An Extensive Interview With StarCraft II's Design Director

Dustin Browder, the design director for Blizzard Entertainment's upcoming mega-RTS StarCraft II: Wings of LIberty, took time out of his busy Zerg-nerfing schedule to speak with us about the philosophy behind the game. Strap yourselves in, because we're in for several pages of exclusive info on one of the biggest PC games of the year.

What's the biggest change you've seen in the way the StarCraft II beta plays compared to high-level StarCraft play?

We're seeing all kinds of crazy strats. We don't know if they're real, if the counters have been found out yet, or if it'll all shake out in the end, or if we need to make fixes. Obviously, StarCraft II is a more mobile game in many ways than the original StarCraft was. This is in some ways a function of the types of units we have. There's a lot of mobility in this game in terms of moving and dancing, attacking and retreating. We're seeing lots and lots of dancing between players as they vie for position, as they try to get in good flank attacks, and as they try to get in behind their enemies and prevent them from escaping. That's probably the single biggest global change that you see, but as I said we're also seeing lots of crazy rushes that we haven't really seen before. We're seeing lots of warping in stuff early on in enemy bases. We're seeing lots of early attacks that weren't really possible in the original game and we're trying to decide, "Are these permanent problems? Are they problems that we need to deal with right now, or will they shake out in the end?

Is that change in overall mobility something you intentionally built into the game?

No. Our creative process on StarCraft II was very different than the creative process I've used on other RTS games I've worked on or even on WarCraft III. We didn't set out with any goals in mind, and I'm sure that this will upset the fans terribly. What we did instead was that we said, "We want to make a bunch of cool units, and we're going to make each unit as cool as we can possibly make it, and then we'll see how it all works together, and we'll tune as necessary from there." So it was never our intention specifically to do anything exactly with the races. Our goal was to make the units as interesting as possible and as different from one another as we possibly could. From there we could see how the races changed and evolved. That's definitely how it has worked out.

The Zerg, I think, are a lot more methodical in StarCraft II than they were in the original StarCraft. The Terrans have a lot more of a commando feel to them in a lot of cases than the original Terrans ever did. I know that's something that a lot of the fans look at and say, "Oh my god, you've changed the Zerg!" but there's really no way to re-make this game and to change the units without fundamentally changing the character of the races. So our goal initially was, "Each unit is as different as it can be from all the other units in the game, and each unit has its own unique role and its own unique abilities, and each unit hopefully feels as cool as it possibly can feel, and then we'll see where the races end up and hopefully it won't feel too different from where they should be conceptually." But we didn't really have a set goal when we started, and right now I'm very happy with how everybody is playing out. Everybody feels very different from one another, the kinds of strats you see in the different race matchups are fairly different. We're obviously looking to continue to push those even further. For example, we're seeing the Roach-Hydralisk combo for Zerg is pretty core to a lot of the matchups, and we're going to do some things to try to break that up a little bit. But against Terrans we're seeing very different strategies as a Protoss player as opposed to playing against a Zerg player.

Were you sad to see anything have to go in the name of balance?

Early on we obviously weren't as worried about balance as we are at this point in time, but for instance the Thor was a unit that went through a lot of changes. We had these big cannons on the back of the Thor that did massive, long-range AOE [area of effect] attacks because that looked cool and it sure was fun to play with. It stepped on the Siege Tank in the worst way. You really didn't need a Siege Tank if you had a Thor, or if you had a Siege Tank you didn't need a Thor; one or the other. So we ended up scaling those back to be a single-target, high-damage stun attack, which feels pretty different from the Siege Tank, but it doesn't quite match the visuals as much as I would like. But it's a much better player-based experience.

I would say that those are some of our biggest struggles with StarCraft, that the art will imply a certain look and feel, like what's supposed to happen, and it doesn't always match the best gameplay goals. We'll often go with the gameplay, which means that the art is sometimes a little bit out of sorts with what you imagine it could do. For example, if I build six Battlecruisers, a lot of players look at that and imagine, "there's nothing in the world that can defeat this. Look how big they are! I will crush all with this! No-one can stand before my might!" And they're quite surprised when they run into 30 or 40 Hydralisks and they get rolled, and they think, "What?! That can't be, that's not right!" But we're not willing to make the Battlecruiser look wimpy just because if he encounters ten times his cost, he's going to lose. Those are some of the challenges we face, and that's the area you'll see we've done something and we have to go back on it a little bit, just because we thought it would look cool and we thought it would play cool. If we thought it didn't play cool, but we still think it looks great, we'll keep some of the visuals but we'll change up the gameplay to make it a better experience.

How do you approach defined counter triangles and other meta design aspects?

We do a lot of work on that stuff. StarCraft is a little different from other RTS games in the sense that we do have some triangles in the game, but we also try to make sure that a lot of the relationships -- between a lot of the early units, especially -- can be changed fundamentally with micro. For example, Zealots counter Stalkers unless the Stalkers kite, and if the Stalkers kite the Zealots will lose. But it all depends on how you manage your units. It's not enough to just build the right units; it also matters how you use that unit. If I bring the wrong units to the party, like if I come storming out into the middle of the map, like "Hah, I've got a thousand Stalkers! Oh my god, that's a lot of Zealots!" you're not just dead. You can, in fact, get some use out of those units.

Terrain is a large factor as well. Marines will absolutely lose to Zealots, unless they get some kind of terrain advantage. If they can catch the Zealots in a tight choke point, or even in some cases just put their backs to a wall and prevent the Zealots from surrounding, they can get a lot of damage off. So for us, it's all about some of those triangles, but it's also about making sure there's a lot of gameplay based on your position and a lot of gameplay based on the terrain around you. Then we just grind away at that stuff for months and months and months trying to make them really interesting and giving a lot of player choice and providing a lot of fun tactical options for players not only in terms of the forces they build, but also in the types of choices they make on the battlefield.

Are you happy with the balance between micro and macro play in the beta?

We see ourselves as doing a lot of everything. Some games favor a macro gameplay, and some games favor a micro gameplay. Our goals are to have both be possible. As a Zerg player, you can choose to be more of a macro Zerg player or more of a micro Zerg player, or anything in between. The reason this is so interesting for us is if we have that kind of choice for players, it means that even in a Zerg versus Zerg game, I could fundamentally be playing a very different race than you're playing because I'm playing a much more economy-based game, and you're playing a much more tactical-based game. Both strategies are viable and possible -- that's our goal. Whether we succeed or not is still up in the air and ultimately up to the fans.

In terms of how happy I am right now, we're in beta. It's certainly the biggest game I've ever had the honor of getting a chance to work on. I'm not happy at all. We're all just scared s**tless. At this point, I just want to make sure we get the best possible product to the fans and we're really waiting on the fans to tell us whether or not we should be happy. I'm hopeful; I'm not seeing huge problems. But that doesn't mean we won't see huge problems any minute now. I could walk back to my desk right after this interview and watch a replay and go, "Oh my god, what are we going to do about that?" [laughs] So we're all on pins and needles over here waiting and seeing how the beta shakes out. Every day brings new sources of terror as we wait to see what the fans think up to do with the game that simply has not occurred to us from playing internally.

How do you plan to introduce newer players to hardcore competitive play?

I'll say two things about that. One is that I would definitely encourage the players out there who got lucky enough to get a [beta] key who are thinking of it as a demo not to think of it as a demo. What we've done with the beta is fundamentally very similar to putting out a beta of level 80 World of Warcraft arena with no buildup before that. That would be the most difficult thing to learn. This is really a chance for us to test our play balance with players of all skill levels, but players who are already committed to learning and playing the multiplayer experience and don't feel too bad if they lose a game. For the final game, we've got twenty-plus hours of campaign, which will get players familiar with and comfortable with the controls, familiar with and comfortable with resourcing, give them a basic understanding of some of the counters in the game, but just give them 20 hours of sort of comfortableness with the game, much as you would when you're playing through 20 hours of questing in WoW. It doesn't teach you how to do high-end raiding, but it gets you comfortable with the UI and it gets you comfortable with your character. So we've got that.

Then we've got our challenges. Our challenges are about ten maps that teach you the fundamentals of online competitive play. They teach you how to defend against rushes. They teach you how to build up your economy. They teach you some of the basic counters for the three races. They teach you how to block your choke. They teach you all kinds of stuff that you will need to understand on some basic level to play in the competitive environment. Then we have five different difficulties of AI that you can play against to learn your race. Obviously the race in solo play for Terrans is a little bit different from the race in multiplayer. If you want to learn Zerg or Protoss, you can go to those AIs and battle it out against them. We've also got an auto-matchmaker for cooperative against the AI that will hopefully get you to a position where you're more comfortable playing online and more comfortable playing in a team environment.

We're definitely going to be looking at Achievements and other tools to encourage our players, once they graduate from cooperative vs. AI to move into team play before they move into 1v1. We're going to offer 4v4 and 3v3 to encourage them to get into a team play environment, get with a team. It's a much more casual environment than the 1v1; you've got somebody to help you play with, you've got allies who can come to your rescue if you get into a little bit of trouble. We really consider the 1v1 the final endgame arena-style gameplay that we hope players will only get to after 30, 40, 50, 60 hours played depending on their enthusiasm for it. That's my hope as a designer, but whether any of that works out is anyone's guess. We're going to try to do anything we can with the UI and with the Achievements to encourage players to follow some kind of arc like that so that by the time they get to 1v1 hopefully they're feeling very comfortable with the game and they're ready for it -- they're not just thrown into the deep end.

Will the Battle.net revamp support co-op or other online gameplay in any meaningful way?

We're going to offer a bunch of Achievements associated with [co-op play] to encourage you to keep playing and getting rewards. We'll have avatars tied to those Achievements so you can show off your ability against the AI specifically, both solo and in a team environment. We've also got the Battle.net party system, which will allow you to get together with your friends very easily and take on AIs in a variety of different formats. That was one of the ways that I always enjoyed -- besides competitive play -- with my friends, playing the original StarCraft that way. We really hope that that is a big part of the experience here for a lot of players.

Ultimately, I imagine that players will start off by playing cooperative against the AI and earning achievements through the matchmaker, but that they'll graduate at some point, having made friends online or having gotten their friends into the game, into playing custom games and trying on 3v4 versus the AI, or 3v5 versus the AI. "What happens if we do it on this kind of map? What happens if we do it on that kind of map? What happens if we're all Zerg?" Those are the kinds of ways that I think are really fun to play StarCraft, where you are continually challenging yourself, not only with new difficulties but with new numbers of opponents on map types that are maybe less friendly to your strategies. Challenging yourself to try races you're less familiar with. It can be a really fun way to play.

Will there be many of these unusual maps at launch?

We're definitely going to have a variety of map types out of the box, but there's really nothing that we can do that will equal the scale or scope of the community's endeavors. We've got hundreds of thousands of map-makers out there making these things. You're going to see a lot more exciting stuff. And of course you saw some really great stuff in Warcraft III. Tower defense maps were created by the community and went on to spawn almost a whole genre of game types that now you can play on Xbox Live or download on Steam. So our community almost invented a genre. You've of course got DotA [Defense of the Ancients] as well, which was a map type that was developed in StarCraft and Warcraft III which went on to be a really great way to play games online. So we're obviously going to continue to pursue that. We're going to try to provide the community with some seed maps when we first launch that will get them started, but we fully expect the mapmaking community to take off running and show us how it's done and make some really great stuff. We'll see what happens.

We're going to try to support some cooperative-style maps taking some of our campaign game mechanics and making cooperative maps out of those mechanics. We have maps where lava rises every few minutes and then recedes, and you can only be on the low ground during the time where it's safe, and if you're on the low ground when the lava rises everybody dies. That could be a really fun cooperative map that's pretty easy for us to put together. We're going to try to ship with some maps like those, but just like we always have in the past – like if you've ever explored some of the custom campaign folders in StarCraft and Warcraft III, before the community even gets their hands on it there are all kinds of crazy map types in there. We're definitely going to try to support some of that at launch, but I fully expect the community to take off running and do a ton of stuff in even the first few weeks and months when StarCraft II goes live.

Is the custom game browser going to be any better?

Our custom game browser is going to be a little bit updated. Instead of just a giant list of maps and good luck finding the one you want to find, the maps are now organized by type. You'll be able to go in there, and you won't see 30 games when in fact there are several hundred games and we just gave you the first 30; you won't see that. You'll see 30 map types that are listed. So you'll be able to say, "I want to play Defense of the Ancients," and if Defense of the Ancients is one of the most popular maps out there -- which, who are we kidding, it probably will be -- then that will be one of the top listed game types and you can just click on that and it will put you into a game of that type. If by weird chance there are no other games out there right now, it'll ask you if you want to start a lobby.

Did you have any radical thoughts early on in StarCraft II's development?

We went through all kinds of crazy stuff. We'd been working on some of our campaign stuff where you've got a much stronger role-playing component, and that has changed shape several times throughout the development process. It was much more of a role-playing game at one point, and now it's settled into a proper RTS with some really strong progression elements to it.

How is StarCraft II a better game now than it was when it was first shown in playable form years ago?

For instance, a year ago we didn't really have a Zerg race that was that much fun to play. At the time, the Roach's regeneration was above ground and all the time. While I appreciated the unit for its simplicity in that role -- a constantly regenerating unit, kind of a Zerg Wolverine character, that was pretty fun -- it ultimately produced some very random results in battles. If the units randomly focus-fired on your Roaches in the correct way, and just by chance they happened to pick off one Roach at a time, then you would suddenly have a big victory against the Roach as opposed to if you let them attack move and they all picked their own targets and you got a little unlucky with the acquisition, then suddenly the Roaches would roll your units. We didn't have a working spellcasting kit for the Zerg at the time. We didn't have the Queen as the proper base defense spellcaster; she was trying to be an offensive caster, and she was trying to be a defensive caster at the same time. She was a little conflicted as to what she was supposed to accomplish. As I said, the Thor was sort of wandering around, trying to figure out what his role on the battlefield was. So a lot of the units were a lot more muddy than they are now; they were a lot less crisp.

The race matchups were also a lot worse. It was very easy to have games where it was just…depending on the week, it could be that Protoss was winning every game versus Protoss losing every game. We had a lot more units in the game. We sometimes had fifteen or sixteen units on the Zerg or Protoss at any given moment before we finally got down to 14 now. Sometimes as many as 17 or 18 on the Zerg, which meant that the race matchups got very confusing. It was difficult to know what to build at any given moment, and there was lots of overlap either between the units on a single race or between the units on a race and another unit. You'd play as the Zerg with something called a Spore Beast and we'd say, "Oh my god, this is just a Banshee, isn't it?" We'd try to tune as like a really fast Banshee, and it was like, "Okay, dude, but that's still a Banshee." It hadn't fundamentally changed its role. We had some stuff that was a lot wackier, that felt more new but wasn't necessarily more fun.

So I would say it was just a much less polished experience. Just not as tight. It was sort of all over the place. You would have some things that were fundamentally crazier and they seemed more fun until you actually played with them. We had a Corruptor at the time that every time it got a kill, it would turn that kill into a flying corrupted thing that would shoot at enemy forces. So you'd see a Viking that was all covered in slime and it'd be shooting at other Vikings and all these spores would be coming off of it. And it was really crazy cool! It felt so Zerg-y, it was so awesome! I definitely miss that vibe from that unit, but I gotta tell you, it was just chaos. You really had no idea of what was happening on the battlefield, because the fliers are allowed to overlap in StarCraft. So I'd have six or seven Vikings kind of on top of each other, and maybe three or four of these corrupted Vikings also on top of the Vikings, and I'd have no idea of what was going on. My Vikings are exploding, but who could say why? Not even the Zerg player knew what was happening. Lots of stuff like that. A lot of clarity issues, a lot of overlap between the units, a lot of playbalance issues.

All of the time, we were just getting it tighter and tighter and tighter, and now we've managed to get it in front of the community, which is really exciting. Now we're starting to see where we can continue to improve it. I've complained about it before, but I thought maybe we were good enough and now I think we're not. In Protoss versus Protoss games, if you get a bunch of Zealots together from two teams with kind of similar colors, it can be really tough to see what's going on. So there are clarity issues, but also the fundamental crispness of the sides and crispness of the units so that when you see an enemy unit, you know immediately, "I know exactly what I have to do to stop that, I have got the perfect unit that's going to roll him. If I can only get it out in enough numbers, I can totally push him back." That kind of precision and crispness really makes the game feel great.

Have any glaring imbalances shown up in the beta so far?

I haven't seen anything in beta yet that's really…. Well, that's not true. We have a proxy Gateway rush that's been causing us a lot of trouble. It's still evolving even now. The basic strat is that you build a Pylon on low ground, outside the guy's base. Then you run inside his base and you build a Gateway next to the cliff edge so it's powered by that Pylon. So the enemy player is unable to get at the Pylon, which is the weak link in that whole setup, because the Pylon is on the low ground. Then you start building Zealots, or at some point warping in Zealots in that power field. This endless Zealot pressing, especially against a Terran player or even a Protoss player, is pretty unstoppable. [laughs] It's pretty grim. That one is one that we've been attacking for quite a while and we're about to attack it again. We're seeing it in Protoss versus Protoss games now. We've beaten it back pretty well in Protoss versus Terran, but Protoss versus Protoss is still really bad. If one Protoss player goes three Gateways, the other Protoss player needs to go three Gateways to counter or he is hosed. So we're probably going to do some changes to the build time on the upgrade for Warp Gates, just to push that a little bit later in the tech tree so that a lot of the early Warp Gate shenanigans that we're seeing in the beta get pushed back a little bit.

It's really tough, because the initial urge from a lot of guys when they see that, is "We just need to take that and move it to Fleet Beacon." They just want to bury it in the tech tree as much as they possibly can. I understand the urge, but at the same time we've got to try to maintain the differentiation between these races, and we've got to try to push everything to feel as powerful as it possibly can without actually being unbalanced. We definitely want to make these things feel mighty. We also want some of these rushes to be possible. A lot of players will look at that and say, "Well, that's just broken. Why don't you just beat that out of the game so it's not even possible?" But for us at Blizzard, we really embrace the rush. For us, the ability to win or lose this game at any time is one of the things that makes it so exciting, that makes it such an intense emotional experience.

There are definitely players out there who don't want that intensity in their lives. I understand. They're like, "I don't want to lose a game, thank you very much, let alone in the first 20 seconds." But we feel like that's part of what makes the game really exciting, and really intense, and such an adrenaline rush, such an intense emotional experience. Because when you see that load screen end, you know it's on. Every move you make for the next fifteen minutes could determine your victory or defeat. We definitely embrace the rush, we just want to know that some of these strategies are as easy to block as they are to do. When we see a strategy that's really easy to do but very, very difficult to block, that's when we have a problem. In the beta right now, we're seeing a lot of players do a Planetary Fortress rush where they do everything possible to get the technology necessary for a planetary fortress as early as they can, and they fly to the enemy base, and they land next to their peon stream and they convert their Command Center into a Planetary Fortress, and it's like, "Haha!" and the guy's dead. All you have to do to stop this, of course, is to put just one unit underneath that Command Center and it can't even land. So it's really very easy to block, but if you're not paying attention you can get caught by it and defeated. So that's an example of a kind of rush that we want to allow. It's okay if players do this kind of thing as long as it's really easy to block.