Hey folks,

I’m brownbear. Today I’ll be responding to a thoughtful write-up from CatZ replying to my last piece on Brood War and SC2. I promised this response several months back, but didn’t get around to writing it until now – my apologies to CatZ.

I recommend you read CatZ’s post prior to reading this article. (And give him a follow while you’re at it, he creates a bunch of cool content.)

Terran Bio and Gating Mechanisms

CatZ makes an excellent point in noting how mechanical differentiation offers substantial incremental power to Terran players:

While I too think that strategic thinking is one of the strongest benefactors upon lowering the mechanical ceiling (by making the game more responsive / intuitive), former mechanical constrains can also continue to shape-shift into different ‘raw mechanics’ as you would classify them (as opposed to ‘strategic mechanics’). I think the most glaring examples are to be found in Terran / Bio play. For example a Terran player’s ability to quickly shift-click and kill incoming banelings, or to split their marines against aoe, to quickly studder-step marines offensively or defensively, to queue / control multiple drops or hellion run bys and so on.. and to attempt do all of this and much more in the most optimal way pertaining each situation is much of the essence of Terran micro mechanics.

…

Now, with this in mind; it is common that many or most of the best Terran players at a given point in time will open with similar strategies / build orders, seemingly foregoing strategical variety. It is in the Terran camp that, in my opinion, the argument made for Mechanical Differentiation and the shift by design from mechanics to higher-level strategy doesn’t seem to have as strong of a case.

Good Bio Terran players have the ability by design to take the game further away from a strategical war to a tug-of-war for speed and attention. In this realm “raw mechanics” from sc1 can often just shift to “other raw mechanics” for sc2, with fe, finer levels of detail in unit control / micro contrasted to the clunky 12 control unit groups you would otherwise be spending much of your actions in a fight on.

I’ve previously dived into the pecularities of the mechanics in StarCraft II versus other games in several different articles  – this one on why StarCraft II feels difficult to play and this one on Protoss race design are good examples. For the purposes of this write-up, I’ll focus specifically on the shift to “other raw mechanics” and how I think there’s more to it than that.

First, I do agree with this to a certain extent. But I think StarCraft II’s raw mechanics feature a substantial difference with older games in that they no longer gate strategic decisions. I mentioned this in my last write-up but I’ll try to make my point clearer here.

Take multi-tasking. In older games, it featured a more severe mechanical trade-off. On the micromanagement side, units would path poorly and easily get lost, requiring a lot of handholding and spam-clicking. Likewise, macromanagement was mechanically more expensive thanks to a lack of multiple building select, intelligent unit queuing, and so on. Any time spent managing a second army was a genuine trade-off in terms of fighting power – not only was it not free, it was often a relatively worse choice compared to paying attention to something else, even if a two-pronged attack was the strategically optimal decision.

StarCraft II changes this dynamic through its powerful user interface. Shift-clicking and good pathing make it easy for anyone beyond a certain skill level to attack in two or even three locations at once. It only takes a couple of actions to put a medivac with 8 marines in a corner of the map, and only a couple more to attack move it into an opponent’s hatchery. Thanks to strong AI, those marines will eventually destroy what’s in front of them until the Zerg player invests in static defense or moves units into position.

Now it’s important to acknowledge that this works both ways, because the Zerg player also doesn’t need to use very many actions to defend an unmanaged attack. This is where I agree with CatZ – excluding attention-asymmetric mechanics for a moment, the power of the Terran player comes not just from attacking in multiple locations at once, but also from actively managing those units. Furthermore, given the Terran’s role as the initiator, their ability to outperform mechanically is relatively more important. But my point is that this mechanical differentiation happens after the optimal strategic decision has been made; it doesn’t necessarily or even usually guide the optimal strategic decision.

Is this just semantics? I don’t think so, because it affects how the metagame develops. In particular, it places stronger incentives on “making the right choices” versus “executing well” – every real time strategy game will always have both, but the relative emphasis depends on the design. In this case, when those “right choices” boil down to a figured-out metagame, that tends not to change without developer intervention. As I noted in my previous piece:

One of the downsides of this approach is that if the game settles into a very one-dimensional state, it’s very difficult to move it elsewhere without developer intervention: players are unable to use their mechanics to improve their way out of the situation. Brood Lord / Infestor and 3-Rax Reaper in TvZ are good examples of this.

I want to make sure I present both sides here, so let me emphasize that this isn’t always the case. CatZ brings up an excellent example in the famous 2-1-1 stim timing:

I think that this is more of what Innovation (T) was referring to when he said “he only needs mechanics to win in SC2”. There have been periods in time where almost all Terrans just went for 2 medivac marine stim drop at 5 minutes… for MONTHS, Innovation included, ByuN included, all the cool kids. It was just the thing to do and yet still at various levels the same strategy would yield different results, not primarily based on the strategic merit of it, but rather the raw mechanics of their driver, the speed, the execution, the multitasking (and of course the smaller decisions that tend to accompany micro).

2-1-1 is not just a strategy with immense mechanical depth: it’s also a great example of the metagame solving itself without developer intervention. Terran players got better at it, Zerg players got better at countering it, and eventually it fell in popularity. But this doesn’t always happen. Consider 3-Rax Reaper, which required multiple balance patches to iron out. The necessity of frequent balance patches, design changes, and other factors I mentioned in my last write-up are good evidence that this kind of situation – players being unable to improve their way out of a bad situation thanks to diminishing returns on mechanical improvement, necessitating developer intervention – has happened consistently across StarCraft II’s lifecycle.

In other words, the removal of gates is more than just semantics – it changes how the metagame develops and forces the developers to be more involved, at the very least to keep things fresh.

From my perspective, neither 2-1-1 nor 3-Rax Reaper is a good representation of Terran’s overall design. The mine drop opener, popular until Legacy 4.0, does that much better – I’ll quote from my in-depth commentary:

We can already see several strategic choices, all potentially viable depending on the situation. The first widow mine can be greedily placed forward in order to catch the scouting Adept if the Terran player’s game sense and scouting indicate it’s not a Stargate opening. Alternatively, the Terran can skip either the second widow mine, build a cyclone, or put down a tech lab and build a tank if they suspect one-base aggression or a Pylon rush. The factory gets skipped and usually replaced by a barracks if they see a cannon or Zealot rush. The six marines don’t have to attack – the Terran can also do a double mine drop and use the marines to defend at home. The Starport can produce a Liberator after the medivac if there are no Phoenixes – if there are, probably another medivac, another add-on, or simply idle for the time being. The timing of the third CC, engineering bays and additional production buildings all vary depending on the game situation.

The strategic depth is pretty compelling, but it gets even better thanks to [the mine drop opener]’s mechanical depth. The six marines and one widow mine have huge damage dealing potential, but that’s all it is – potential. How much they get done depends on the player’s ability to control them in addition to their game sense and scouting of the Protoss base. Effective use of the Medivac boost both maximizes the surprise factor of the attack and employs it early enough to be usable upon exiting. The widow mine can be burrowed, unborrowed and targeted depending on what the Terran wants to attack and when. The marines should ideally fire in an arc, be pulled back if they’re damaged and, at the highest levels, dropped in and out of the medivac to dodge shots. Finally, all of this is highly positional: it needs to consider the placement of enemy Pylons and the simcity of the Protoss base.

In the worst case, the Terran makes all the wrong decisions and executes badly: they lose seven workers to an Oracle and get their Medivac killed by a Phoenix. The game is effectively lost. In the best case, the Protoss fails to clean up the drop for minutes on end, the macro behind is pristine and the game is effectively won.

What’s beautiful is that both the worst- and best-cases are rare, and really only happen if a player is severely outplayed or outplaying their opponent. In your average ladder game, the result is different every time: you do slightly better in one game, slightly worse in the next. You learn something about controlling marines, you tighten up your build. Your opponent makes a one in a million mistake and your mine kills 12 probes. And so on and so on.

The strategic and mechanical depth of this interaction are what make it fun and enable it to stand the test of time. Players can execute mine drop openers for years and years and still find them enjoyable, still learn a little something. It’s a good mix of fresh and familiar.

I include this quote for two reasons. One, I think it better captures the overall design and balance of Terran than 2-1-1. Two, it was one of the most popular, well-understood, and well-liked openers in the game – but the developers happily nerfed it to push the metagame in a different direction (in fairness, they didn’t target this opener specifically but the widow mine in general). That’s additional evidence, to me at least, of an emphasis on strategic diversity and evolution over the refinement of raw mechanics.

Don’t get me wrong – I agree to a certain extent that there has been a shift from one set of raw mechanics to another in StarCraft II. Partly I think this is inevitable, regardless of design intent – I discuss this in depth in my mechanics piece. I also agree that this is particularly pronounced in the case of Terran. But I think there’s more to it than that.

Design Intent and Trajectory

Next I’d like to discuss StarCraft II’s design philosophy and how it’s evolved since launch. CatZ makes this point later on in his post:

So in essence I agree that sc2 sees a shift in favor of strategy and a mental game over a physical one, I just don’t know that there is a glaring design intend to shift away from ‘raw’ mechanics, as much as there is just a newer cleaner engine where that naturally happens, from there, all races seem to have wildly different design philosophies rather than adhere themselves to the same strict principles pertaining mechanical constrains.

I don’t agree with this – I think the design shift here was intentional rather than something that naturally happened. First, I’ll talk about the move from StarCraft I to StarCraft II. Here’s a quote from Dustin Browder I reference pretty frequently:

Q: That leads us back to the balance between Micro and Macro. Since Blizzcon 2008, you have changed the economy system again. Back then there were already two vespene geysers in each base, but they would shut down for a short time after having collected a certain amount of gas. Therefore you would have to check the status of your source of income frequently, forcing a large amount of Micromanagement. Why this change?

A: Oh dear, we are thinking about how to modify the geysers since forever. We want you to have to manage your economy more. And the geysers would be a perfect start point, since they were quite unspectacular in the past: You sent three workers there, and that’s it. So we decided to change the mechanic, which hasn’t succeeded thus far. It was extremely hard to balance the new system. Had we decided to regulate the gas supply necessarily by hand, to collect the regular amount of resources, we would have severerly disadvantaged the newer players, since they couldn’t afford expensive units like Battle Cruisers and Templar. But just these units have the most appeal to casual players. Therefore, we would have to modify the mechanic in that way, that you still earn enough gas if you leave the geysers to themselves. But then, Micro experts would collect by far more resources and would produce only very mighty units like Carriers and Archons. That would also be unfair. In addition, the constant geyser-checking would become annoying very quickly. We want to reward the players, not annoy them.

When I read this quote, what I hear is “we originally wanted this interaction to be more mechanical, but it ended up conflicting with our underlying design philosophy, so we changed it”. This isn’t the only place you see this sentiment – for instance, decisions around pathing and AI evolved for years before the game was released (also confirmed via a Reddit AMA). Another example – Rob Pardo spent six months convincing the development team to put in unlimited unit selection. Here’s the full quote (thank you to Lalush for transcribing):

It made the team very timid, is what it did. I think it made it very hard to try to diverge too much from the original. Because it was so successful, especially in Korea it was this huge esport game. There was so much anticipation, especially in Asia for the game to come out, and the team didn’t want to let them down. We used to always call it: it’s like trying to make baseball 2.0.

I think what that did was it created a lot of fear. I feel like there was a lot of decisions that were very challenging to push through the team, I’ll give you a very micro one that I find kind of interesting:

So in both Starcraft and in Warcraft previously, when you group-selected units, you’d only get a limited number. So in Starcraft you could group-select 12 units max. So if you had 50 zerglings, you’d have to group-select them in smaller groups.

One thing I wanted to change in Starcraft 2 was to get rid of that restriction. My reasoning was that I felt like RTS games in the genre had evolved to a place where users just felt — it just felt like the UI was broken. People didn’t understand why it was there. I think people forgave it in earlier games.

That is one of those decisions, though, where the game is a little bit more tactical because you have that limitation. And you could also argue it even has game-balance ramifications in Zerg versus Protoss, because Protoss has less units and that sort of stuff.

But I didn’t really feel like it was that load-bearing. I was like: “You know, it’s going to be a little different. But I just think we put this in and we just design the game knowing now we have unlimited selection.” But there was all this fear that that was the secret sauce of Starcraft. “We can’t undo that”.

I literally… I was getting pulled into meetings over a period of six months before I finally got them to put in unlimited selection. I mean it was crazy. A little micro decision took months before it went in.

To me this is all evidence that the choices in the engine – better pathing, better AI, unlimited unit selection, etc – were made intentionally after a great deal of deliberation. You could maybe argue that Rob Pardo, at least from the above quote, came from a place of simply wanting to modernize the engine, but the sheer amount of pushback from other developers is evidence that it wasn’t that simple for the team as a whole.*

More evidence around design intent comes from how StarCraft II evolved after release. Let me pull out one of my favorite quotes from David Kim:

While we like to have more harassment options in the game, we don’t want to have harassment options like Hellbat drops that are a little too easy to execute. So for example if you see Marine drops executed by the highest level of Terran players versus a average level pro player, you can clearly see the difference, while Hellbat drops doesn’t look like that because it’s really easy to execute. That’s why we want to see less of it.

This approach, from 2013, differs substantially from today’s design philosophy. One of the points I highlighted in my interview with Thorin was the push in Legacy toward attention-asymmetrical mechanics. I’ll just quote directly from my follow-up piece:

[With regard to attention-asymmetric mechanics], with each expansion the designers introduced mechanics that require less attention than what’s required to defend them. By now, there are so many of these that even the best players in the world forget that they exist or get overwhelmed when they see them – burrowed banelings and shift-clicked liberators are two good examples. Rogue’s combination of Nydus Worms and Swarm Hosts against Neeb also come to mind, although this should be asterisked as a counter-build. The sheer number of game-changing plays in combination with how relatively easy they are to execute means that the game struggles to settle into something predictable.

Frequent balance updates, a continuously changing map pool, large design patches, an emphasis on attention-asymmetric interactions – these are all consistent from a design perspective in that they focus on “high-level strategy” over the fine-grained mechanics emphasized by David Kim five years prior. I’d further argue that they’re necessary – the underlying game design’s shift in emphasis toward strategy necessitates frequent developer intervention to keep the metagame fresh.

* WarCraft III’s development cycle was more straight-forward: the 3D engine was chosen first, and design decisions were made afterward. See the interview with Rob Pardo in this book for more information.

Unit Counters

Here CatZ and I take different interpretations of a quote from Dustin Browder. Again, I recommend you read his whole write-up; for the sake of brevity, I’ll pull out the key quotation:

I believe what Dustin Browder was trying to say is: unit counters alone don’t have as heavy an impact (compared to SC1) and micro plays a big role as well.

I actually agree with this interpretation. The portion of Mr. Browder’s quote that I care about is further down:

For example, let’s look at a situation where we have banelings fighting against marines. If they were to just clash against each other without any micro, the banelings will definitely kill off a lot of marines. However, if the marines have stim, I believe you can use micro to come out ahead in the engagement. Let’s look at another situation, where we have banshees against marines. In a straight up fight, the marines will definitely win the fight. Yet, if the banshee has cloak, the situation would be different. Let’s look at yet another situation, marauders against stalkers. If both sides a-move, clearly the marauder will win. However, if the stalker has blink, and uses blink well, the situation might turn out different as well.

I don’t think it’s an accident that all of his micromanagement examples relate to abilities, nor do I think it’s an accident that StarCraft II has increasingly emphasized ability-based micro as it evolved from Wings to Legacy to today. My point here was to highlight this as additional evidence for a shift in design philosophy toward abilities, but I didn’t do a good job in articulating this. I definitely concede that Dustin’s sentiment about the micro being qualitatively “better” in StarCraft II is not clear from the sourced quote. Per CatZ:

Seems to me that he felt that neither Micro nor Unit Countering were as determining in their own right hence alluding to a “better balance” between the two, he was never saying “micro is better than in StarCraft 1” unless we make those last 7 or so words stand on their own.

Maps

One area of curiosity for me was CatZ’s point here:

This is a big part of why, it is much more common for Zerg players for example to find themselves playing “reactive” styles than Terran players who are instead most often dictating the pace of the game. While every race can (and will inevitably) try to seize all 3 advantages for themselves (basically winning the game), bio terrans don’t usually care too much about the surprise element. Traditionally speaking the “Terran Meta” especially when it comes to bio is the most stagnant / varies the least in play-style.

I’d like to see this qualified against Legacy’s approach to the map pool – changing it regularly, experimenting with new map layouts and resource configurations, featuring wildly different maps within the same pool, and giving players minimal practice time before hosting a tournament on a new pool. This isn’t really a disagreement as it is a point of curiosity about how the “three fronts” perspective takes maps into account.

Final Thoughts

CatZ’s post gave me a ton to think about, and I’d like to apply more of the differential race design perspective to my upcoming work for StarCraft II. I really appreciate him taking the time to put it together, hence why I wanted to write a proper response rather than a quick throwaway on Reddit. Sadly my failure to prioritize meant my response was five months delayed – my apologies again.

If you’re interested in hearing more from me on this topic, I recommend reading through some of my older work (e.g. on mechanics, on BW/SC2, on balance).

Thanks for reading! If you enjoyed this post, I’d love it if you followed me on Twitter and Facebook and checked out my game-design focused YouTube and Twitch channels. All the best and see you next time.