This is the second commentary article on the Faster is not better article I published recently. It is prepared by brownbear.
brownbear is a renowned writer in the StarCraft community. He produces articles, videos, and other content on computer game design. His focus is on esports, particularly StarCraft II.
Please note that I asked brownbear to write this commentary before I published my article.
I’m brownbear. Today I’ll be responding to an article from Max regarding Legacy of the Void’s shift from twelve to six workers. I highly recommend you give his article a read prior to reading this one.
On Lower Opportunity Cost Of Strategic Decisions
I agree with Max’s assessment of the consequences of doubling the starting workers – a shortened early game, a change to opening resource ratios, and faster worker saturation. However, I don’t agree with all of the identified byproducts. The biggest disagreement comes around this point regarding the opportunity cost of strategic decisions:
This opportunity cost in strategic decision making is nicely illustrated in the builds in HotS. If you build two proxy Barracks and continue Marine production, you cannot practically expand or tech. In contrast, building two proxy Barracks in LotV is an opening that applies early pressure and later transition to a standard macro set up comfortably. Using the two recent vods below as examples, Maru built two proxy Barracks to apply pressure with Marines, and at the same time he is putting down a Factory (and Starport; i.e., tech) and a Command Centre (i.e., economy) in his base. Hence, the additional workers in LotV allow players to get units, tech, and economy simultaneously. One may argue that the proxy two Barracks of HotS is more comparable to proxy four Barracks than proxy two Barracks of LotV. While this is correct, such argument shifts away from the difference in opportunity cost for build choices. There is no build in HotS to the best of my knowledge that allows players to get units, tech, and economy simultaneously.
I’d like to offer a different interpretation; Legacy places the same importance on decision making and ordering as Heart of the Swarm, but it does so on a significantly shortened timescale.
Optimized build orders call for players to build their production in a strict order. For example, in the classic Terran 1-1-1 expand into bio, the ordering of the third CC, two engineering bays, and additional two barracks determine how aggressive the Terran will play in the following two to three minutes. Deviating from the ordering dictated by your build is always going to be sub-optimal – if it weren’t, it would have been part of the build – so players should only stray from their plans in response to compelling scouting information.
A good example of this is Maru vs. TIME, a recent series from the last WCS Finals. In the first series (see vod below), Maru and TIME both leveraged aggressive playstyles, creating an opening for TIME’s mid-game focused style to find significant damage after holding Maru’s early attacks. In Game 3 this resulted in Maru losing his upgrades, arguably a game-ending moment in bio vs. bio.
By the end of the first series and throughout the decider series (see vod below), Maru adapted – he moved to more defensive builds with earlier upgrades, but with less early tech and no ability to attack. These modifications enabled him to hold TIME’s aggression more effectively. For example, in Game 2 of the decider series, TIME gambled with a medivac drop on the natural around the 7:00 mark. Had Maru followed his approach from the first series, he would have lost both of his upgrades again; but because he had stim this time, he was able to push in and save his second engineering bay.
Below I’ve mapped out the builds Maru used in both series. I’ve omitted Game 2 of the first series because it was not a macro game; I’ve also omitted the initial 1-1-1 opening, because it’s common across all games.
The timescale is short – in all cases, across the span of about three minutes, Maru gets all three of economy, tech, and units. However, just because the timescale is short does not imply that the ordering of his decisions doesn’t matter. The difference between researching stim at 4:30 and at 6:00 is the difference between winning and losing.
It’s worth noting that these are not the only trade-offs – Maru also had to give up map control and the ability to find early damage. Map control is important for obtaining scouting information and holding onto the tempo of a series, while the ability to find early damage is one of the key advantages an experienced tournament player like Maru has against a younger player like TIME. Aggressive “skill check” builds – the sort Maru employed in the first series – are often used by veterans to tilt newer players and reduce their confidence in winning. By favoring economy and tech over units, Maru traded away this potential advantage.
Consider some alternatives to Maru’s approach, which consistently builds a third CC, 2x barracks, and 2x engineering bays, in that order. This is the safest mid-game in TvT, enabling upgrades and economy with enough units to defend aggression. A greedier opening would prioritize the engineering bays before the barracks; the greediest opening would skip critical early tech, like the cyclone, in favor of slightly faster timings. A two-base timing would prioritize 2x barracks and 1x engineering bay followed by a third CC and an engineering bay + armory. A two-base all-in would skip the third CC in favor of two more barracks.
My point here is that Legacy still requires making careful trade-offs in designing and executing build orders. However, because trade-offs occur on a micro scale, they can be difficult to meaningfully perceive by most players, especially at the lower levels. An entire game’s worth of tech progression occurring in ten minutes is really freaking fast.
I think you can apply similar logic to the discussion of gas-first vs. barracks first openings as Terran. Here are a few compelling differences between the two:
- Barracks-first allows for an extra early marine, which helps with probe harass, pushes away overlords, and can be used to defend against early pools.
- Gas-first allows for a slightly faster factory, incentivizing aggressive play with hellions and/or liberators.
- Gas-first is good against certain aggressive builds like the ravager rush because the factory is out in time to defend; barracks first requires much more delicate micro around bunkers to buy time.
The difference between the two openings is small, but the implications are significant.
Build Order Variety
In the latter half of Max’s article, he discusses how Heart of the Swarm featured more strategic variation than Legacy does.
The key takeaway you get from the information is whether the opponent is doing anything “non-standard”. In contrast, in HotS, we went beyond that and paid attention to when the first geyser was taken. There was so much depth to whether the first Refinery was taken at 11, 12, 13, 14, 15, 16, or 18 supply. These are essentially different openings, and they are meaningfully distinctive enough that you can deduce the plausible builds. For example, in TvZ, a 12 Barracks 12 Refinery Reaper expand usually has two Reapers, and the ratio of mineral and gas makes it smoother to go for a Reactor Hellion and early third Command Centre with two more Barracks follow up. But a 11 Barracks 11 Refinery Reaper expand usually has three Reapers, and the resource ratio smoothly goes into a Reactor Hellion and early third Command Centre with defensive Banshee. Knowledgeable players would scout for these minor differences to further deduce what the opponent could do and planned accordingly.
I never played Heart of the Swarm, but I can imagine that it featured more strategic variety than Legacy. Strategic variety in real-time strategy games almost always decreases over time. Consider, for example, all of the following TvZ builds:
- 6 hellbat + 1 medivac + 3 marine all-in
- 6 hellbat + 1 medivac + 2 marauder all-in
- 2-1-1 stim timing into 3rd CC and macro
- 2-1-1 stim timing into 3-1-1 + tank all-in
- 6 helion + 1 lib into 3rd CC and macro
- 6 helion + 1 bc into 3rd CC and macro (viable with either mech or bio)
- 4 helion + medivac drop into 3rd CC and macro
- 4 helion + proxy starport + medivac drop into 2-base all-in
- Proxy 2 rax marines
- Proxy 4 rax marines
- Proxy 2 rax reapers
- 3 CC before 2nd refinery into macro
I’ve chosen these builds because at one point or another they were all in the pro-level meta, and in the right hands, they are viable up to and including the Grandmaster level. However, you no longer see most of them in top-level games, because Zerg players figured out how to scout and counter them effectively.
As players figure out a real-time strategy game, they cull it of its strategic variety. Openings and builds become more predictable because players realize they can cut corners and still defend various cheeses and aggressive plays. I believe that, over time, the same effect would have occurred in Heart of the Swarm. One of the design challenges in StarCraft II is its age: after almost ten years of competitive play, people have figured out how it works. I think that this occurred simultaneously with the shift to twelve starting workers, but I don’t think the two are related.
I do agree with Max that moving from six to twelve workers does reduce the total number of viable openings, simply because the time period of having six to eleven workers no longer exists. This is most notable in reducing the strength and variety of various early game cheeses (although, paradoxically, may have had the opposite effect in mirror matchups). However, I think the predictability of the mid- and late-game meta is something that would have naturally occurred over time.
On Moving Back to Six Workers
I agree with the suggestion to shift back to six workers, for two reasons:
- While I don’t think it will increase strategic variety, I think it will make decisions feel more meaningful because they occur on a longer timescale.
- Revamping the game on a regular basis helps keep it fresh.
Max’s article focuses on the tangible aspects to six and twelve starting workers, but to me a more important consideration is the experience they offer to the player. I think it’s reasonable to say that a real-time strategy game doesn’t just need to offer players meaningful choices; it needs to make those choices feel meaningful as players make them. Stretching out the decision making timeline ought to help with this.
With respect to the latter point regarding regular revamps, I have said several times that I do not agree with the developers’ decision to prioritize stability and make smaller and smaller changes to the game. To me, the numerous major design changes from the launch of Legacy to the end of 2017 are what enabled it to have a growing and successful competitive scene.
There are several aspects to six starting workers that make it an ideal vehicle for this year’s “game refresh”. One, it will help bring back players StarCraft has lost over the years by returning them to a familiar start to the game. This audience is the biggest untapped reservoir of new players; the majority of the increase in the playerbase after free-to-play was thanks to returning players, not brand new ones.
Two, it features a good balance of risk and reward with respect to game balance. There will likely be new openings, especially early game cheeses, that will require attention from the developers. However, the game should continue to feature the same overall strategic trajectory, just on a different timescale. This is safer than making major changes to existing units.
Three, it retains the qualities that make Legacy special. The changes that reduced deathballing and turtling, like reducing minerals and gas per base and the intentional inclusion of asymmetric interactions, will still exist. Reverting back to six workers allows us to preserve the core gameplay of Legacy while at the same time making its decision making feel more meaningful.
I had never seriously considered moving back to six workers, because it felt integral to the improvements Legacy made to StarCraft II. However, having read Max’s article and considered the idea for some time, I’ve decided it’s an excellent suggestion. From my very first interactions with the game, I observed that seemingly small mistakes – delaying a command center by thirty seconds or a minute – had major in-game consequences. There was a disconnect between how decisions felt and how they actually played out, something I always wrote off as just another design quirk. This idea makes me think that this problem can actually be effectively improved, without harming the other positive changes Legacy brought to the table.
Thanks for reading! I’d love it if you followed me on Twitter and Facebook and subscribed to my game design focused YouTube channel. One of these days I’ll stream regularly, so be sure to give me a follow on Twitch as well. All the best and see you next time.
This article sparked discussions in the community on r/StarCraft.
If you enjoyed this article, I’d love you to share it with one friend. You can follow me on Twitter and Facebook. If you really like my work, you can help to sustain the site by contributing via PayPal and Patreon. You can also support me and enjoy quality tea with a 15% discount at AFKTea by using the “TERRAN” code. See you in the next article!