Before he joined Blizzard and became the Lead Multiplayer Designer for StarCraft II, David Kim would take part in StarCraft and Warcraft III tournaments. Although quick to downplay his skill compared to other professional players of the day, it’s hard to look past the fact that he spent years dedicated to the competitive StarCraft circuit -- honing his skills as a Protoss player. And then when Warcraft III was released, the Alliance. David was in high school when the original StarCraft came out, and after each day would join other students in the computer rooms to play multiplayer. He took to the game almost instantly, and it wasn’t long before he ran school-wide StarCraft tournaments, and in the process become known as “the StarCraft guy.”


Travelling between Korea and Canada, a lot of David’s college years were spent playing Warcraft III and StarCraft, taking part in various tournaments. It was at this time that Relic Entertainment scouted and recruited David to join the team that would go on to work on classic RTS titles like Company of Heroes and Dawn of War. But, once Blizzard announced the existence of StarCraft II at the Blizzard Worldwide Invitational in Seoul, he immediately applied to join the team working on the sequel to one of the most popular and enduring strategy games of all time. And truly, live out a dream.

At BlizzCon I had the chance to sit down with David to chat about his background and all things StarCraft II multiplayer.

StarCraft II is essentially a trilogy of games that began with 2011’s Wings of Liberty, with each entry not only offering up a new cinematic story to experience but also big changes to the multiplayer. How did the team approach changes from title to title?

The trilogy side of the StarCraft II games is mostly related to the campaign. From the multiplayer side, you can probably see them as one game with each release representing a new expansion. And now, with the major patch work that we’re doing, we see that as another iteration or expansion to the StarCraft II experience. Our mindset when working on the multiplayer was less like we were creating a full game each time, but more along the lines of finding the problems that exist, fix those, and then look at cool new strategies we could evolve from what was already there. Which is basically an expansion point of view.

That makes sense. But even as expansions you guys had all this additional time between releases, how did that effect the development process?

With that extra time, we tried our best to think outside the box and experiment a lot. Looking back, some of the ideas we had sound kind of stupid now and are a bit out there in how crazy we were thinking. But, in order to see the big picture and to realise and know exactly why something isn’t working you do need a lot of time to experiment. For example, we had this Terran unit where you had a launch pad that you had to build in your base to use it. And so, you would send out this bomber and pick an area that you were going to bombard. The bomber would then fly out there, bomb the location, and fly back. But you couldn’t control the unit during that time, it was all automated. And the counter was if the enemy killed it then you’d have to re-build it at the launch pad.

But, even though we got a version of it into the game it didn’t really work for StarCraft. Because in StarCraft the cool thing is that with every unit there’s a potential for me to do something with it in combat. So, this had none of that. And someone might say, well yeah, that’s obvious. Why did you even try that? But we wanted to try as many different things as possible, so we’re more fully aware of the wide spectrum of possibilities.

Was there anything like that on the multiplayer side that informed the campaign side? Or vice-versa.

Generally speaking, it was one way. We would try something on the multiplayer side, and because we go in stages -- in speaking of the bomber unit, that got cut before it went to the art pass or the effects pass or the audio pass. But, there were some times when there were some units that would receive the full treatment from art to audio that weren’t the rift fit. So instead of just throwing those away we would see if they could be used in the campaign. And then later we started to joke a little bit more about that whole process -- if the unit is dead, let’s give it to the campaign team. Now, that was more of a troll because if the unit didn’t make sense for the campaign then they’d throw it out too.


This happened more with Wings of Liberty than the other two games though, because after that we learned our lesson. We realised that we should leave stuff like art and audio and effects to the very end, once we were sure that these units would actually be used. One of the units that did make the jump was a fast-moving factory unit that fired on the move, the Diamondback.

One of the great things about Legacy of the Void was the new co-op mode which kind of bridged the gap between the campaign and multiplayer side. Was there a sense that StarCraft II was two games?

Certainly. Up until Legacy of the Void we had two modes, Campaign and Multiplayer. And if you were to look at the downside of each, where with the Campaign mode once you’re finished you’re done, that always felt wrong. We put in so much work into it, it’s an awesome experience, so it’s a shame that it doesn’t have the same sort of replay value that the Multiplayer side has. And the downside to the multiplayer, which is kind of still there, is that the barrier of entry to new players is quite high. And the reason for that is we wanted to make the most competitive game in the world. Something that if you spent two years mastering, you would really feel that you were substantially better. So we wanted to take the good parts of both the campaign and the multiplayer – where on the multiplayer side it was the replay value, and on the campaign side the more relaxed, stress-free narrative type of environment. Which was how the co-op missions got started.

They’re fantastic, and my go to mode for StarCraft II multiplayer.

Even though it’s only been a year we believe that they’ve been a huge success within the community so that’s why next year we’ll be pushing both the traditional multiplayer and the co-op missions. Which are primarily there for players who just want to have fun with their friends and not worry about the mastery or the competitive side of StarCraft multiplayer.

It’s great to know that StarCraft lore is still getting some love, but with something like the War Chest micro-transactional feature there’s a sense that the focus now it almost entirely multiplayer.

The thing you should know about the War Chest is that it was requested by our community. We would often see posts by players asking for things that they could purchase, like skins, voice packs, and stuff like that. Which is pretty rare. It’s the first instance that I’ve seen, where the player base actually wants to pay for more stuff in a game. So we thought, if that’s the case and that’s what they want then why don’t we develop something so we can focus on that whilst also looking to support the Esports side of the game. And that’s how that system came to be.


Speaking of Esports, how do you guys manage to patch the game when there’s tournaments happening all the time.

We try to make the smallest changes as possible with each incremental patch because of that reason. There are StarCraft tournaments going on all the time. And even though our next major patch is launching in the off-season of the World Championship Series, there are still cups and tournaments going on and casters are always quick to remind me of that. But, you have to pick some point right.

How does community feedback factor into the patching process and balance changes being made?

For the past couple of years, we’ve been working closely with our community members to make sure that we’re making the right decisions for StarCraft II. What we do know is, when issues come up with, we simply create a post saying, “Hey, this is what we’re seeing, our reasons for thinking this way are so and so, here’s what we want to do, but if you guys have more thoughts, then let us know.” After that there’s probably a week or so of community discussion on Reddit, Team Liquid, and the Battle.net forums which we try to stay involved with. Letting them know when they make some good points but also letting them know when we don’t agree with their proposals.

So we go back and forth a lot and then once we reach a solution and we have the numbers and changes, we then put them in an online balance test map. We did use PTR’s before but we noticed that not a lot of people participate there because you have to download a separate PTR client. With the test map, that’s in-game, it has a matchmaker and that’s a lot easier to get people to participate. From there we look at the data, get more feedback, and even enlist the help of some of the pro players and casters. But even after we push the changes to the live game, we still monitor them very closely for a while before we move on to the next thing. Just to double check and make sure nothing goes wrong. Because even though a lot of people participate in testing these changes, it’s still only a small percentage of the user base.

There’s no doubt that StarCraft II’s multiplayer is quite complex. Shifting gears a bit, with the recent announcement of Blizzard’s partnership with AI researchers Deepmind, do you think its feasible for an AI to beat a pro human player?

There seems to be a split opinion, because some people think along those lines. That StarCraft II is such a complicated game that it would be hard for an AI to play it. And I kind of lean on that side. On the other hand, if you had no restrictions on the AI I don’t think a human could beat it. For example, no matter how good a human is it’s physically impossible for a person to micro across multiple different screens. Because you can only really look at one screen. But with an AI, if you have no restrictions then it could be micro-ing perfectly in ten different places. So, there would have to be some sort of level playing field. Let’s say an AI was carrying out two different marine drops across the battlefield and micro-ing them all at the same time perfectly, then to viewers that would look like…

Like cheating?

Exactly, so as long as restrictions and rules are set then I think it would be very difficult for an AI to beat the best human player in the world.


How does that relate to the existing “AI” in StarCraft II? Because, there’s no way I could beat the AI on Insane difficulty.

Well, Insane cheats. Because it knows what’s happening always and it has a number or reaction and counters to draw from. As for the other AI difficulty levels, we try to have them scout and then react accordingly. We also have quite a few different build orders that they can choose from at the beginning. And when players figure out ways to exploit the AI, we have to specifically hand craft their reactions to things like Canon rush and Bunker rush. But that’s all very basic stuff compared to what these guys are trying to. They seem to be trying to create the best StarCraft II player in the world. Whereas for us, we just need an AI that someone can practise against before they head into the real challenge -- PvP.

Where do you see StarCraft II’s multiplayer in the years ahead? Do you see it as a game that will last for over a decade like the original?

For sure. For us the goal we have on the design side is that we want the game to be in a place where major changes like the ones that are going in now aren’t going to happen anymore. Where the game is strategically and diverse enough that it could last for a decade or longer. That’s the goal, but right now I don’t think we’re there yet. But hopefully after the next major patch goes out we can use the next year to evaluate the changes, look to see if there are any missing holes, and hopefully get closer and closer to that ultimate goal.

Thanks to David Kim and Blizzard for the chance to sit down and chat.