StarCraft II: Building On The Beta

For developers creating multiplayer games, the beta is invaluable -- data on every aspect of how the game is played comes rushing in. But Blizzard's StarCraft II design director Dustin Browder warns that caution is required when sifting through the data for meaningful changes to make.

He also says that it's not Blizzard's place to create an innovative competitive PC RTS with StarCraft II, as it fills a very specific niche in the market -- shaped by both the series' broad popularity and its place at the forefront of the e-sports movement.

These and other topics are touched on in this interview, which encompasses the single and multiplayer modes, both of which Browder has responsibility for.

As design director, are you equally responsible for single-player and multiplayer design? They seem like very different games, more so than in the original StarCraft.

Dustin Browder: I am both single and multi. I have a lot of folks who help me out with that, though. I have a staff of very skilled designers in both areas. They do a lot of the heavy lifting.

Having been in the beta test for a couple months, I would imagine the amount of balance and design data you have to work with has skyrocketed.

DB: It's pretty high. We can learn a lot. The danger with a lot of this data is that you have to be very careful how you use it. With unit stats, I can tell you that, for example, in a Protoss versus Terran game, 12 percent of the time the Protoss build carriers. And when they build carriers, they win 70 percent of the time. You could say, "That must mean carriers are overpowered!"

That's not really true, though. It could just be that as you get towards the end of the game, if the Protoss have the extra resources to waste on a bunch of carriers, they're probably going to win anyway.

In other words, if your opponent hasn't managed to stop you being able to build 24 carriers...

DB: Right. Of course, it doesn't mean the carriers aren't overpowered either. That stat alone actually tells you nothing. It's a very dangerous stat. If you listen to that stat, you can make all kinds of mistakes. The real challenge for us is to continue to sort the wheat from the chaff, to determine which stats are real, which stats are meaningful, and which ones we should be looking at to make a meaningful change.

So you get that empirical data based on stat tracking, you get feedback from the beta forums, and then you get the high-level meta-game that exists largely in the minds and strategic evolution of experienced players.

DB: Yup. And our personal play experience, which is probably the most important. We take from whichever one is providing the correct information.

If we look at the stats and we say, "This doesn't actually back anything we're experiencing online," I'm very suspicious of that number. We get information from a lot of different sources, and then we use the other sources to refute or corroborate. We look at another source and say, "You know what? What they're saying online matches my play experience, and it matches the stats. This seems real. Let's talk about what some possible fixes can be."

On the other hand, "You know, what they're saying online does not match my play experience and it does not match the stats. Let's put this on a watch list, and if we see more information that can prove it to us, we'll make a fix. If we don't, maybe it will go away and was never real."

How much do you make a point to follow the meta-game that surrounds high-level players? Do you explicitly try to track that scene? It can be like going down the rabbit hole.

DB: We do track what's going on, but we don't make changes to try to push it in any direction. We're not trying to manipulate the meta-game so much as we're trying to make sure the overall experience is a positive one.

We saw, through StarCraft: Brood War especially, how many times the meta-game changed when we didn't even make any balance changes. One of the challenges for us is to know when not to touch it. I don't know how we're going to hit that.

That's the challenge for us down the road, and that's where we lean on guys like [Blizzard game design VP] Rob Pardo, guys who have been through this for many, many years. We can go to these guys and say, "Hey, you've done this before. Do you have any opinion of what seems real to you here?" We can use that vast well of experience Blizzard has at this point for this kind of work to help us out.

Looking historically, even after StarCraft came out, it wasn't really until Brood War that it really took off and something clicked.

DB: Yeah.

Blizzard had a while to ease into that. Now, you guys are expected to hit that right out of the gate. Do you think you're operating in a development design mentality as a result than that team was?

DB: Absolutely. We have so much more information available to us than they had then, both in terms of stat-tracking and in terms of contact with the community. And the community is smarter. These guys learned how to play this game much faster than they learned how to play StarCraft or Brood War. We have the advantage of their additional knowledge as well as the advantage of our additional knowledge.

And of course, we're shooting for e-sports, which didn't exist in the way it does now. We're targeting something that simply was never there. It's definitely a brave new world in that respect. There's all kinds of new information out there for us to get, and there's all kinds of new targets to hit that simply weren't there before.

On the notion of the e-sports phenomenon, there are some people for whom StarCraft is a fundamental competitive dynamic; it makes sense to them that StarCraft II still relies on that same dynamic, in the same way football hasn't changed its basic competitive form in many, many years. Then there are the people who are more dismissive, who say, "Well, it's the same thing I played ten years ago, but in 3D." How do you address that second group?

DB: For us, we're not trying to be innovative. We're not trying to change for change's sake. We're just trying to make a quality product. We definitely felt like there were some things from the previous game that were very high quality that we weren't super confident we could do a ton better. I wouldn't have a lot of enthusiasm to make Siege Tank 2.0. Siege Tank is good. It's a good unit.

It's much the same as the guys who are making Civilization or Team Fortress. They're making iterative changes to a quality product to do something really, really great.

But for the guys who are saying, "I just need new," we've created a whole solo play experience which we feel really scratches that itch, that is a brand new experience.

It gives them a new way to play RTS games that they haven't played in our games before. I know other games have done a similar experience, but we feel we've got a very high quality version of a non-linear RTS experience.

We really think that's an area where players who are just bored of RTS can have a lot of fun. But for the multiplayer experience, we're much more about quality than in showing off, if you know what I'm saying.

I'd heard you actually tried to implement a Relic-style cover system, like in Company of Heroes or Dawn of War.

DB: We did actually try a cover system. We tried it frequently, and what it did to our game was prevent a lot of movement from happening on the battlefield. It slowed the game down. Players would move to cover and stop. The other player is like, "Can I get around it? Screw it. I'll just fight." Or, "I'll just build up." The game stagnated.

Our game is about dancing: advance, retreat, advance, using the choke point -- until, "Oh no, the enemy went air, the choke is useless!" It's about give and take. For our game, it was a disaster.

It wasn't a perfect cover system. We never got that far, but the early indications were very poor. A lot of players view RTS as a continuum: RTS was this, and now changes have been made, and now RTS must start from there. We don't view it that way.

We think each game has its own style and flavor. Each game has its own strengths and weaknesses. What works for us would never work for a Dawn of War, and what works for Dawn of War would never work for us. They're different games, and that's the way it should be.

It seems encouraging that in 2010, after something of a lull in RTS, that we have such a broad spectrum now -- those games, as well as things like Total War and so on.

DB: Right. Totally different gameplay experiences, as it should be. [Those developers] are making a game that really speaks to their consumers and makes their people really happy, which is awesome.

How do you look at that market? Do you feel more confident when you see more RTS out there?

DB: It's always great to see games, as a gamer. There have been a decent number of Diablo clones, but we haven't seen a ton of them. And yet, we're making another one. So, if we were the only RTS out there, I'd be happy to make an RTS. We don't need thousands of people making this genre for it to feel like we should make a game. We can make a game.

You don't see a ton of clones of The Sims, but you still see The Sims turning out a product that's obviously very compelling to the users. But as a player, yeah. It is cool. It's fun to play these games, and I love a lot of them.

You were saying before the interview that you recently split the single- and multiplayer trunks, in terms of unit design and balance.

DB: Sure.

When did you do that, and how significant a change has it been?

DB: It was a couple [months] ago. We had made a change to multi, and it broke four missions, and I said, "We're splitting!" [laughs] It was like, "It's over! Oh my God! I can't believe we just did that!"

We tried to make it to the last possible moment. We probably actually went a couple moments past the last possible moment. We probably should have done it a couple days earlier.

We were pushing it, trying to get them as close as we could together, and we'll still roll [multiplayer] changes back into the campaign if we find that they're safe for the campaign.

Most of the fixes in that respect should hopefully be minor. But if we come up with some major changes to balance from multi, we're going to have some decisions to make, and it's going to be tough.

The danger there is that single-player is supposed to train you for multiplayer. But in practical terms, that never occurs.

DB: It never has.

Are you then trying to avoid creating a false impression for players that single-player will prepare them?

DB: Right. And yeah, we've had that feedback from people as well, especially Terran players who go online and are looking for their medics [which exist in StarCraft II's campaign but not its multiplayer mode]. For such a long time, they were a crutch in the solo campaign.

We had some balance problems. People were building nothing but marines and medics all the time. We've since addressed that issue. The medic has been a little bit nerfed for campaign, so you get more diversity in the campaign, which is almost as important to us as unit diversity in the multiplayer experience.

We want you trying different units, playing with different toys. People who would play nothing but campaign, and go to play online, would be like, "Where's my stuff?" We've been trying to work out language in the load screens and so on, to say, "Hey, you know, this is a little different."

It's definitely a challenge for us. We have ways to guide the players through the experience right from the launch screen. "Hey, you've just finished the campaign. Now you should play some challenges." You'll know that you need to learn some new things before you jump into the multiplayer experience. Then, "Hey, you've played some challenges? You should play versus AI. We've got some achievements. Come get the achievements. Come on, buddy. Come over here."

One of the problems that we have over a game like World of Warcraft is that WoW limits you by level cap. You're not going to blunder into Molten Core or Karazhan or whatever at level 15. You're just not. Those mechanics are way more complicated than you're ready for at level 15, so they control it through content. But we don't want to lock players out of multi just because they haven't finished the campaign. That would make a lot of people just very angry with us.

That means we have to try and control it through UI. We're going to use our achievement system as a carrot to pull players through the systems and hopefully get them to do them in an order that will make sense for them -- or something like an order that will make sense for them. It doesn't have to be exact. I just don't want them to play mission one, then jump into a one-on-one match, and then be angry, which could very easily happen.

I found it fascinating earlier when you mentioned unusual cases like the Zerg, for example, being very strong in Korea, but less strong in North America. How widespread is that kind of thing?

DB: That's the biggest example of that kind of thing. The other match-ups are more even across all the nationalities that we're tracking. So, how to deal with that? Right now, we're just not going to. It's too strange to make a fix based on that information.

What we need to do is go out there and use other tools now to discover what the differences are out there, and which one is real. Or are none of them real? And do we change nothing? That one is simply too strange to make a move on it. It's just terrifying, frankly. It's like, "Oh my God! What do we do? Better do nothing." Because any move you make could be horribly, horribly wrong.

So we're just going to watch and wait, let the communities exchange information, let the meta-game settle down in some places, and see what happens.

If we see things that are making the game more boring through less diversity, if we see rush problems, and if we see things we know are balance problems -- because we have played them and we cannot for the life of us figure out how to beat them, and we see other people not figuring out how to beat them -- then we'll deal with them. But that particular issue, we're going to have to let that one ride and see how it goes.

The challenge modes try explicitly to create situations that will mimic things people would need to know online. How did you approach designing those?

DB: Both as a tutorial, and as a game. We will say, "Look. You need to learn how to use our hotkeys. Here's out hotkey challenge. We've disabled the [on-screen] UI for this challenge. You need to learn some hotkeys if you're gonna beat this." Then you can beat it at higher and higher levels of difficulty. You can say, " I've learned enough. I'm moving on," or you can say, "I want to get the highest level because there's an achievement for that. I'm going to learn a lot of hotkeys in order to do this."

It's actually pretty fun mission. You just go nuts, kill as much stuff as you can, hitting hotkeys. It's just you destroying a whole map full of stuff with hotkeys as fast as you can, but it's challenging to keep up. There's a lot of adrenaline.

Some of the other ones, the "counter-challenges," are less of a game and more of a tutorial. We'll say, "You need to figure out which units need to go where, to counter these enemy forces."

Once you've figured it out, there's some micro involved in limiting casualties, but it's not as difficult or as rewarding to simply power through it, as it would be with a hotkey challenge. It's a mixture of the two. It's largely to learn the game, but it's got a component that allows you to push yourself and have fun while doing it, so it's not just learning.