TvP Concepts guide (Part 2 of 3(?))

Disclaimer: Concepts in this guide are a general guideline, not an absolute rule. They likely won't apply in all situations, or at all skill levels. There are likely some lies in this guide, or not entirely factual truths, or generalities that have exceptions not explained. If some things that I say don't feel like they apply to your rank, then you can ignore them safely. The useful information likely is in what you haven't thought about or are unsure about. The idea behind this guide is to get you to focus on what matters and to learn how to think about the matchup. You will need to use your brain. Individual results may vary.

Getting out of the Cave

Currently, most existing BW guides cover builds. The goal behind build guides is usually to get you some powerful army at a particular timing in a safe and stable manner; however they usually don't go into what you do with that army. Typically it'll end with something like "just go attack your opponent", but it doesn't say how. We followed the guide to the exact letter but we lost anyways!?! This guide will attempt to address some parts of that gap.

This guide will aim for the opposite, we will attempt to cover general engagement principles in the Terran vs Protoss matchup. When we talk about army movement, we talk about where your units are positioned, relative to the opponent. This guide will attempt to cover what you should be doing with your army, what the relative value of your bases vs Protoss' bases and how you should be maneuvering in each stage of the game. We will not be covering how to fight in this guide (though there is definitely some overlap), or consider if we should be moving out at all.

The hope is after you read this guide you gain some understanding about how to review your own replays and what you should consider when thinking about the game. As this is a very theoretical guide, it will take a lot of practice to put the concepts here into your own games. While TvP is a slower matchup, it can easily be overwhelming so it's important to both think inside and outside the game. It's good to treat methods and examples in this guide as a toolbox rather than a definitive rule that you should be following. There usually isn't one right answer in all situations. Every game is inevitably and invariably going to be different.

Additionally, I do not recommend using this guide until you have your build fundamentals down. If you can't make the required units to fight to begin with, there's no point in worrying about what to do with the little units you have; it's just better that you make more units first.

First, we will establish some facts about the matchup, then we will look at the win conditions, then we will look at some basic army movement principles. We first will look at the matchup from a ground army perspective, and then look at how elements may change when Carriers are involved.

Throughout this guide, I may use terminology like high unit count and low unit count. These should be relative to the current stage of the game. For example, in late game, low unit count might be 12 units, but in early game it can be just a single unit. Also, we'll loosely define early game to up until 7 minutes (1-4 factories, around 1 control group of units), middle game 7-12 minutes (5-6 factories, 3 control groups of units), late game as 12 minutes or later (7-9 factories, 4+ control groups of units).

Terran vs Protoss

If we were to imagine a classic TvP fight, what do we envision? Is it multiple skirmishes going on at once in several unrelated locations, like TvZ? Or maybe your whole army is spread across the map and we just use part of an army to fight like in TvT? Obviously, neither of these usually happen in TvP; we usually think of one big battle, where Terran is sitting in siege mode with all the support units around it while Protoss is shoving their entire army at Terran. And this will be generally true: when we fight in TvP/PvT, both Protoss and Terran wants to bring their entire army to the fight. Conversely, it's bad if you don't bring your whole army to a fight. Let's examine why this is the case.

1. Terran army scales extremely well with unit count. Every unit is a ranged unit, including the tank with 12 range.
2. Against a large Terran army, if Protoss engages with just a small portion of their army, it is likely to do little or no damage, resulting in a loss.
3. On the other hand, in low unit count fights, Protoss can trade relatively well.

So, because of points 1 and 3 it's not ideal for Terran to scatter their army like in TvZ, and because of point 3, defending all over the map isn't ideal like in TvT (although it can be done, we will discuss this later).

I clarify here that the above points made here apply if both sides wants to fight each other. However, Terran's army is slow (due to the tank), and since Protoss typically has map control, Protoss usually gets to decide if they want to take a fight or not. So while as Terran we almost always want to bring all our units to a fight, Protoss may decide not to fight at all! Finally, due to point 3 above, Protoss typically benefits from finding fights where unit count is low, so they actually benefit from trying to starting multiple fights around the map.

Now that we've established that both sides wants to bring most of their units to any given fight, then the questions that we want to address with army movement is:
- How do we get an advantaged fight vs Protoss?
- How do we prevent Protoss from taking an advantaged fight vs us?
- When we move out, where should we attack?

Win Conditions

The two basic win conditions that we should always think about in this matchup, for both races: 1. Prevent the opponent from mining enough so that they can no longer produce enough units to fight us
2. Kill the opponents army and a gain a position where they can no longer contest our army.
Usually the two win conditions are tied in some way, if we can kill their workers that usually leads to a unit advantage, which we can use to kill their army. Alternatively, we can kill their army and then proceed to remove their economy.

Here is some examples of win conditions for Terran: Early game:
- Kill a bunch of goons and camping their gateway production
- Kill a bunch of goons leading into several probe kills
- Run in/drop vultures and kill a lot of probes, leading into a deadly mid game push
Middle game
- Kill their entire army once, leading us to camping a lot of tanks outside their nat, allowing us to kill the rest of their bases
Late game
- Kill multiple nexuses, preventing Protoss from getting enough money to trade with our army efficiently
- Keep Protoss on a set a number of base, preventing them from expanding. Trade off armies to run them out of money
- Camp half the map and depend on efficient trades, betting on having an army while Protoss is unable to take more bases.

Here's some notable loss conditions beyond just kill army => kill worker (or reverse) that Protoss can use, so we should try to bear these in mind when playing:
- Protoss makes invisible units (dt/arb) and Terran has no method of detection
- Protoss camps out our production, preventing us from making any army units (Early goon shoves/Recall/mass drop/carrier)
- Protoss resets our tank count over and over while expanding, preventing us from ever taking efficient trades

When playing and reviewing replays, try to think about the relative values of your and your opponents assets:
- Army
- Main
- Nat
- Expos
- Controlling a segment of the map (Control of middle, or access to additional bases)

Finally, we should think about what happens in a base trade (main base for main base). This is the main mechanic we think about when we think about if a counter attack is possible. When we attack, we need to ask ourselves: **"Is it possible to base trade here? Does a base trade favor us or the protoss? If a base trade favors us, then we can attack and ignore the possibility of a counter attack. If it doesn't favor us then we need to be prepared for that possibility. ** In TvP, base trades usually will favor Protoss starting from when they get templar tech. This is because in base trades, comsat scanners are frequently killed off, so detection becomes an issue for Terran. In late game, it becomes increasingly hard for Terran to base trade, due to multiple gateway locations (+ arbiter and possibly carrier). In the early game, base trades are usually good for Terran, because goons have horrible dps vs SCVs whereas tank/vulture/marine swiftly kills probes. Additionally, Terran always has the possibility of lifting their buildings away to mine or produce elsewhere. Against carrier, it depends on how much of the Protoss economy you can kill and if you have enough anti-air. Bear in mind that these are general rules. I've experienced crazy sorts of game states where a base trade can be Terran favored, even late game.

A further category of base trade is if Terran and Protoss both kill an expansion. These usually depend on contents of the bases and what actually gets killed. In general, a single Terran bases are worth more to Terran than a single Protoss base is worth to Protoss; however, Terrans can also lift away their command center, preventing them from losing it and just suffering a temporary mining loss instead. Since a lot of the value of the expansion is in the workers, so if you can kill a nexus + workers for temporary mining loss, it's frequently a move Terran's can opt for.

Army movement resources

These are resources that aren't the traditional minerals/gas/supply resources, but are nonetheless resources we have for aiding our push or defense. These resources are very volatile and if you have an advantage in one of these resources it can be very brief, so it's important to pay attention when you have an advantage and utilize it when the opportunity arises.

All the resources are concepts that we should all have an intuitive understanding of.

A lot of the resources here are not specific to TvP, but rather can be applied to any matchup. The questions you should be asking are the same, just that the answers will be different.

Vision/Information

Imagine if you installed map hacks and could see anywhere ~~and anytime on a click of a button (d:^)~~ on the map you wanted to. Wouldn't it be overpowered to be able to see how exactly your opponent is positioned? That's exactly what you're allowing when you allow an observer to sit over your entire army. Protoss will know what you want to do, when you move out and how you're positioned. Meanwhile, we don't know where Protoss's units are positioned. How in the world do we expect to win a fight unless if we have an overwhelming unit advantage?

Now imagine if we could just teleport our entire army to outside the Protoss's natural while their army was away (But not their main, that'd be OP d:^)). We'd pretty much win the game on the spot right?

That's the power of vision. When Protoss has vision of our army, they know exactly where we are on the map, how we're positioned, so they can plan countermeasures. They can run at us when they know we're unsieged. But if you take their vision away, they don't know how many units we have, whether we're sieged or unsieged. All they see is a couple tanks at the front shelling away at their Nexus. Protoss can only run at you with their current army only once. If they're wrong, that's the game for them. Taking away their vision is akin to playing without comsat as Terran. It's immensely uncomfortable for them and a form of pressure.

Therefore, when able, kill observers whenever possible. This cannot be overstated and bears repeating. Since we usually need to expend scans to kill observer, there's an opportunity cost that we cannot use the same scan energy to look at our opponents tech or use them to detect invisible units. If we need to scan to fight invisible units or scan tech, we should prioritize that first. In late game, we can usually clear them walking a vessel and goliaths around. We can also deny vision by planting turrets (spaces where observers can't go)

On the other hand, it's equally important for Terran to figure out how Protoss is positioned. How many times have you lost getting ambushed by a Protoss army that was just sitting on a hill or came from a random angle? If we know how far our armies are apart, if we know if Protoss is ready to fight, then we can position ourselves with certainty. If we know where Protoss's army is, and they don't know where we are, then from their perspective, our army has teleported.

What tools do Terran have to see opponents army? We have scans, mines, and sending small groups of units. I will elaborate on usage later.

You can think of seeing the opponents army as the army movement equivalent of scouting your opponent's build in the early game. Sometimes it won't matter, but most of the time it absolutely will. Ignore doing it at your own risk.

Finally, a game mechanical note about vision. In Broodwar, unit vision updates at a set interval, usually a couple of seconds. This means that if a unit dies or leaves an area, you will still maintain vision over the area for a couple seconds. Practical applications :
- When moving your units, you naturally have more vision than if they just sit there
- When killing Protoss observer or scouting units, it can be good to wait 2-3 seconds before starting to move your units
- Patrolling your vessels can be good to stop recall to increase the amount of vision you see. (Another benefit is that when we go to EMP with that vessel, it might already be in full speed, avoiding acceleration)

Space

Let's say we want to take our third. How easy is it to move our army if Protoss is sitting at home compared to them sitting in an arc directly outside our base? Obviously, when they are sitting in an arc, we need to be more careful. They are closer to our base, so they pose a more direct threat. Conversely, if we are on their side of the map we are threatening to kill their bases.

This also affects how we can act when we lose or win a fight. If we win a fight on Protoss's side of the map, we can usually push a base directly because they don't have time to rebuild. If we win a fight on our side of the map, they have time to rebuild parts of their army. This same principle applies in reverse as well. If we lose a fight on Protoss's side, we have time to rebuild, but on our side of the map, we usually need to retreat to high ground.

For every race in Starcraft, the less space you have, the less choice you have when declining a fight. If you back Protoss against a wall, they have no choice but to fight you.

Positioning

Space is just about how much distance we can walk (distance from point a to point b). While related to space, positioning is about where you are on the map (x,y coordinate).

We want our units to be positioned in a place where they protect something of value or are threatening something of value.
You might think this is some blatantly obvious statement, but when's the last time you've defended an already dead base or we watched Protoss walk to attack a base and we're way late on defending it? Or we watched Protoss use their entire army to chase two vultures all the way to an empty corner while you just watch it happen and sat there?

When we review our replays, when we move to the middle of the map, we should ask ourselves "What am I threatening by being here?"
When in game, we should ask ourselves "How fast can Protoss engage my army if they're positioned here? How fast can Protoss attack me if they're in this part of the map?". If the answer is that they're in some relatively worthless spot and takes a long time to walk back across the map, then we should take it as a free invitation to walk across the map as they're not in a position to engage us. If they are relatively close to us, then we need to be careful in controlling our army.

Map control

The main way Terran takes map control is primarily laying mines on the map. When we lay mines somewhere on the map, we're basically saying to the Protoss: "Hey, here's a chore. You can spend time and army to clear the mines (giving Terran information on unit positioning), or risk losing units to them." When Protoss clears the mines, they also move slower than normal; they need to make sure zealots don't run ahead (and dying to the mines) and dragoons need to pause each time to attack the mines. It is a way of restricting Protosses movement and also to buy time in certain cases.

Traditionally, we've thought TvP as a matchup where Protoss has control of the map. These days though, I'm not sure that it's entirely true. Terran can also deny protoss movement by sieging tanks in the middle of the map and force Protoss to walk around or deny movements to certain parts of the map. Terrans have realized that just sitting in the middle of the map with tanks and laying mines everywhere is an incredibly powerful concept. In a sense, this isn't a push to target the any particular Protoss asset, but rather just a push to gain more map control. After we're happy with the map control we have, we can try to force a fight where Protoss has nowhere to go but fight us.

If we decide to sit in the middle, we should be ready to fight at any time though, so be careful about trading off too many vultures for probes and be sure to reinforce periodically if we aren't maxed.

One other very important form of map control is controlling how the units coming out of production buildings can move. If Protoss just sits in an arc outside of Terran's base, it's hard for vultures to get out. If Terran camps outside of Protoss's natural, then they can't access the rest of the map. If there's already army out on the map, it can't get bigger. Even if the force can't directly threaten anything, it's a huge threat generator. Cutting reinforcements off is a form of time pressure, which we will talk about next.

Time

Fights don't happen in a vacuum like in a micro map where we start with a set amount of units. We're constantly making more units, getting upgrades, expanding our economy. Therefore, we also need to consider how we're currently scaling compared to our opponent. If the scaling is relatively even on both sides, then there isn't a immediate need to rush to fight. But if Protoss is scaling faster than us then we are under pressure to do something relatively quickly. Anytime we are pushing or defending, we should ask ourselves: "If we do nothing and waste time, will the game state get better for me or is it better for Protoss?"

What situations do Protoss scale faster than us?
- If it's an early game push and Protoss is waiting for additional gateways to kick in or reaver to arrive.
- If it's a 5-6 factory push and Protoss is waiting for additional gateways to kick in or legs to complete
- If it's mid or late game and Protoss is waiting for additional Carriers

What situations do we scale faster than Protoss?
- When we have a scary contain up
- When Protoss is maxed but we are not maxed and Protoss is not setting up additional bases (Even if they are setting up additional bases, we're scaling in the short term).
- When we have a significant production lead and are not maxed
- We are slowly shelling at their base

This is rather simplistic; these are just some common examples. Actual in game situations will have many answers, so it's something every good player should be thinking about when looking at their own games.
How we scale is usually determined by our build and Protoss's build, so this requires a good understanding of build interactions. It won't be covered in this guide (and hard to generalize).

Army

A lot of Terran players tend to rush fights, as we feel that we are constantly under time pressure. The feeling is that if we don't attack, Protoss army will grow too big, they'll take too many expansions and the game will become unmanageable.

The reality is that it is usually far better to keep your army alive than it is to rush into a bad fight. You can always decline to fight right now and fight later. While we usually don't talk about your army as a resource, it is most definitely a resource. After all, how are you supposed to fight without an army? I like to think of my as a resource which gets spent by fighting. In the same way that it's absurd to spend all your minerals on 20 ebays, it's absurd to just use it by move commanding up a hill straight into the Protoss's army.

While Protoss taking another base is a form of time pressure, it is overvalued by Terrans at low rank. We see Protoss on 5-6 base and start panicking. But Protoss still needs to saturate each base and scale out their production before any of those bases start to pay off for them. If we just take a good fight (eliminate their army) then none of those things matter.

We can also trade army for other resources. We can sacrifice a vulture to find the Protoss army. If we need to slow Protoss down, we can throw our army in between what they intend to attack. We can push across the map extremely fast at the cost of our army.

Attention

Because our army is so important, we need to take great care of it. We shouldn't be looking away from our army when we can be engaged upon or when we're in danger in taking a bad trade. How can we possibly justify making an additional two SCVs if 10 tanks get stormed when we're looking away? The point of our build is to build our army to kill and defend, not we build army for the sake of building army.

Luckily as Terran we have a decent enough solution. If our tanks are sieged (not on mines) with our entire army supporting it while maintaining vision of anything running in (think: arbiter), then it's unlikely that we will trade poorly. Then we can look away and do our management tasks (When we are 200 supply, we shouldn't even do any management tasks, just focus entirely on our army). When we do our management tasks, try to do everything we need to do and more: overbuild depots, queue multiple units. Then we should add units from our rally. This is an easy set of steps. It won't guarantee that something bad won't happen. It's possible Protoss can still win a fight, but it'll at least they'll trade into a scary sieged Terran army. Since your army is also planted in place, so it also could be easier for Protoss to attack somewhere else. But we should know where the Protoss army is....right?

On the flip side, if we are looking for a fight, Protoss also needs to spend attention on their army. So it's also difficult for them to do macro chores when we are on the map. They may bank minerals, but it should be hard for them to add many gateways. If we can catch Protoss looking elsewhere, we should take advantage of it by jumping them or moving quickly into a lethal position.

Army movement fundamentals

Now that we've established what the basic concepts of army movements are, let's just try to tie everything together. As another reminder, a lot of the principles here assume we're in some sort of neutral situation, that it's possible to attack or defend, and not some already critically lost situation. If we get caught up in a hopelessly lost fight, it's likely we didn't execute properly, or we don't understand how our build or Protoss's build works.

How do we get an advantaged fight vs Protoss? (Offensive army movement)

First, to even be able to fight as Terran, our army needs to be positioned to be able to fight. As established earlier, this means that Terran's army is completely together. Additionally, our unit needs to be sufficient to fight. While obvious, we shouldn't push if we have no army to push with. If we've traded or taken damage earlier, we should look at our unit count. If it's too low or our composition is broken, think about it if it's worth it to attack or if we should try to gain an advantage in other way (map control/expo) or even consider just sitting in the middle.

If "making workers" is the golden rule for thinking about a build, then "fight with your entire army together" is the golden rule for army movement in TvP. If you do this, you shouldn't get run over. You still might lose the fight, you might get pulled apart by recall/counters/carriers, but you won't get run over.

Similarly, Protoss needs their entire army to fight Terran. If Protoss is not ready to fight but Terran is and both armies are relatively close to each other, then Terran can initiate a fight immediately, either picking off units for free and having a army advantage or Protoss will fight by trickling in units, which is horrible for Protoss.

Why would Protoss not be ready for a fight? It is a combination of the following:
- They don't have vision of you
- They are out of position because you have pulled them out of position
- They are defending vs vulture threats
(They might've also just not expected you to attack but that's not something to be relied upon d:^))

Therefore, to get an advantaged fights we must:
1a. Erase Protoss's vision
1b. Pull their units out of position by forcing a reaction
3. Find their army
4. Fight

It turns out that throwing sand in their eyes and telling them to look behind them while punching them in the head is pretty good!

Note here that we can erase Protoss vision to help pull units out of position, or pull units out of position and erase vision; both are good. The other benefit of pulling units out of position while pushing is you generally force a choice onto them: Deal with a small threat now and risk losing to a big push or fight the big push and risk losing a significant something (usually econ).
Pulling their units out of position usually does cost time and possibly units. In early game, one unit is a significant chunk of your army so this may not be worth it, and in middle game, you usually are under heavy time pressure, so it also may not be worth it. This is all to say it's game state dependent and it's possible to skip steps if the situation calls for it. If we know that Protoss has no units at all, there's no point in taking strange paths and the value of killing their vision is low.

At this point, I'm sure you're asking, "How do we pull units out of position?" but instead of going over that, let's just go over some detailed examples of the above principles in different stages in the game. We'll start by working backwards from late to early game, as I think it's easier to conceptualize.

Late Game - 200/200 push

Vision: Kill vision by using goliath + vessel, or goliath + scan. To find Protoss's army, we should use a combination of scanners, mines, and running vultures around. Anywhere mines are, we can rule out Protoss's army being there. If we don't know have any map vision at all, then we should first start scanning the areas closest to us. Check middle, and then further out on the map. If we still can't find it after 4 or so scans we can at least know that their army is nowhere near us and we can start moving anyways. We can always send one vulture ahead to find the army.

Outside of battling, scans should be used to check for tech switches, confirming army movement and checking bases vultures can't check (a main after P has taken a nat). It's usually a waste to check bases vultures can check, as vultures do this very easily.

How do we know Protoss is ready to fight us?
Protoss needs to have their arbs/goons/zealots/shuttles all together. The key unit here is zealots. Protoss usually wants to have goons sitting around the map for map control, but zealots are not very good at chasing vultures and are prone to running into mines.

Very frequently we will see goons clearing mines and we can't tell if zealots with them. In this case, it's important to scan behind the goons to see if the zealots are there.

Pulling units out of position:
We can run around 8-12 vultures, or we can try dropping tank + 2 vulture or tank tank (and optionally support this drop with additional vultures). One alternate late game tactic we can do is to drop a killer EMP on their army/arbs.
When running vultures around, we should get down mines wherever we don't see Protoss units. Here are some possible targets:
- We should be able to raid a Protoss base, looking for a base with 2 cannons or less. Before we go in, it can be good to lay mines in front of the base but out of vision so that we can stop panic reinforcements with ease. Although we usually think of the main/nat of the Protoss as well defended because that's where their rally is, sometimes it's a viable target when it feels like their spread out too well.
- If all the bases are too well protected, we should try to see if we can fight lone high value targets (small groups of goons/ht).
- Otherwise, just try to see how Protoss is positioned and lay mines. Remember, Protoss needs to spread across the entire map to stop vultures, and if they're spread out, they're not ready to fight you. Even just threatening a push by inching out a bit can be good to buy more breathing room.
- Frequently, you can use one vulture group to buy space for another vulture group. If Protoss tries to corner one group too hard the other group should have a lot of space to plant mines.
- In early late game, if Protoss takes a base, but has not started mining from them, you can just leave 4 vultures or so on the transfer path. Protoss will need to be diligent about escorting their probes (and when they do so, they are usually not ready to fight)

If we can't pull them out of position, getting rid of their vision and finding their army is also ok. It's a waste of time and units if we just keep losing vulture for not much gain. In general Protoss has to find a balance between splitting their goons to defend vs vulture move out and keeping their vultures together. If they're too defensive, then we already get the breathing space we want from just walking across the map.

Middle game - 5-6 factory push

Vision: Your only method of detection at this stage is scans. Therefore, we need to be very careful with what we do with them. Your first scans usually should go towards checking the opponents tech and gateway count. If you're pushing regardless of what their doing, it's better to get rid of opponents vision. We should be expending scans around where observers are most likely to be. These locations are : right outside our natural, our rally point, and on the direct path between the two bases. Also, if you see shuttle/reaver, associate it with an observer nearby as reaver/shuttle have poor vision. If there is heavy threat of dt/arb, make sure you bring scv for building turrets and save a couple of scans.

To check for protoss army, your vultures which also act as pulling them out of position should be scouting how they're positioned.
Protoss army is ready as long as it looks like the majority of their forces are together.

Pulling units out of position:
Here, we should not expect to be able to lure Protoss into chasing our units way into their own base or into some random corner of the map. Rather, we want to be able to force a situation either Protoss has insufficient units forward to deal with our initial push, where they run the risk of getting run over because their units are split, or they are pushed back into their base where we can just walk up for free. It's relatively bad for Terran to try to crawl up the map from their own base as it's slow; we want to be able to get as much ground as possible. Remember, in this scenario time is not on our side, Protoss generally is expanding gateway count, getting legs, getting arb/carrier/storm/reaver.

We want to run around 4 vultures. Here's some possible targets:
1. See if we can run into the third or nat if goons are pushed too far forward.
2. Aim to lay mines in the direct ground path between your two bases. You threaten two things: If they have forces in your face, then when they retreat they will run into mines or when they reinforce the forward forces, the reinforcements run into mines.
3. If you can't find anything, laying mines in the middle of the map can be good to protect from flanks and counterattacks.
4. Camp their 4th, to force them to use goons to escort a probe, though usually it's better just lay a couple of mines for vision.
5. If we're fighting goons only, we can threaten to run at them with all tanks/vultures together, planting mines behind the goons

We can also consider taking a weird ground path to their base if we kill their vision. For example, on Radeon vertical spawns, run through your third and into their third. Or on eclipse, going through the very bottom low ground route.

Fighting:
We want to walk as much as we can for free, then when we think Protoss are capable of starting to engage us, we want to slowly crawl tanks forward, leaving at least 3 tanks sieged at all times. Make sure to not miss macro cycles and reinforce. Once we reach outside the natural (or even places like between the nat and 3rd on Radeon, or the high ground leading into the low ground on deja vu) we should consider just sitting there and waiting for Protoss to ram into us.

Early game - 1/3/5 tank pushes

Early game is a bit different than the other two stages of the game. First, in terms of vision, we don't have scans, but Protoss also don't have observer. This means that we need to be using units to scout. Here, when using vultures to scout, we also have to be careful not to take damage on them (or even worse lose them), since doing so will directly affect the strength of our push. On the flip side, Protoss also needs to rely on goons to gather information, and they don't really doesn't want to take hull damage on their units, since it's non recoverable. If it's possible, we should also try to hide if we're building tanks or vultures.

One important thing about vision in this phase of the game is that for 3-5 tank pushes where goon count gets above 4 (enough to two shot tank), we need to take care to not get ambushed from high ground. It's often a good idea to send an SCV ahead of your forces to check or tank the first hit.

In the early game, it's also difficult for us to pull Protoss out of position or distract them, since before their third goes up the only place of value is the natural. However, Protoss typically wants to position to take their third so you might find goons spread out more than normal, in which case we should try to pounce on them.

For the most part, early game pushes are a micro battle, so I will go over this in a later guide.

How do we prevent Protoss from taking an advantaged fight vs us? (Defensive army movement)

Defensively, we care most about vision and how we're positioned. If we can see attack coming, then we can position ourselves correctly to absorb the attack. Of course, if we don't have the units to defend to begin with, none of this matters.

In this section we will be purely considering pure defense.

Vision
Similarly to how we want to take away Protoss's vision before we attack, Protoss wants to take away our own vision before they attack us. While they can't really do anything about our scanners, Protoss can clear out our mines before they attack. On most maps, it's not possible for us to defend four or even three bases at once, so once we lose vision of an incoming attack we'll be late to rotate once Protoss's force is around attacking. Therefore, we should be placing mines on possible attack paths to expansions whenever possible so we can be prepared to move once they attack us.

Additionally, mines on the ground paths that they can attack from are all useful. If we place a bunch of mines halfway on the map and they haven't cleared them, then we can safely ignore any possibility of a frontal bust at the current moment. We can then spend our army resources on trying to harass or shore up defenses elsewhere.

If Protoss is clearing our mines, we also don't know what's behind it. It can be good to scan behind the goons to see what their composition to determine if they're just merely clearing mines or if they are planning to attack us.

If we're sure that it's possible for them to attack us, then we should make sure that all our tanks are sieged and spread out and gear up for the incoming attack.

If possible, try to replace the mines that were cleared after the goons leave.

Another source of vision is floating buildings. This is particularly useful when defending drops or arbiter recall. We typically want to float either the rax or ebay above our natural and another building to the air path leading into our main. We can move out the buildings later when we go and take the third. Even when we can't stop a recall or drop, if we can see it, we can take action against it before it hits us (Transfer workers ahead of time, start moving your army as soon as you see it, etc)

Finally, we should take away vision from Protoss by killing off their observers. The less they can see how we're set up, the better. Common spots they will have observers are over our army, around our army (usually in front of it) and in our base. Against drops/recall, killing the observer in our base is useful since it's easy to have a weak spot here or there even if we mine it up.

Build a wall, and make Protoss pay for it
When we go out and crawl to take a new base or position, that is when we are most vulnerable as we don't have the defensive tools in place yet. We want to make sure that we safely secure our expansions as much as reasonably possible. This means we want to wall off attack paths whenever we can, lay mines in front of tanks, lay mines around our buildings. At expos, we can also leave 1-2 siege tanks for defense. If we build a wall with nothing behind it, there's no pressure on Protoss to not attack our wall without a tank. We can't control where Protoss attacks us, but we sure can make it unpalatable for Protoss to attack us. In other words, do your chores.

Sometimes we get into situations where we need to prioritize where to lay mines. Usually, the order you want to do this is first protect vs frontal bust, then we lay mines in base to protect vs recall (as recall takes more time than a stasis bust), and then finally at our nat and third. This should also give you an idea of what the most valuable locations are for Terran (Let's remember our loss conditions here). If we lose our army in a bad fight, then we're vulnerable to either getting overrun (and having production camped) or we'll be put on a treadmill where our tank count gets reset over and over. If we get recalled on our production, what we risk is losing some production, perhaps workers + cc, key upgrade buildings (armory/sci facility/starport). However, we still have our army alive (and likely nearby) and reinforcing units spawning from our factories. If they recall their army on nat, then we lose mining at the natural, possibly cc, and our reinforcements are cut off. If they recall onto an expo, then we lose mining there.

What to do when Protoss attacks us somewhere where we have no units
Let's say we get recalled/dropped/attacked when we have no vision. What should we do?
1. Look at how many units they're attacking with, and what the composition of units are.
2. Determine:
- If we need to run worker away/lift CC
- If we should be defending at all
- If we should defend, how many units we should send

The composition of the army that attacks matters because we want to see if it's goon heavy, zealot heavy or a balanced mix. Zealot heavy armies cannot kill CCs but kill worker lines pretty easily. So against these, it's important to run away worker and lift CC. It's also "easier" to defend zealot only recalls as you can send vultures, but be careful with positioning; if you don't micro, it's possible they can still get a good trade vs vultures, especially around the ramp. Goon heavy armies kill buildings pretty well, but are also a bit more risky/committal for Protoss (If it's part of a recall, losing goons for free is fairly costly).
Finally, we should touch on one property of Protoss army: Aside from reaver, Protoss doesn't kill buildings very easily. If we trade poorly rushing in with vultures, it's usually better to wait for more units or when our units are in a better formation/composition. I plan on going into this into a bit more detailed in a future guide.

If they attacked us at a 3rd/4th/5th/etc, we should consider if it's worth temporarily losing mining from it and instead counterattacking them. If it's a significant force, then Protoss shouldn't have much to contest us running across the map either. Note that this is only possible because Terran can lift their buildings, a luxury that other races don't have. Just have a plan for after they cleanup your base. With the force that attacked us, their options are to continue attacking us (countering us), keep some units there, or pull their units back.

If they attacked us at our main then we usually have to go and defend, unless if we can defend with units producing from our factories. This is because if our production gets killed, unless if we can pick up massive compensation, we generally will get bleed to death from lack of units.

We should also pay attention to how many units they're attacking with. If they're attacking with a low unit count, we should not over-defend as they may want to hit us with more units elsewhere (multi recalls, chain recalls, drop in main into bust, etc). My opinion is that usually when you get dropped/recalled, bringing up to 1-1.5x as many units is usually sufficient (as you are the attacker, sometimes up a hill into a chokepoint). Conversely, if they attack with everything, we should defend with everything (given that we are choosing to defend).

Where do we attack?

Up until now, we've gone over how to attack, but we haven't touched upon where to. Even if we win every fight, if we proceed to run around like a headless chicken and can't capitalize on good positions, it's very possible we still end up losing the game. Here, a lot of how we act is down to personal playstyle and the actual game state that you're in. I'll try to discuss merits and demerits of taking certain actions, but don't just take my word for it, play the game, and then think about what feels easier or harder.

In the late game, we usually have multiple targets of value:
1. Protoss's nat/main
2. Protoss's army
3. Protoss's side expansions (little or no gateway possible)
4. Another nat/main (possibly with multiple gateways)
5. Securing an area of the map

In early game there's only 1 and 2, and in middle game there will be one, possible two expansions.

First, let's remember that the Protoss army itself has value. Killing it and winning hard allows us to walk up to whatever target we want to. We also get the obvious benefit that Protoss needs to spend additional money on their army. We also need to spend money to replace our army as well. However, depending on where the fight takes place, and how many units we have left over, we may or may not be able to follow up on the attack and we may not be able to have the time necessary to get to the target we want. On the plus side, if we fought all their army, you shouldn't have to worry about the possibility of a counter attack. Just make sure you fight well. After we win a fight, the first question that should come to mind is what should we convert the win into? We can take a base, or we can try killing a base or setting a contain, or possibly both. At worst, we should be able to find better positioning for our army.

In terms of targets, up until late game, there is the main/nat, and then there's everything else. The main contains the Protoss's production buildings, meaning that they need to walk out of their nat to access the map. This means that even if we just contain them outside of the nat and we don't hit anything, there is value in that portion of the map. By attacking the main/nat, we cut off Protoss's reinforcements. This is important enough we consider it as leading to a win condition for most of the game. Additionally, there is usually an added benefit in that Protoss needs to spend attention to re-rally their gateways. Finally, if we want to kill all the production buildings, containing their base is the first step towards that. However, going into the main can be quite difficult, though sometimes we're forced into it.

There's a couple situations where setting up this contain can be less valuable:
1. When they set up a second source of gateway production. This diminishes the value of setting up a contain because they can send units out in a separate place.
2. When they have carriers.
3. Similarly to a second gateway farm, when Protoss has access to recall or lots of shuttles. However, recall is expensive to tech to and expensive to use (150 mana + arbiter), and shuttles take a lot of time to build in large numbers. You should usually feel pretty good if you get a contain up in this situation.
4. When the nat is mined out.

Points 3 and 4 are usually minor reasons not to push the nat, but points 1 and 2 can be major reasons why you wouldn't want to.

When we attack a Nexus with probes mining from it, we are attacking their economy. How much damage it actually does depends on: 1. How many probes we kill.
2. How many mining bases Protoss currently has
3. How rich Protoss is.
If we kill all the probes, killing a mining base is quite damaging. 15 probes is usually 20-25% of the Protoss economy. If they're rich or have a lot of Nexus, there's less relative value as they can remake probes quickly. However, if they don't have much mining or are poor, then it's more relative value, as they're forced to choose between economy and production. If we kill a Nexus with no probes that we kill, they lose potential mining time from there. It's good if we can get it for free, but if we take damage in return, it usually won't feel good aside from low eco situations.

In longer games, as the number of gateway farms go up, and as bases get mined out, the value of pushing the main/nat or the other main tends go down while killing expansions tends to go up. So even if Protoss has a lot of healthy bases with a huge bank, don't discount the value of killing an expansion. We have to start somewhere to destroying their economy. Additionally, sometimes we don't have the army and/or time to go set up a contain, but we have enough army in position to go kill an expansion, so we might as well do it.

After we kill an expansion, your options are to leave the army there and hope you get whatever value you can, go attack somewhere else, or consolidate with the rest of your army. If Protoss can contest us at all, and we can do it, I think consolidating your army is usually a solid move. After this, we reset back to the beginning, but with (hopefully) more of an advantage.

In a lot of late game pushes, how we push also affects where we want to take our 4th/5th. We generally want to push along the same path that we're expanding along. For example, if we spawned 7 o'clock on polypoid and Protoss at 5 o'clock, if we push directly to their base, then 6 o'clock is a nice expansion to expand to since our push protects it. Compare this if we want to take the 11 o'clock main/nat/min only. On the flip side, if we choose to push cross map, we can safely take one of those bases, and it would set up nicely if we also needed future expansions.

We can also push to secure an area of the map. This is especially important when we are aiming to split the map, or if we're on the verge of mining out. Of course, we can also yolo push somewhere and defend by attacking, but sometimes we don't have the firepower to do so. Even if we do have the firepower, think about the cost/benefit. If we lose our push, then we just straight up lose the game. If we win our push...what would we get off of it? It may just be better to secure another base, mine up more of the map, and then when we're stabilized, we can consider pushing. As a bonus, Protoss may come and fight you anyways, but it should at least be a defensive fight.

The last two factors when deciding where to push is we spawned close or cross, and whether the map is 2 player. If the map is 2 player, usually there will be only place to place production buildings, making pushing the main/nat almost always viable.

Cross spawn is has some notable differences compared to close spawn:
1. Rush distance is longer, giving the defender more space to work with.
2. In close spawn, the direct ground route is also the shortest path, but in cross spawn, there usually are multiple paths of roughly equal length. This matters a lot for counter attacks. In close spawn, Protoss usually need to walk along a longer path in order to counter attack Terran, but in cross spawn, they can just position themselves along the path Terran isn't in.
3. In late game split map scenarios, it is difficult to attack and defend along the same path. Instead, you're usually forced to choose between attacking Protoss or protecting your 5th/6th.

If we assume we're attacking a location (including the main/nat!) when we didn't get rid of Protoss's army, Protoss has 3 options. If they sit and do nothing they will lose something valuable and we'll be happy. In any case, we should be tracking their army carefully.
- They will position to defend themselves
- They will counter attack us
- They will threaten to counter us and cut off our reinforcements

If they are committed to defense, we need to be ready to fight. If we are there first, then fighting should favor us as we're no longer attacking into the Protoss.

Dealing with counters

To begin with, to deal with counter we want to make sure we do our chores (lay mines in our main, wall off our expansions, build turrets, lay mines on counter attack paths). We do these chores ahead of time because we want to invest our APM when nothing is going on in the game so that we don't get overtaxed when the game state turns chaotic and crazy. We don't really want to need to focus on micro on 3 places at once when we get double recalled and we're attacking at the same time. This mostly applies to late game, but even in the mid game it's useful to lay mines along the flank paths to both protect against them running into us and have mines for counter.

We can't really get countered if we just run directly at their army. Dragoons don't walk that much faster than tanks, so if Protoss is already on the ground path we're pushing on, it's difficult for them to just side step us (short of recall) if we're already on top of them. We should also be ready for if they run at us. Of course, if they start running to some random corner we might not want to chase them because we don't have time to. We also can't get countered if we're just sitting there defensively d:^).

In the early game, we might push some dragoons away with a tank push when Protoss's goons are split and we can lay some mines to dissuade them from joining the fight.
In late game, I think the concept is pretty similar. If we can lock them into a corner, we can lay a minefield, put 2-3 tanks behind it, and Protoss will have a lot of trouble trading efficiently into it.
In the mid game where their army is goon heavy, we can wall them off with mines (and maybe a tank). We also should consider hunting the goons down by running vultures behind their goons and laying mines. Even if we spend a minute doing this, if we kill a full group of goons, they need to spend a whole minute remaking it from 6 goons, and their composition is temporarily broken.

Another common situation in the late game is that we go out onto the map, and Protoss is far away from our army, and then starts making a beeline to our base. In 200 v 200 situations, we probably brought most or all of our units to fight, so we're unlikely to be able to defend with whatever we have at home. However, usually base trade does not favor us at this stage, so what should we do? We need to turn back with our army. Now, there's a couple of ways to extract value of the situation. Since vultures are faster than tanks, we can turn back the tanks first while laying a bunch of mines across the map. We can also split off a small group to kill an expansion, and turn back with the rest of the army. When Protoss counters, they usually are stuck between whether they want to move further into our territory or trying to attack our forces, so it should be possible to trade well. If we need to cope, we can call our opponent a clown. While it might not feel great to turn our entire army around, the alternative is letting Protoss getting on top of our production and then our army just slowly bleeds out.

If they go for a side base instead, it's similar to recall. Of course, we can follow what we did above, but we don't need to turn back immediately; we can consider evacuating the base and go wreak havoc on 2+ bases and then turn back. We want to try to get more back in compensation than what we lose.

In the mid game, there's two crucial differences: We're not at the population cap (meaning we can make more units), and we have less bases to protect. Dealing with a straight counter is a matter of not letting Protoss into our main. When we see it coming, we should prepare for it by moving our units into the main, laying mines, and blocking our ramp with SCVs. In the meantime, we can charge forward with our army and set up a contain. One other tricky thing that can happen is they can threaten a counter, but instead just cut off reinforcements. I'm still researching what possible options in this case, but I believe it should be possible to trap the army and kill it with your main army and your reinforcing units (remember, Protoss's units are also cut off from their own reinforcements). You might wonder why this isn't an option for Protoss in the late game or early game. In the late game, if we're already maxed at 200, there isn't much value for Protoss to cut off reinforcements; we're not going to be reinforcing with anything. They're better off either engaging our army or just committing to a backstab. In the early game, it's possible in situations where Protoss wants to cut off tank or marine reinforcements (think 2 fact siege push or 12 nexus cheese situations) from an early contain. In these cases, you don't have a choice but to wait for more units and just bully through.

When Protoss cuts off your reinforcements, make sure you retreat your existing units at the rally to a more safe place and re-rally.

Late game win patterns/conditions

So, given that I've talked about this, where should we attack? When Protoss is on one production location, I believe that attacking their main/nat is by far the easiest and most consistent way to force a win. But when they start expanding over the place late game, it's kind of complicated and usually will depend on game to game, so let's go over some common ways I see high Terrans win, which you can try to apply to your own game.

Fight => Set up contain on nat/main => Go kill a couple bases Temporarily trap the opponent into their main/nat while we go kill their econ. It's ok if the contain gets broken. It's good if we prevent them from seeing the contents of the contain, as it's difficult for Protoss to gauge what they need to break out.

Kill army once => kill an expo (or two) => repeat Pretty simple right? One concept here that I'd like to explain is that it can be helpful to look at which bases are easy to kill and which are not easy to kill. Easy to kill bases tend to be on low ground and can only be defended by cannons. Hard to kill bases are ones with gateways at them, and on high ground. It's valid to just starve them out by letting them have the hard to kill bases, but not letting them have any of the hard to kill expos.

Trade army => deny additional Protoss expo => expand ourselves Sort of similar to split map, but the idea is we just kill their army over and over and then get map control. Eventually Protoss starves out.

Fight => Kill their main => win another fight => kill expo Kind of difficult to do, but if we kill their main, it's a lot of money that Protoss needs to spend on stuff that isn't their army. It's better if there's a massive amount of gateways in their main.

Split the map out Just don't attack. Let Protoss trade into us. Maintain map vision, mirror Protoss's army movements, don't let them get carriers.

Carriers

Carriers sort of turn the game upside down. Carriers do the following:
- Split the Protoss army value into two parts, the ground army and the carriers.
- Carriers have the most value together and are relatively immobile; by themselves, they have difficulty dealing with Terran's army splitting up.
- Therefore, Protoss needs a ground army to go along with their carriers. Additionally, Protoss's ground army tends (though not always) to be weaker than normal due to lack of upgrades.
- Carriers also require an active economy to use; if Protoss has no mining then it's possible for them to run out of money to build interceptors.
- Carriers in the early game threatens to run Terran over if Terran has insufficient anti air production.
- 8+ carriers is considered a loss condition; it becomes difficult for Terran to trade efficiently enough.

Compared to a normal game, there's at least a few differences:
- Since it's bad if Protoss builds carriers/ground army/economy all at once, killing expansions have more value than usual
- It's difficult to kill Protoss's expansions economically if we don't kill their army (particularly ground), their army has more value than usual
- Cutting off ground army from the rest of the map has less value since Protoss has mobility in the form of carriers. If Protoss only still has 2 bases, this can still be good.
- Killing Protoss's carriers will revert the game state back to ground army vs ground army (this is usually advantageous for Terran)

Therefore, most of the times when fighting carriers, we want to attack Protoss and cut down their economy before they are powerful. If Protoss has all three of high carrier count + an economy + ground army, it becomes extremely difficult for Terran to defend all the possible threats. If Protoss only has 2 of the 3, it should be playable. If they only have 1 of the 3, it should be a win (assuming we're not crippled ourselves).

When we attack, we want to kill their ground army so that we eliminate counter threats, and it allows us to solely focus on building anti carrier units. After that, we want to attack Protoss's economy. We want to kill as many expos as possible and kill as many probes as we can. It becomes viable for Terran to also expand to different corners of the map when the ground army is dead and doing so forces carriers to fly long distances without doing damage. We should try to preserve tank count (longer build time, 2x gas cost compared to the goliath).

When Protoss does not have a ground army, your goliaths should try to follow the carriers, and your tank/vulture can try to find value by killing expansions away from the carriers. In extreme cases where Protoss has no money, they can even kill threaten elimination of all of Protoss's buildings, forcing carriers to come back and defend.

In late game, if Protoss hasn't gotten carriers and don't seem to be doing much, please make sure to check if Protoss is carrier switching. Protoss usually wants to avoid Terran getting too big with too many tanks. So if they're not doing anything, ask yourself "why? what about the situation makes Protoss wants to do nothing?"

Miscellaneous

What to if the game situation is hopeless

There's basically two options. We can try all-inning and hope that Protoss expanded too fast and underinvested in their army. If this is not possible, then we should try to play very defensively, crawl across the map, and hope we can split it out.

What to do if the game gets extremely hectic (multiple drops/recalls/counter attacks while attacking etc)

My general recommendation in these types of situation is to try to prioritize what needs to be done first, then try to deal with each problem as well as you can in order (instead of trying to deal with everything at once). In my experience, if you try to deal with everything at once, you will do nothing well and lose a lot more than if you approach the situation slowly and methodically.

An example would be if we got recalled in our main and our 4th at once. Assuming my army wasn't already driving off somewhere, I would first run away the scvs to my nat from both locations, lift and run away my cc from my 4th, then go clear out the recall in my main before dealing with the recall at my 4th.

What did we learn?

I hope we've learned the main points:
1. Your army is very important. Please give it the attention that it needs
2. What we should be thinking about when we commit to a fight (Where is the Protoss army? Am I ready to fight? Is he ready to fight?)
3. Our goals when we attack Protoss.

A lot of the concepts here in this guide are not TvP specific. The army resources are pretty universal for any matchup, though the value they carry will differ a lot from matchup to matchup. Concepts like distracting them and punching in them in the face is very powerful, even in TvZ and TvT.

Some final pointers:
1. Please watch your replays with a clear mind and try to see what each player sees. It can be helpful to watch them with only your vision turned on, or with just your opponents vision turned on. This can help you understand 2. Consider reverse racing. This can help you get perspective on when the other race is strong or weak, or what their mentality is. This is also helpful for looking at how builds work.
3. I do not actually recommend calling your opponent a clown, that was just an attention check d:^).
4. As stated in the introduction, this guide is mostly conceptual. Like executing your build, it will take time to practice and implement army movement concepts. Please play the game. Just knowing what to do is not enough.

When I originally set out to write this guide, I wanted to just go over on some basic army movement patterns. However, I felt it was lacking the appropriate context, so it kinda ballooned... into whatever this guide is. I realize that this guide contains a lot of information. My hope is that this is the kind of guide where you read, gain some insights, apply it to your game, and then come back and re-read it and repeat the cycle again. In some ways, I think this guide was helpful to me just to understand how I see the game. To be perfectly honest, I can benefit from actually following the things in this guide. Thank you for reading, and please let me know (feel free to DM me) if this helped you grow as a player, or if it was a complete waste of time.