• Newsletters
  • Submit News
  • SCLegacy RSS


Dustin Browder: So we're going to make a presentation for you guys that sort of describes some of the processes that we go through, and gives you a little bit of insight into the stuff that we're looking at right now, maybe some of the things we have planned for some of our future patches. Is there anybody out there that has any questions at all about StarCraft II multiplayer? Anybody? Just a few people, all right. So get your questions ready, because we're hoping we have a lot of time for you guys to ask us whatever you want about what we've done or what we're going to do down the road. All right, so I'm going to talk a bit about how we balance the game. So, we have a lot of different tools that we use to balance this game. We have community feedback, we have pro feedback, we have our own play experience, we have a lot of stats that we gather, but the important thing to know about these tools is that they all fundamentally fail us at some critical point. Every one of these tools, at some point, we thought was the promised land. We had finally found the tool that would save us, that would be the perfect answer to everything we need, and then we found out it kind of sucked. So the only trick that we've got is we have to use all of these tools together. We want multiple tools to agree about what the direction is, what we should do, and only if these tools sort of fight together as a team be of any real hope of finding the right answer about what we want to do in the balance of the game.



Matt Cooper: So I'm going to talk to you guys today about our first tool, and this is kind of our first line of defense. And it's you guys, it's our community. And one of the best things about this is, hundreds of thousands of players play millions of games, and of course this is across all skill levels, all regions, and we get some great debates from you guys. It's really not that useful for us if we have an internal tool that tells us, "Building Battlecruisers makes your win percentage go up by 3 points." That doesn't really tell us anything. Maybe the players making the Battlecruisers have already won the game. But when we get these great debates and we can see in what situation a Battlecruiser might be overpowered, against what race, how players are using it. And of course, like any good debate, you get the opposite, people arguing, "It's not overpowered, here's how to beat it." Of course, like any of our tools, there's some problems. One of the biggest problems of player feedback is, "The loudest voice tends to win." It's not necessarily about who's right or who's wrong, and not everyone chooses to post on the forums. And of course, we get a snowball effect where one player comes along, makes a thread, and, you know, a bunch of people jump on that, and by the time someone who disagrees of this or has good debate to bring to the subject, you know, there's twenty posts in a row about how he's wrong, and he doesn't really want to get in an argument with twenty players. But you guys are great at - you're one of our best tools for finding bugs, and a lot of the stuff that's blatantly wrong that we can't necessarily find in our internal playtesting because we don't have enough time, you guys find almost immediately and it helps make this the best play experience we can for you guys.



Dustin Browder: Yeah, I think a good example is, from the beta, we had many times where we - we played tens of thousands of games, right, and we had not seen a lot of the stuff that you guys literally found within like, three hours of the beta going live. Like, sitting there playing on Battle.net, I'm getting overrun by Gateway pushes, and I'm like, "Oh my God, how did we not see this?" Right? And then, you know, patches were incoming almost immediately.

David Kim: Okay so, the second tool that we have is pro feedback. Pros really know the game and really lead the strategies that everyone else on Battle.net follows. One example is in the IEM tournament finals, Morrow, a great Terran European player, played against Idra, and beat him pretty easily using this one specific strategy of going mass Barracks with mass Reapers early on. And, within a couple days, all over Battle.net we were seeing this exact same strategy. Even in Gold, even in Platinum, and especially among the Diamond-level players. And because pros really know the game, they give great critical feedback. For example, Maka, one of the best players currently in the world right now, he's actually here today at BlizzCon as well, he really took his time to refine the Zealot/Sentry/Void Ray strategy. And he not only sent us replays, but also a very detailed explanation on when to build every unit, as well as how to use these specific units in combat when you engage the enemy. And, using this feedback, we were able to make the changes to the Void Ray in the patch that we released last week. So the one downside to pro feedback is, pros generally only favor one race because they only play one race. So pros have a difficult time judging victories or defeats. "Did I win this game because I was more skillful than the other player, or was there a race imbalance?" That's always the hard question for pros to ask themselves. So second downside to pro feedback is, pro feedback doesn't necessarily represent all skill levels. Even though people try to follow pro strategies and execute it on their own, the level of execution is not the same, meaning we get different balance across different skill levels. And even though our focus in balancing our game is for the competitive level, we do want the game to be fair and balanced across all levels as well.



Dustin Browder: So one of our other great tools, of course, is tournaments. We're getting a ton of these here at BlizzCon. If you guys haven't checked out the RTS Stage you really should.

Matt Cooper: They're so awesome.

Dustin Browder: We've got our homage there to Fruit Dealer, the winner of the first season of the GSL. [crowd cheers] Absolutely, give it up for Fruit. So the advantage of tournaments games, of course, these are great, great games with great players playing them. And, this is the ultimate test of our balance. If there is a tiny crack in our game balance, these guys are going to drive a truck right through it. Right? Everything is on the line for these guys. They've got tens of thousands of dollars on the line, their career, their pride, everything, and they've been practicing eight to twelve hours a day for months on end. So if there's any problems, we're going to see it. We don't tend to look too much at tournament results. And that's because the skill between all these players is not necessarily even. There's no perfect matchmaking in tournaments, right? Sometimes, one player just practiced a little bit harder. So I know there was a lot of concern in the community when Fruit Dealer actually won those matches, because there was a feeling that Zerg needed a buff. And we actually kind of agreed with that at the time, and then Fruit Dealer went ahead and won, and we're seeing a lot of posts, like, "Oh my God, now they're not going to do anything because they have proof that Zerg is okay." That wasn't proof. Right? Fruit Dealer was on the top of his game. He had practiced, he knew everything he needed to know. I think Fruit Dealer would have done great with any race that he practiced on that day. He was top, top, top form for that matchup. So, you don't know what players also didn't show up for the match, you don't know what players were sick that day, you don't know players maybe got unlucky in an early round and got eliminated by a matchup they weren't quite as comfortable with, there's lots of variables in tournaments. So, we do look at the games, and that's a very powerful tool, but the exact results would only concern us if we saw something like, "Oh my God, it's just all Terran all the time, no one's even choosing the other races."

Matt Cooper: And of course, another tool we use is - we play our game. We play our game a lot. And, everyone who works on balance is a Diamond, random player, so we get to see all the matchups on all the maps, so if there's a six-pool flavor-of-the-week or something you guys are seeing in your games, we're probably seeing that same strategy in our games, and this is great for us. And of course we have tons of people across Blizzard giving us feedback at all different skill levels. Diamond players, Bronze players - and we do focus for the e-sports level, for the best play, but if we can fix something at a Bronze level that's not going to affect e-sports, we'll try and do this. We want everyone to have a good experience when playing our game. But, again, this is a very small sample size. Like we mentioned earlier, we had played tens of thousands of games going into the beta, but we still couldn't catch some of the Warpgate or Gateway rush strategies, and, you know, we need a bigger sample size to catch a lot of that stuff.

David Kim: So we've taken a look at all the opinion-based tools that we use to balance our game, so now we'll look at some concrete tools that we have. First is spreadsheets, and they're great for making simple calculations such as DPS. And in StarCraft II, not only the damage per second matters, but the damage per shot as well as how often a unit fires matters as well. And it's great to just lay out all these different numbers for every single unit in the game in order to balance our game. Another great thing about spreadsheets is that, they're great for analyzing buildtimes and costs. We don't try to equalize the costs of different tech choices, and we also don't try to make every production building needed for every one of the races symmetric, but we do try to make note of all these numbers so that the balance on these numbers can be perfect. The major downside to spreadsheets is that, it actually doesn't take into account a lot of the other variables that are going on in the game, such as pathing, terrain, or strategies, or even micro. And to take these into account, we'll actually look at our other tool.

Dustin Browder: So we have a tool, that's an in-game tool, called "Make Combat". And this is sort of a debug tool that we can use to create battles very quickly and run tests rapidly to see what's going on. And of course, what's great about this, it's in the software so it accounts for movement, it accounts for pathing, they can account for a lot of the variables that something like a spreadsheet wouldn't use. Like we often have seen, changing something as small as a unit size, as I'm sure you all know, has a significant impact on how well that unit performs. So you can, you know, make a small tweak to the size of the Zealot, quickly use Make Combat, and see what's actually going on. And it also can account for unit cost. Like, StarCraft is often balanced for cost. Right? We're not balancing unit-to-unit, we don't care if seven Zealots beat seven Marines, that doesn't make any sense. The Zealots cost a lot more. So we can actually set up battles in Make Combat that allows to test for unit cost to make sure that we're getting it right. So here's an example of Make Combat in action. You can see, we're going to type in a command there, we're going to do it for cost, and we're going to speed up the game, so we can watch really quickly what's going on. So those Zealots and Stalkers are going to fight those Roaches and Hydras and, as you'd expect, it's not going to go very well for our Protoss brothers there. They're going to be beaten down, get a little report there at the end, how the units are left, what percentage is left, what the leftover cost is. And we run it again just to make sure that, "Yeah this is what's happening, nothing's going on." Let's change it up. Let's put in ten Immortals, in [inaudible] those Stalkers and let's see how it goes. And so, as you'd imagine, this is going much better now for the Protoss, the Immortal being so strong against the Roach. But like everything in StarCraft, a lot of these relationships are very touch-and-go. And so we want to run it a couple of times to make sure. And we'll run it again just to see what happens, and because of small variations in pathing, and small variations in unit target acquisition, suddenly, what was an absolute Protoss victory, suddenly turns into a narrow Zerg victory. And these are the kinds of relationships we're looking for, because this means terrain and micro will be the deciding factor in this particular battle. So if we talk about - so, Make Combat, of course, we were in love with this tool. This was the greatest tool ever made. We had it, we could just test the game fifty times an hour, we could see whatever we wanted to see, and we could prove to ourselves that everything was great. We were so happy with this. The downside of this tool, of course, is, it's sometimes difficult to account for terrain. I mean, you saw that battlefield. When was the last time you were in a battlefield that was that open? Now, we do have some terrain on the debug map, but it's easy to forget to use it as a designer. And reality is, battles are rarely that perfect. It also fails to analyze the current strategies, what we call the metagame. It fails to analyze what was going - what can actually happen in a real battle. Really? Ten Immortals and ten Zealots? If four Mutalisks flew over that formation, those Protoss would be dogmeat. I can convince myself with that battle that, "Oh, hey, look, this is easy for the Protoss, they can counter this with Immortal-Zealot." And actually when you get into a live game it's a total disaster. This is what we found when we went into beta. We've been using this a tool a lot before we went into beta, and when we got to play a lot of games against people of our own skill level, suddenly we were having lots of arguments. We would say, "Make Combat says this," and I'm saying, "Hey, I get it. I can see what Make Combat does too, but I'm telling you, I played last night, and it doesn't work, and that's not going to happen." And so the danger of Make Combat we discovered was it creates a false sense of security. It makes you believe you know what's going on, and instead of using the other tools, you can start to rely on it to give you proof of what you want to believe. "Hey, I know, we got the answer, it's Immortal-Zealot. It's going to stop it every time." When you get into a game, it's not going to happen. And even look at that scenario I showed you. Ten Immortals. How many Robotics Facilities did this fictional Protoss player have to put out ten Immortals at that stage of the game? It's just not practical. And that's why our other tools are so important, and why this is just one small piece of the arsenal we use to try to understand what's going on.



Matt Cooper: And of course, we have Battle.net stats, and the great thing about this is it's everyone playing the ladder. There's a lot of players who might not post on forums, a lot of great players, but we have all their games on Battle.net. And this takes the subjectivity out of what they're doing. A lot of players - if you ask someone, "How did your last thirty games go?" Most players probably couldn't answer that concisely. Even internally, we'll get someone coming up to us saying, "Man, I can't beat Protoss as Zerg." And, you know, we'll give them some advice and then we'll go look at their match history and we'll notice that they actually have a seventy win percentage in that situation. And it's not that they're trying to mislead us or anything like that, it's just, players tend to remember their losses a lot more, and it's really hard to think back what happened in a bunch of your last games. So, let's take a look at some Battle.net stats, and this is for North America across all leagues. So, Terran versus Zerg, we can see that Terrans are edging out Zerg very slightly, fifty-one to forty-nine in favor of the Terrans. Protoss versus Terran, fifty-three to forty-seven in favor of the Protoss, still a really good number for us. And Protoss versus Zerg, fifty-one to forty-nine in favor of the Protoss. So I mean, these numbers all look great to me, I think we're done.

Dustin Browder: It's perfect.

Matt Cooper: So we can just go home, right?

Dustin Browder: What could possibly go wrong with these numbers?

Matt Cooper: Let's see.

David Kim: But with there's every other tool that we have, this has a huge downside to it as well. And the main reason for this is our matchmaker. Our matchmaker does an excellent job of matching two players, that are evenly skilled, that will go fifty-fifty against each other. However, the matchmaker doesn't know where the race imbalances are, so it doesn't take this into consideration. So, in order for us to use these win percentages as a balance tool, we actually have to know what the race win percentages are for each matchup when two equal level players play, and not with perfect matchmaking. So to explain how we get this actual number that we use, we have our math genius, Doctor Josh Menke.

Josh Menke: Thanks David. Yeah exactly. The matchmaker is trying to make sure every player has a fifty-fifty win percentage, but it's also making sure every race has as close to a fifty-fifty win percentage as well. And so because of that, we can't really trust those raw race win percentages, so we have come up with an adjusted win percentage that accounts for the skill of the players. And not only that, we can use this to track the metagame day by day, week by week, we can know what Protoss versus Zerg looks like at the top level, at the middle, and at the bottom. And this is a great tool for us to be able to know what the current metagame really is. Now, you're probably all wondering, "How is this possible?" So, to kind of boil this all down, we throw up the calculation here, and just let you guys see that and use that whenever you want to understand how we do this.



Dustin Browder: You guys got that? Seems pretty basic, right?

David Kim: I got it, I got it.

Matt Cooper: Pretty basic math, I think.

Dustin Browder: Yeah, simple stuff.

Josh Menke: Seriously, this is how we do it. I didn't throw on the - how we handle the different levels of skill, but that'll give you guys a clear idea. Now, let's run through a simple example though of how we actually fit that equation there. So let's say we have two players: we've got a Zerg player, and a Protoss player. Before the match, we kind of think Bob's probably the better player; we've seen him play other players, we've seen Frank play other players, so we kind of have an idea about their skill. We've also have seen Zerg play against Protoss in the past, so we also have an idea that we think, in this case, that Zerg, at this point, is better than Protoss - slightly. So, then we have those two teams play, we observe the results, we saw the Protoss player won the match. So now, we need to change what we think about the skill of both the players and the balance of the matchup.

Dustin Browder: With that cheesy Void Ray rush too, that's terrible. That's terrible. I can't believe he did that.

Josh Menke: Yeah, we don't [inaudible] how he did that. So here we go, before we thought Bob was better by about five percent, but at the end of that we're going to slightly update that, change what we think, and now we're thinking, "Maybe Bob's only slightly better than Frank." And before we thought Zerg had a three percent advantage over Protoss, we're going to kind of back that off a little bit now because we just saw Zerg lose a match. And we do this over millions of matches, over hundreds of thousands of players. We don't really update win percentages, we're just showing this to make it easier. It's a much more complicated system than that. But this does give a flavor of how we're able to do this and get these adjusted win percentages that gives a more realistic view of what the race balance is.

Matt Cooper: So now that we know we have to factor in the skill when we're looking at our Battle.net stats, let's look at the adjusted win percentage for North America, again, across all leagues. So for Terran versus Zerg, Terrans are still doing pretty close against Zerg, fifty-one to forty-nine. This is a really good number for us. Protoss versus Zerg, fifty-three to forty-seven in favor of the Protoss. Now, a number about a fifty-five to forty-five means there's no statistical evidence of imbalance. If that number starts going closer to sixty, forty, then that starts getting in our danger zone and this is a number we definitely want to look out a little bit more. So, Protoss versus Terran, what do you know - sixty, forty. This is definitely a number we're concerned about. So, we're going to investigate this a little bit further, and we'll look at Diamond specifically in North America. So for Diamond, we have Terran versus Zerg, forty-nine to fifty-one, still a good number. Protoss versus Zerg, fifty-five to forty-five; it's borderline, but we still consider this a pretty solid number. And Protoss versus Terran, you can see it's a lot better: fifty-three to forty-seven in favor of the Protoss. Now, this does agree with what we're seeing across all leagues, but it's a much better number. So, we're still going to want to look at this a little bit further.



David Kim: So before you get into the race win percentages of Korea, I just want to talk about the Korean region in general. Korea, we've seen in the past, they always lead the world in terms of strategies. So, for example, the highest - we've seen in the past that highest level players in Korea always develop these strategies that, say, top American or European players tend to follow, and, for example, just internally for myself as well, a lot of people come to ask me if I have a plus ten APM advantage just because I'm a Korean person. And even though we haven't really scientifically proven this or anything, but, questions like these arise because there's the highest concentration of top players in Korea. So, one of the examples we saw in the beta was, in Korea, during the beta, at the highest levels, Zerg was dominating Protoss. However, in US or even in Europe, this was not the case, it was actually the opposite: Protoss was dominating Zerg. So, we took a look at the strategies that are being used, and in Korea, the top Zerg players were using this mass Spine Crawler plus mass Mutalisk strategy against Protoss that Protoss was just unable to beat. And within a matter of couple weeks we saw that, American and European Zerg players were using the exact same strategy, and a couple weeks after that, we saw the stats kind of align across all regions. So, let's take a look at the stats for Korea across all leagues. Terran versus Zerg, Terrans have a four percent advantage, which is pretty good. Protoss versus Zerg, Protoss has a one percent advantage, which is even better. And, if we take a look at Protoss versus Terran, Protoss has a six percent advantage. So, this number actually agrees with what we saw in North America, but let's take a greater look at Protoss versus Terran in specific. So we've just seen that Protoss, in Korea, Protoss has a six percent advantage. In North America, Protoss had a ten percent advantage. But, if we look specifically at Diamond Korea, which is the highest level right now, the relationship is actually flipped: Terrans have a four percent advantage over Protoss. And we think the reason for this is because top Terran Korean players are really amazing with their Stimpack kiting micro as well as multitasking Medivac drops all over the place, and perfectly microing all these things all together at once. But, as I said before, we can't just go by our past experiences, so let's take a look at our other tools to see if that's what they say also.



Dustin Browder: So, Battle.net has identified a problem for us very clearly. We can see that - we're seeing Protoss dominating Terrans almost across the board except in this one place that makes us nervous, this one place that we've seen in the past can be different and it may change our strategies. So if we look at our other tools, and now to go to the community. Now we've got about a dozen guys I think, all over the world, guys and gals, who are gathering feedback for us, from our own websites, in languages none of us speak, right? And gathering feedback from fans from other forums, other websites that are made by the fans, and they're bringing this to us as best they can, trying to give us a summary, of kind of what people are talking about, what the most popular threads are, and what people are actually saying. So the community says that Terran is too strong. The community says we should probably nerf Marauders. This is something we've been hearing since mid-beta, that Marauders might be too powerful. Nerfing stim, usually focused again on the Marauder, though sometimes on the Marine. Though none of this agrees with the stats, except in that one league in Korea.



Matt Cooper: So if we look at our Make Combat tool, Marauder Stim is not overpowered. Now, you do get a lot of speed movement out of it, but the actual bonus damage you get from using Marauder Stim, versus the health you pay to use it, it's very borderline. We've tried a couple small tweaks on this, and it actually, in some situations, it's actually bad to use Stim. Another situation is where, Marauders against Zerg. They're very borderline. A couple small tweaks to Marauders and Roaches start beating the Marauders. And this is not what we want; we want the Marauder to be a counter unit to the Zerg. And of course, if we go by our Make Combat tool, Marine Stim is one of the most broken things out there. If we went by this only, we would have to nerf marines, and - according to our play experience, and what the community is saying something, I don't think this is what we want to do.

David Kim: So let's take a look at what pros have to say about Protoss versus Terran. We actually asked five different pros - five of the best Protoss and Terran pros out there, to see what they have to say about this matchup, and the first player actually says, "Terran wins eighty percent of the time versus Protoss." The second player says, "Toss has no chance versus Terran." The third player disagrees by saying, "Templar with the energy upgrade is too much for - too much." And the fourth person says, "Terran has no chance versus the Phoenix plus Colossi combination." And finally, the last person says, "PvT feels very balanced." So, two people have said Terran is overpowered, the other two players have said Protoss is overpowered, and one person says the balance is perfect. But it's very rare for us to get these kinds of mixed opinions, so there must be something going on besides just one strategy being overpowered, or one race just dominating the other race, so let's take a look at another tool.

Dustin Browder: So this is where it gets really exciting. We think we might bit a situation where it's not down to just one or two units, it might be a large combination of factors, and depending on how you choose to play, giving you a great time in this matchup, crushing your opponent, or you're getting rolled, depending on what race you're playing. So, let's look at our sort of last line of defense here it's our personal experience. We do find Terrans to be very strong at the beginning of the game. I often get run over early on with stim timing pushes. It wrecks my day every time I have massive amounts of nerdrage every time it happens. We do see in many cases the Protoss being a little strong in the endgame. If the Terrans can hold out it seems like a combination of Colossus and maybe more importantly Psi Storm suddenly gives Protoss an advantage. And this leads us to believe that there might be a more fundamental design problem here. That while we could maybe be to a place here pretty soon where technically they're going fifty-fifty all that really matters is "Can the Terrans win in the first twelve minutes or not?" And if they don't win then the Protoss win. And while "I suppose we could say that's fifty-fifty good enough." That's not the game we want to make and I'm pretty sure that's not the game you want to play. So that means we do have to look at this issue we have to try to figure exactly what the combination is that's causing our problems.



Matt Cooper: So going forward we're definitely going to keep looking at the Protoss versus Terran matchup. We think we identified a problem between our Battle.net stats and all our other tools so we want to keep looking at this matchup closer. And one thing we're definitely going to look at regarding this matchup is Stim balance on Marines and Marauders. We do think it gives the Terran a little bit of - too much of an advantage early in the game and they're probably winning a lot of their games with this strategy. But of course we're going to look at Psi Storm balance and how it affects the late-game. I know personally once I have both High Templar upgrades I got Warpgates I can warp anywhere across the map I really don't have that much trouble against Terran. And we're going to look at a number of other things: Thor repair how the SCVs can be untargetable and as you guys keep bringing us new issues and as we notice new things in our games we'll continue to add to our list.

Dustin Browder: Yeah so I want to thank everybody again for all the feedback you've given us everybody out there who's posted everybody out there who's never posted but has played a game has helped us at some point" make some decisions on the balance. And that's it for our presentation; let's get the questions started.





Q: Hello. StarCraft is amazing" I love it. One of the things I thought was peculiar was the fact that there's no option to not play online. Do you see that coming in the future? To where you'll be able to work on achievements offline then link up to Battle.net and have the stats queue up?

Dustin Browder: It's possible. That's something we talked about quite a bit with if we want to do that or not. Ultimately it let us more exposed to people hacking the achievements if they can do them offline then log them in but I'm not actually sure how secure we are now anyway. But it's definitely something we can look at. You obviously can play offline now as you say but we don't give you achievement access. But I'll take it as feedback and we'll definitely talk about it. Thank you.


Q: So the Mutalisk seems a little bit... Well the Thor and the Marine both hard-counter the Mutalisk. And the whole reason they gave the Thor the splash damage was to counter the Mutalisk. Maybe like give the Mutalisk +3 to light? What do you guys think about that?

Dustin Browder: Well I think we feel - I mean you guys can back me up or say whatever you want about this - I think we feel like the Mutalisk's speed is his greatest asset. His ability to be pretty much wherever he wants to be and hunt whatever he wants to hunt makes him a very dangerous weapon. And we're also very nervous about any weapon that can attack air and ground. That's a very flexible weapon in any unit's arsenal. Something that moves faster than almost anything in the game and can attack air and ground is something we're definitely going to be very careful of giving too many buffs to. I think we feel pretty good about the Muta right now. You guys want to say anything on the Muta?

David Kim: Yeah um because the Muta speed is so fast and the other two units you mentioned the Marine and the Marauder are both very attack-move-friendly units. So there will be a difference in - at the start I think - in using these units. But as players learn to use - make use of the speed advantage of the Mutalisk we feel the balance might actually be better.


Q: Hi. I'd like to find out - if you can tell us - when will fully customizable key bindings be added to the game and why didn't it have them in the first place? Because I've tried the different key layouts all of them and I think I could do better if I could set the keys up my way.

Dustin Browder: Yeah no I totally agree with that. I think we've had that feedback from a lot of players. The reason it didn't make it for the game was I mean while we were six plus years in development we didn't want to hold it any longer. And we didn't feel like customizable key bindings was a reason to delay the game. So we held that back and we've been working on it I've actually seen the design in progress it is being worked on right now there is UI currently in our internal builds where we're testing it. As you can imagine the UI for custom key bindings is not simple. It's - actually I would say it's insane. It's absolutely nuts. And our biggest fear is there's this giant landmine that new players will step into. They'll accidently bind a key and then they'll have no idea how to do anything anymore. So we've been working really hard on the design of it and on the engineering. I don't have a date for you exactly but I can tell you it is under construction - it's actually pretty far along. I think we should have something at you guys pretty soon.


Q: It's just funny that you asked that because I just released a hotkey trainer like one week ago. So check out for that on North America if you want to train on that. But anyway - okay um my question is slightly off-topic it's about custom maps. I know you were talking a little bit at BlizzCon 2009 about the custom map marketplace that might eventually come out and maybe premium maps and stuff like that. And I put a team together with the hopes of you know making the absolute best maps out there eventually when that marketplace opens up and release some really awesome stuff and - I'm just wondering if you have any more details about when it might be coming out. I don't want to hold you to anything that you say just any hunches that you have -

Dustin Browder: No I can tell you it's not coming out in the immediate future. It's probably closer to a patch in Heart of the Swarm or Heart of the Swarm to be the soonest to be able to get that out to you. There's a lot of challenges we have to get that going. But I love that you're putting a team together and we know that the fires are burning we know that people want it. We'll try to get to you absolutely as soon as we can it's one of our top priorities coming forward.

Q: Awesome thank you very much. And also the ten megabyte map limit is there a chance for that to change in the future as well?

Dustin Browder: There're many arguments - there's throwing of chairs it's been really rough in the office over this issue. We are consuming a lot of our storage very quickly but we'll see. It's definitely something we're always talking about I think I have like a weekly conversation about that at this point. It's definitely a topic for us so we'll take a look at it.


Q: Yeah by the way love your game - your guys' game been playing since I was a kid for the original StarCraft. But what I'm kind of asking about is like the early gimmicks of the games specifically the proxy rush for the Protoss as well as like the early Thor rush with the single Thor with the repairing. How do you guys usually account for that - with catching those? And it felt like it took a while for those to get fixed. I mean do you see [indecipherable] those come out you try to get those fixed as possible or does it take a while for you guys to kind of understand what's going on?

Matt Cooper: Those are some of the easiest strategies for us to fix so when we're playtesting internally we're more focusing on the more complicated scenarios where you have multiple bases running what the 200 versus 200 pop armies are doing against each other and making sure that you know all races can stop rushes and stuff like that. But internally we don't try as many you know proxy rushes or six-pool rushes and stuff like that because they are very easy for us to fix once we see there is a problem.

Dustin Browder: Sometimes. There - I agree with that 95 percent of the time. I think there's the occasional - and we've had some that were really touch-and-go because a lot of the units used in the early game to rush are also used to block other rushes. Right? So it can get really challenging. But for the most part I think we've been able to crank those out pretty quick. I think there was a Terran SCV rush that was actually probably our scariest one with Marines behind the SCVs the SCVs the wall and that led - I think that was in the late beta - led to actually a nerf on the SCV health. But the same time the SCV was essential to stopping some of the rushes that can happen against them. Definitely one of the challenges we have is we want those rushes to be somewhat viable. We want them to be available we feel like that makes the game scary from the first minute you click on anything in the game and to the last. A lot of games are trying to get rid of those early rushes and they say "Rushing is bad we don't want it" right? I don't know that's my other developer voice I guess. But we don't like that we want the rush to be possible so it's always a tough call. And we definitely welcome your guys' feedback whether a particular rush is still too easy to do and still too hard to block. In many cases it's just a key of detecting it and using your workers to fight back but it's definitely always sort of a touchy one because a lot of new players get rolled by these and they're pretty unhappy when that happens.

David Kim: And because they're such a core part of the game we try our best to teach players how to block these rushes and our multiplayer challenges is actually a good tool for you to use to learn to counter these rushes.


Q: Hi. I think you guys did an awesome job balancing the game so thank you so much with all the races. My question's very simple. I always go into a bit of a nerdrage when I get mass Hydra rushed. I play Protoss so I'm wondering if you guys can just real quick give me some strategies and - don't say Templar because - I don't know I have better luck against Terran but when it comes to Zerg I just get destroyed when I get mass Hydra-ed.

David Kim: Well the main downside of mass Hydras especially on the offensive is that they're really slow so they can't kite anything off of creep. So uh did you say you were playing Protoss? Yeah so even just using mass Zealots with Stalkers shooting - over the shoulder shooting or even combined with Sentries with Guardian Shield that cuts down the Hydra DPS by quite a bit. Should be good enough for you. But eventually you're going to have to get out the splash options. If the scale becomes really big say if he brings in 50 Hydras obviously a clumped pack of 50 Hydras is a little too much for only tier 1 units to handle. So if you don't like using Psi Storm then maybe you can probably try using Colossi.

Dustin Browder: Yeah Colossi aren't too bad they work pretty well in attack-move it can be effective especially like you say if he doesn't have a bunch of Corruptors out that's also an option for you.


Q: Hello. I love this game it's great. You guys in all of your history of all of Blizzard games you are great at balancing everything. So one of my questions really is basically as it goes to the score screen are you ever going to expose at a more atomic level how you calculate what the scores are on that screen or allow users to look at that data in their own replays to sort of analyze that stuff more statistically sort of the way that you guys do with the matchmaking and stuff?

Dustin Browder: Are you talking about like your hidden rating or are you talking about just what happens on the score screen?

Q: Well both things. Like whether or not - like how many units of each type you built in the game how many units did how much damage to what in the game things like that so -

Dustin Browder: There is not enough room in here for me to talk about that score screen. We had a lot of plans for that score screen which just didn't make it. Again we made decisions. "Do we want to hold this game even longer or do we want to make a better score screen?" At least like you know - we're just going to put it out with the score screen we've got but there's a lot more we can do with the score screen I totally agree with that feedback. I mean lots of things we sort of haven't planned. We want to give you better analysis of your game we want to provide you with some more hints as to what you could do better try to provide you with some more intelligence that's built into the game and like you say give you access to some of the more fundamental numbers. Now most of those numbers you're seeing on the screen are just for cost. So if you see a unit value that's not a specific unit value it's just how many gas plus minerals you did. So that stuff is kind of expressed but I agree I think we can do better with that score screen. I think you'd look for that stuff in Heart of the Swarm that's something we're definitely going to be pushing forward. There has been some talk about some of that stuff making a patch but I'm not sure. It might be more of a Heart of the Swarm kind of revision to that screen to give you a lot more details to help make you a better player. So you finish a match you lose you want to know right? Right before you launch another game you don't want to watch the replay right away. I think that's a great bit of feedback and I totally agree with it.

Q: I guess as a second part follow on to that are you ever planning on an ability to sort see what the score screen is in a replay without having to watch through all the end to get to the score screen? So you could kind of see the - it's like more games all at once rather than just one at a time?

Dustin Browder: We might be able to figure something like that out. The challenge in a lot of cases with these sort of suggestions is the UI issues. How insanely crazy does our UI really look does it end up like kind of a WoW raid UI where it's totally nuts? And that's not what we're going for. But I think it's a great idea I'd love to enhance the replay feature to give you more insight into what's going on in the game. I think it's a cool suggestion.


Q: Thanks. I love this game anyway - hopefully it's a simple question but I play mainly in Diamond leagues - I mean not that I'm great I'm not pro or anything but - one thing I've noticed in 1v1 and 2v2 is that Terran and Toss - and I'm sure this has been brought up so sorry if it's been brought up before - but Terran and Toss seem to have not initially a great advantage but a huge advantage with choking off at the early game. And I play mainly Zerg that's kind of the only race I play and I was just wondering if kind of the design for Zerg was that Zerg must out-macro their Protoss and Terran counterparts? And also is the creep kind of a complementary feature for the Zerg that they can move faster on so that they can defend a little bit better and have a slightly better - so kind of a weird question but just wondering if that was a conscious decision for every Blizzard map to have a choke point that Zerg must out-macro and the Terran and Toss can kind of defend against the early fast rush from speedlings like that.

David Kim: Yeah so we tried to make every race as unique as possible and in this case as you said we want the Terran to be the choke-y race and that's why we added options such as lowering Supply Depots to make the walling-off easier. And on the case of the Zerg we actually made a conscious decision - for example Spine Crawlers can be uprooted and be moved but we made a conscious decision so that Zerg can't block off a choke point by making it so that the Spine Crawlers' radius is actually smaller than the footprint which means there's no way for Spine Crawlers to block off a choke point right? So that was our goal. And in terms of maps we try to balance - we try to make the maps so that they're all balanced and fair for all different types of races that we have. And I guess having those choke points was a result of this much more so than about what the maps wanted to be I guess.

Dustin Browder: But I think your question you've already answered it for us. You're totally right. I mean our goal was to say "Zerg are the swarm race." We want them to out-macro we want them to win we want them to seem like they cheated like they just came in with way too much garbage and just overran me. And you're also totally correct that creating those creep highways is a core part of how we are trying to set up Zerg. To be honest we had no idea if it would work when we went to beta and even beyond but we're pleased to see that the sort of systems we set up are working the way that we sort of thought that they would. At the end of the day you guys are - it's really all in your hands now how it actually plays but that was sort of a goal to say that Zerg at tier 2 couldn't easily push a choke point right that was well-defended. So at that point they would have to out-macro and then once they get to tier 3 hopefully they either have some of the tools that they need in their units or they just got way too much stuff and it really puts the onus on the other races to keep the pressure up.


Q: Hi my name is [redacted] and good afternoon. I just have a question about the balance and I just wanted to ask: is it going to become like World of Warcraft where every single Tuesday we get another patch that makes one race overpowered? I mean the Roach thing was great for some players but for others I can defend a choke point with ten carriers and the Roaches destroy one building and go straight in and kill the Nexus. You know just because they have so much HP now. And - you know I know you guys are working hard but I just want to know in the future -

Dustin Browder: I think that's a very fair question. So we've done I think a couple of balance patches this month which is a little more than I think we would like and we had a few bugs and we did a few more patches as a result of that. We are going to slow down and I think we've already slowed down a little bit relative to beta. Beta I think we were just machine-gunning patches out as fast as we wanted to because well because we could. Nobody had paid for the product yet everything was free right we didn't have a customer base that was playing the game in tournaments even though they were in some cases that wasn't the deal we had made with them it's like "This is for testing." But as we go forward we're hoping to slow that rate. Now if we find out we got balance problems we are not going to be chicken and I'm going to back-patch once a week if I have to to get the balance to be correct. But that's not our goal. Our goal is to try to stabilize to try to settle down. We're feeling pretty good as we showed in the presentation the Protoss versus Terran matchup is making us a little nervous and that's one we're going to be looking at but we're hoping to slow it down so that we're not hitting you so often. I don't know exactly but our goal is definitely not that - the other thing I'll say too is WoW has challenges that we don't even begin to have. Right? WoW is balancing for PvE they're balancing for PvP at the same time they have a ton of classes in play they've got tons of builds on those classes that are possible. We've intentionally kept our game a little smaller. We've kept our game to three races. We've kept the unit counts from 12 to 14 per race on purpose so we could do this. We could very easily provide more game: "Hey here's five races and seventeen units per race" and then we'd probably be doing more of what you're describing and that would in our perspective not be such a great gameplay experience for a game that's designed only - for principally beyond the solo campaign - to be a PvP experience. So hopefully not but we'll see.

David Kim: And just to comment on the current state of the balance we feel the balance is very tight and the example you brought up is actually really good because right before this last patch came out a lot of people felt Terrans can't - I mean Zerg can't beat Terrans Zerg can't beat Protoss. But we made this small tweak of making Roach range from 3 to 4 and that small change made it so that players are actually saying the opposite especially the pro players. So that actually shows our balance is really tight and because of that we're only going to make small changes at this point.

Dustin Browder: We'll be cautious going forward. Unless we find out something's horrible. But assuming that doesn't happen we're going to be a little cautious.


Q: Hi. I'm ExcaliburZ. I made the long ladder analysis thread on Battle.net forums I'm the "Multiplayer in eSports" MVP and it's a pleasure to finally meet you guys. My question is what besides the hidden matchmaking rating influences player ranking promotion and matchmaking.

Dustin Browder: Want to take this one Josh?

Josh Menke: Yeah. So - I think a lot of what you've already posted is pretty - I can't come out and say everything is accurate I guess to put a stamp of approval on it. But it's close enough to make pretty good guesses at what's going on there. As Greg said yesterday when you play a match - and similar to what you've seen in WoW arenas - we look at your ladder points and we compare that to that skill rating in the backend. In addition something you may not know or you may have guessed at is bonus points don't have any role at all in that update. Aside from that we do have to do some other bits of math to look at the fact that you are in different leagues and the fact that the skill of your division actually does come into play. But other than that you pretty much have a good - you have enough knowledge there to understand what's going on within your division which is where we want you to focus. In addition as we come out with the Masters league and the Grandmasters league something that you might think is really cool about the Masters league is you - so right now when you rank people across all of Diamond league those rankings are not exactly accurate. In the Masters league if you do that you can trust those a lot more.


Q: Hello. I was sitting behind you yesterday in the HuK / Loner game and I started thinking "When you guys are out here watching the games are you guys watching it as the developers and the balancers or are you guys trying to separate yourselves and just sit back and enjoy the game?"

Dustin Browder: I'll speak for myself I'm just trying to be a fan at that moment. If I see something horribly go wrong I would probably switch to developer mode immediately and freak out but other than that I'm just screaming and cheering and enjoying the game like everybody else. What about you guys?

Matt Cooper: As a balance perspective I think we're always kind of turned on in the sense like we're always looking for stuff that's breaking what we have going on in our heads you know like "Ugh how is he using Roaches like we didn't intend that at all." But I definitely watch as a fan and I mean I didn't see anything go horribly wrong so I'm pretty happy.

David Kim: For me personally I watch a lot of pro games. I not only watch all the major tournament games but pros actually send us replays for us to watch as well. So because of that when I'm watching these games I'm not really surprised by the tactics they use so I'm just watching to confirm the balance of our game.


Q: I think that guy's word is "snuggy" I'm not sure. I was going to say "terrible terrible powerpoints" but they're not those are awesome awesome powerpoints you guys have. That's really good. My question is and I'd listen to the Protoss versus Terran balance you guys are looking at do you guys think the Warp Gate technology and how fast the Toss player gets it is somewhere you want it to be? Because I get the feeling that in oftentimes the Protoss can force a macro game very early and very quickly and I don't want anybody to mess with the gate stuff because I think it's awesome it's a unique ability but at the same time I'm like "Man I got to fight all that stuff again. I just built six Barracks and I still can't keep up with four Warp Gates." So I'm curious what's your opinion on the Warp Gates? Are they where you want them to be?

Dustin Browder: Well we're always trying to make everybody feel as broken as we can. Right? Like every race should feel insanely overpowered and yet somehow it all works. So I'm hopeful that we're there - they've actually been nerfed quite a bit over the last year I would say. So I'm hopeful but - I think they're okay. What do you guys think?

Matt Cooper: I think every race has a broken mechanic. Like Zerg puts down a Spire a Hydra Den and a Roach Warren and then you're fighting 200 pop Roaches and then two minutes later you're fighting 200 pop Mutas. Like you just don't know what's coming. So I think it's actually in a pretty good spot right now.

David Kim: What we're noticing at the especially pro level is we're seeing less and less of all these Warp-In rushes that you're speaking about. And I think the reason for this is because Warp-In rush is a very different type of all-in rush that people aren't used to from StarCraft or even in War 3. And top players are actually figuring out ways to counter it pretty easily. And I think eventually the skill required to master this will trickle down to the lower levels as well and hopefully it will be balanced across all levels of play.


Q: Hi. Do you think the Hunter Seeker Missile and their use - well it was a very cool spell before but now I think it cannot deal any more terrible terrible damage.

David Kim: Hunter Seeker was actually a spell that was causing a lot of trouble in the late game. Terran players were actually able to just take maybe one or two expansions lock those two down with mass Planetary Fortresses as well as mass Siege Tanks and all you do is just make Ravens on the offensive to trade their mana for enemy resources. Just being able to kill everything right? So because of that we had actually to nerf it because that kind of degenerate gameplay is not what we're looking for. And at the same time not every single spell in the game has to be generally useful in every single situation. We feel that in specific situations for example if you're playing against Zerg and he makes a lot of Mutalisks having a few Ravens casting Hunter Seeker is very useful. So we feel that as long as different abilities are very powerful in specific situations that happen on a fairly regular basis that's a very good spot for that spell.

Dustin Browder: We're definitely not - to your point - we're definitely not saying "used a lot". I don't think we're going to buff it in the immediate future because of the problems David talked about. We're going to kind of wait and see on that one to see what the community does with it. We're often surprised by you guys when you suddenly start using something that nobody ever used before and you're doing just a ton of damage with it. So we're going to wait and see on that one but there's no plans to buff it immediately even though I tend to agree it only sees uses in a couple of scenarios.


Q: Alright my question has to deal with the life cycle of a patch and how much leeway you give in to professional gamers in order to develop new strategies. Say the Roach range from three to four right before GSL. How do you give them enough time to say "Okay I'm going to change this now in order to develop better strategies."

Dustin Browder: Well I think the way to answer that is we really need a server that's sort of separate for those guys to play on so they can choose if they want to play the old patch or the new patch and that's the technology we're looking into. I think it's our long-term solution. We have announced in the past what we're going to do in advance and then it took us like a month to get that patch out at the end of the day. This last one we kind of sprung on people but we did have David and Matt Cooper went out there and made sure that we were looking at only the earliest versions of some of these tournaments. Right? The GSL I think was in the earliest rounds when that one first went out?

David Kim: Yeah it was right after the preliminaries but we were aiming for right before the preliminaries.

Dustin Browder: Right so we missed a little bit on that one. But we're trying to get it - obviously until we get that tournament server where people can choose which patch they want to play on we're definitely right now trying to target times where we'll do the least amount of damage to any major tournaments. Of course there will come a day where there are no - I hope there will come a day - where there are no tournaments out there that - there's no time for us there's always a tournament. Right? And there's always something major going on. And by that time we better have that tournament server up and running.


Q: In the last few patches you've made a number of changes to both Roaches and Reapers that have almost completely killed off the early Reaper attacks against Zerg but do you think that might have been too hasty and not given the community enough chance to really figure out strategies to counter that? Because in the recent GSL FruitDealer kind of showed a lot of Zerg players that we may be playing the race wrong. And this patch - or these series of patches may have just been too hasty. Do you feel that way?

David Kim: For the Reapers change specifically as we talked about in this presentation our main focus is on the competitive level but we do care about other levels as well as other gametypes as well. And in this case specifically especially in team games we're seeing that at the very highest level Terran plus Zerg teams were almost fifty percent of the top teams in both US and Korea. And these teams are actually using this very specific fast mass Reaper with mass speed upgraded Zergling strategy and if the number gets that high if the number becomes -

Dustin Browder: And their win ratios were insane right? Like these guys could not be stopped. And actually it wasn't even just at the highest level of play it was a lot levels of play. 2v2 has just been trashed specifically by the Reaper.

David Kim: And it's almost to the point where you can't play 2v2s at a high level unless you pick this combination right? So that's why we decided to nerf the Reaper. And on the flip side for one-on-one we felt that there's too many opening options for the Terran before the opponents can actually scout them. For example all these different proxies such as Reaper proxy double Barracks Marine Bunker rush as well as like a fast Hellion push or even a drop Banshees Thor repair rush all these options. So we actually wanted to limit these options and because Reapers were a problems in twos we were actually able to hit that in 1v1 as well.

Dustin Browder: I'll also say as well if you've watched any of the matches that have been going on on the stage over there I've seen a shocking number of Reapers today against Zerg players so - I don't know we'll wait and see. I agree the damage was high but we did it a little bit on purpose knowing that we were hurting the Reaper maybe more than some of our Terran brothers and sisters would have liked but I'm not sure the Reaper's just out yet we're going to have to wait and see what happens.


Q: First of all big fan. Dustin I love it when you cast games. I think if the Blizzard thing doesn't work out for you you have a career in casting. My question is regarding the - what I perceive to be an imbalance between the Thor and Ultralisk. Both these units cost the same both are kind of like the end weapon for these races but the Thor can be repaired it can shoot air and as far as I know it can beat the Ultralisk one-on-one. Is this correct? I saw this on a YouTube video. It was like three Ultralisks on three Thors and the Thors won. Do you feel like there's any kind of imbalance there?

David Kim: The Thors are actually supposed to win because Ultralisks are a splash unit and we don't try to make those two units identical. Thors are actually good - are the best against heavy groups of clumped-up air units light air units such as Phoenix or Mutalisk whereas the main role for the Ultralisk is to kill large groups of small ground units. So that's why you're seeing the difference there. And for every single unit in the game our main focus is to try to give it a unique and specific role and we don't try to make two units in different races roughly equal.

Dustin Browder: Yeah and the other thing too is a lot of Zerg units are a little weaker ultimately than their counterparts and that's largely because of Zerg's ability to build such a powerful economy. Zerg in many cases do need to engage with a larger force maybe even a significantly larger force to be successful. And since the Zerg can also save up a huge amount of larva which was not the case in Brood War you can often hit the enemy with a 200 pop army and follow it immediately with a 120 to 150 pop army right before he can even think about replacing his losses. And so that's sort of the strength of the Zerg. The Ultralisk doesn't need to win head-to-head against anyone he's got seven buddies with him. And that's what he wants to do to sort of run over the opposition.


Q: Hey guys I loved the SlayerS_Boxer FruitDealer match last night it was awesome. When I was going home I kind of thought to myself "Hey if Boxer was maybe spawned in cross-map locations he might have done a little bit better." I was wondering do you take stats that are relevant to player spawn locations? And I mean I know kind of large numbers maybe it kind of works itself out but I couldn't help but think maybe that has some impact on the metagame.

Dustin Browder: That's a really great question. I don't know if we have those stats I'm not certainly seeing them in my daily reports about the balance but map balance is extremely real. Right? It's a big factor what's going on. And we are going to - we could make the game perfectly balanced I suppose; we picked one map and never moved it from that. But we do want this versatility and variability of gameplay so you do have to learn different matchups not only by race but also by spawn position. So we do create variable maps with different balance points on them but we are in constant watch-out for maps that sort of get a little too far off the line. Right? Where it's a little too easy for one race to have an advantage - we've actually recently caught a couple maps from ladder play. Even though we love these maps but specifically because there was just too many imbalances with specific matchups on those matches we swapped other maps in. So yeah map can have a significant impact and it's definitely one of the variables we watch when doing a lot of balance.

Josh Menke: Yeah. We do track - in those adjusted win percentages we track an adjusted win percentage per matchup per map. And if that goes crazy we would take a much closer look. When we were first looking at what type of stats to gather in order to make sure the maps were fair I worked closely with the map designers and I asked them "Do you want location-based win percentage type stuff also" and they said "You know that might be drilling down too far" but if we got into a situation where it was obviously favoring a race then we could drill down more.


Ventrilo Servers
Contact Us About Us

SCLegacy is hosted by DarkStar Communications, home to high quality Ventrilo Servers.