This is the sixth entry of The Elementary Series. The topic is about the different stages of the game, and how they can be used as a rough guide to our strategy planning.
The target audience of The Elementary Series is mainly the beginners. There are exceptions to the things I discuss below, but I leave them out to ensure the target audience understand the main points.
Broadly, we can categorise different stages of the game into three parts: early, mid, and late. We often hear players and casters use these terms loosely to describe how the battle field may change as the game moves from one stage to another. But how exactly do we know where to draw the line to divide the categories? You probably will get different answers when you ask different people.
There is no “right” answer, but I will break this down in a way that is meaningful for practical implications.
Before I can convince you that my way is meaningful, we must first agree on one assumption. It is that both players enter the next stage together regardless of their races. That is, if player A is in the mid game, then player B has to be in the mid game too.
How to categorise them
First, we can all agree on one thing, which is the three stages are progressive in such a way that a later stage cannot occur without the earlier stage. Since it is a form of temporal information, one reasonable way is to use a precise time to divide the stages. For example, early is 0:00 to 6:00, mid is 6:01 to 10:00, and late is 10:01 and beyond. The obvious challenge is how we decide what the precise time is (e.g., why 6:00 and not 7:00), and it is difficult to have a consensus. Another way is to categorise based on the availability of units, tech, or upgrade. For example, it is late game when Zerg have access to Ultralisk. However, the racial asymmetry could be problematic as, for example, Terran could have access to high tech units quickly if they want to. Further, does it mean a specific stage does not occur if a player does not tech to that specific unit? Thus, while these different categorisation methods make intuitive sense, they are not practical.
In my opinion, in line with the assumption stated above, the most meaningful way is to categorise according to the relative goals of the two players. I will explain it using examples.
Applying in game via examples
Let’s use TvP as the example. Many TvP goes along something like this. Terran go for an expansion build (Reaper expand or Command Centre first), and Protoss go for a one Gateway expand. Then, Terran go for either a 1-1-1 or 2-1-1, while Protoss pick a tech path. A recurring pattern is the first major turning point in the game is when Terran have their first move out with the main army. The relative goals of both sides are different before and after this timing. Terran are generally defensive before they can commence their first attack, and their goal is to take minimum damage as they grow their army to maximise the potential of that planned attack. On the other hand, Protoss would have map control, and look for opportunity to deal damage. Terran would place defensive structures at tactical places (e.g., Missile Turret in mineral line), and position units accordingly to defend against Protoss. Thus, Protoss basically have the offender priority until the Terran push out, and hence, their goal is to maximise their benefits from the early game control. This is evident in the early development of the previously standard Stargate opening into a quick third Nexus.
The relative goals shift once Terran have gathered enough to make the first push out. This is usually triggered by the completion of Stim upgrade, which provides Terran a sudden power spike that Protoss cannot afford to face head on in the middle of the map under normal circumstances. Consequently, Terran’s role moves from being defensive to being offensive, and the two sides’ goals change accordingly. Terran would attempt to gain map control by getting active on the map, and the goal is to threaten Protoss’ third base while they establish their own. From Protoss’ perspective, they would position themselves to anticipate attacks at various locations, and their earlier decisions made are justified by how well they are prepared for this role transition. Therefore, due to the shift in the relative goals of Terran and Protoss, the first move out from Terran is the dividing line that categorises the early and mid game in the match up.
Builds are designed with this shift in relative goals in mind. For example, in INnoVation’s 6:00 timing attack build, he gets Viking and Missile Turrets before the move out. He does not get a Liberator for harassment, as that does not align with the goal. Stats’ Adept Phoenix build is another good example, because his power spike (i.e., Gateways, Resonating Glaives, and +1 attack upgrades) is timed to match with the then metagame push timing.
But what happens when the players use less standard builds? A short answer is that it simply changes the length and intensity of the early game. Let’s use proxy Barracks as an example. Unless the game ends with the attack, it would eventually move back to the point when Terran would gather enough to push out again with Stim. In the vod below, Maru proxy Barracks to Bunker rush Stats. Stats held the attack, and both sides transition into macro game. During this transition phase, Maru positioned himself defensively even though he had some units harassing the Protoss. Then, similar to almost every TvP, Maru moved out once he had Stim and Medivacs. Due to the unorthodox opening with proxy Barracks, the upgrade got delayed, and the push out timing got delayed as a result. Nevertheless, the overarching goals remain the same. Therefore, this shows that while builds could be less orthodox, there are still specific “triggers” in a match up that shift the relative goals of the players. This also explains why using a specific time (e.g., 6:00) is not practical in understanding the categorisation.
Then, what about mirror match ups? Although the goals of the two players are the same at various stages of the game, they still change. Let’s use Marine Tank in TvT as the example. In the early game, both players try to deal damage to the other player and not take damage themselves. Importantly, the real goal is to have an economy advantage heading into the mid game. Consistent with this goal, tech units are always favored in the early game, because high tech units are simply superior in early engagements. Cyclone and Banshee are better than Marine. The mid game goal is to be active on map to control important tactical locations, and use that to pile up damage (e.g., siege up at places that can deny mining). Then, what is the trigger that moves the game from early to mid game? The obvious one is the focus of unit production moves from those tech units to mainly Marine, Siege Tank, and Medivac.
It is important to note that this is conceptually distinct to the suggestion of using access to units as a dividing criteria. It is not uncommon for Terran to get an early Siege Tank for defence due to strategy preferences (read the comparison between TY and GuMiho), and Medivac is an option for many TvT builds. Though Siege Tank and Medivac are used in these examples, the game does not immediately move to mid game.
I timestamped the vod below when the players are transitioning from early to mid game.
Implications of the categorisation
The main implication of the categorisation is strategic planning. Most of us would go into a game with a prepared build order. Even if we do not have a precise build order, we should still have a framework. Since the notation of build orders generally end at a certain point in the early game, we do not have an as precise plan for the mid game and beyond (an example on reddit). Usually, in the mid game of each match up, there are fixed goals which are well understood by both sides.
Attacking versus defending in general
Following up with the TvP example earlier, Terran would try to find ways to do damage to Protoss, while Protoss would prevent that from happening. Players have their own interpretation of how to play the mid game based on this broad goal. For example, Maru’s way of dealing damage to Protoss in mid game is to constantly piling up pressure at different places using the mobility of the bio forces. On the other hand, TY uses small pokes to find a small crack in defence before he builds on that to force opponent to fall apart. Through different ways of “dealing damage”, Terran’s more tangible goal is to delay or take down the fourth Nexus. That means, based on the game plan, Terran should not be sitting back and worry about whether Protoss are trying to kill them. Conversely, Protoss should not try to walk their army out there in the middle of the map to try to attack, because we know Terran have an advantage in that area in the mid game.
With that being said, although such general understanding is important, it is also crucial to point out that there are situations we need to re-evaluate certain assumptions we have about the match up. A great example is Phoenix Adept in TvP. I was arguing in an article about the reason Terran players consistently got destroyed by Phoenix Adept composition in mid game is Terran players fixate their mid game plan to attacking only. The Phoenix Adept composition is meant to power spike in the mid game to shut down attacks. This is a little advance for beginners (i.e., target audience of the Elementary Series), but it is important to put it out here.
After you learn who is attacking and who is defending, it is easier to figure out the tactical details. For example, Zerg tend to be defensive against Terran in the mid game, so their tactical choices are designed around the fact that they can defend. To defend well, you need to know where the opponent is going to attack, so Zerg are actively deducing where Terran would attack in the mid game. Overlord placement and Creep spread are obviously key to getting such information. The other less obvious but common tactical move is to have a small mobile force (e.g., Zerglings) moving around the map. The units are not used to attack, but they are used to run around the edge of opponent’s base to deduce where the army is heading to. See how Elazer does it in the vod below.
He sent a few Zerglings up north to see how Terran react. If Terran do not have units of the expected size coming over to deal with the Zerglings, you could deduce that Terran might have moved their main army out. Moreover, since Zerg are defending, they always find opportunity to counter attack. In the vod above (continue watching from the timestamp), you can see Elazer kept some Zerglings at the north, and sent them in to harass Terran as they moved out. All these choices make sense because Zerg know it has reached mid game based on a certain trigger, and they know what their overall goal is. You do not see Terran do that as their goal is different.
Building and production choices
For beginners, a good implication of game stages is to know what is a good time to build static defence (e.g., Missle Turrets). Do not build static defence in the early game in general unless it is to counter something. Generally, whenever possible, it is always better to defend with units than with static defence. You rather spend the resource on an additional Stalker than on a Photon Cannon. Units are more value for money because you can use them for other purposes other than defending a fixed location. Thus, you always spend your resource in unit production in early game instead of static defence, unless it is absolutely necessary. For example, in TvT, you only build Missile Turret in the early game because you need it against Banshee.
Static defence become more valuable as you progress to mid game, and even more valuable as you progress to late game. It is very common to see players build a wall of static defence buildings in the late. There are multiple reasons. First, you are already max out on supply, so static defence is a good way to spend the resource. Second, players take many bases in the late game, and it is difficult to defend multiple locations at once. Static defence buildings fill this gap. Third, static defence buildings are cost effect in a prolong battle. You do not want to fight a wall of Photon Cannons and the Protoss army at the same time. Fourth, a wall of static defence buildings allow players to establish key tactical areas in the late game. Below vod is a good example.
Players generally have a core army composition in a certain match up for the different stages. Let’s use TvZ mech as an example. In the early game, Terran’s main goal is to expand and control the map. Hence, despite Siege Tank being the main unit for a mech army in general, it is not produced in the early game as it does not provide map control. Thus, consistent with the early game goal, Hellions and Banshees are usually the units produced in the early game (see vod below). Alternatively, Terran can go for Hellions and Cyclones, as these units allow Terran to be aggressive and control the map.
Hellion, Cyclone, and Banshee do not scale well as the game progresses (I’ll discuss more about this in the next Elementary Series article), and hence, Terran would have to move to a different composition. Again, the composition should be aligned with the mid game goal. Terran usually push out to kill Zerg with a +2/+2 timing using mech, so the units produced for this goal should do well in a big scale frontal battle. Thus, Siege Tanks are the main units produced. Hellbats are also made because they provide the meat shield for the Tanks, and they are also a good way to spend minerals for a gas heavy mech army. Vikings are produced to complement the Siege Tanks, which are vulnerable to Blinding Cloud. The vod below is a perfect example of how it looks like.
The goal in the late game is relatively hard to define. Every race more or less has the same goal in late game, and that is to (a) take as many bases as possible, (b) create the most supply efficient army composition, and (c) trade as cost effectively as possible with opponent’s expensive units. There is usually a core late composition, but players would switch to different tech to beat opponent’s composition in a rock-paper-scissors battle. For example, if Terran make many Siege tanks to defeat Zerg’s ground army, then Zerg would switch to Mutalisks. Subsequently, Terran would switch to Thors, and the cycle continues. Therefore, it is important to know what tech the other players is getting in the late game.
The early, mid, and late game stages are categorised based on the relative goals the two players have.
It is always a good idea to have a general game plan for a match up, whereby you have specific goals in the different stages. These goals have implications for every decision a player makes, and they provide structures for things like build order, production, and positioning. In fact, I have applied this logic to my design of the “one build for three match ups“.
The next Elementary Series article discusses unit composition and utility.
A shout out to Harry Hurley for being a Code-A supporter on Patreon. Thank you very much. It helps to sustain the site.
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. See you in the next article!