draw a Lifesteal, Rush and I Deathrattle. So now you have to think, "Ooh, I wasn't gonna put a Rush minion in my deck but if I do then I have enough of these guys". It really changes the way you build your decks and think about which keywords you want to include.

We're excited that Toki, Time-Tinker is the first Standard card that can draw cards from the Wild pool, but she’s the only card in this set that can interact with Wild cards right?

Yes. We really wanted Toki to feel special and highlight her time travel powers—she's a mage who is an incredible engineer-slash-tinkerer, and she makes all these gadgets that help her travel through time. She's one of the heroes in the Monster Hunt, and her power is to restart the turn. So if you have random effects you don't like, like you cast Arcane Missiles and it doesn't go where you want it to go, you can restart and try it again. She's a really fun hero to play, she's super funny.

And each of the heroes has a nemesis, so the eighth boss in each run is specifically the nemesis of the hero you're playing. And Toki's nemesis is Infinite Toki From The Future, and she's come back in time to stop herself from messing with the timeline too much because she's aware of events in the future that are not good. So Toki fights herself. We love the character, we love her storyline, and really wanted her to be a legendary card in the set and have a unique effect that really calls to that time travelling thing, so that's why she goes back to the past and gets cards from Wild.

Check our list of the best 25 legendary cards in Standard, as voted for by Hearthstone pros. We'll be updating it based on a new poll once the meta settles.