I'm going to reiterate some concepts from the ZvP thread here, but I would also encourage you to read my last game theory thread, including the discussion with Bakuryu.
One of the most mystifying things about new players coming to Brood War is how to play on certain maps. For example, if someone is a Terran player and is having a difficult time playing against Zerg, they'll often ask for advice about their build order and/or unit composition. Not bad stuff to inquire about. The reason build orders even exist, though, is because of the way maps are designed. As maps change, build orders are also going to change.
As far as unit composition goes, this is also largely determined by the map as well. If you have a build order that works really well on one map, you may try the same strategy on a different map, and it feels really weak. To compensate for this weakness, you may be advised to change your approach to playing on the map and learn a whole new build order. If you're stubborn about playing one style in order to "master it", you need to know a few things: First, no one can 'master' any strategy or play style. You can always get better. Second, no 'good' player plays only one style or strategy 100% of the time. Third, you're practicing a build order and strategy because you want to improve so you can win. If you're thinking critically about this, you made the decision to pursue that strategy in order to win, and you need to make logical decisions about what is going to lead you to being successful in specific circumstances. There's a saying for this, which is: "Trying to ram a round peg into a square hole."
Before someone screams at me, "How can I git gud at my build if I'm practicing a bunch of different builds?", let me say this: I am not saying don't practice your favorite build. Yes, practice it. Consistency is very key to breaking out of being a low-level player into a mid-level player. But if you hit the ladder with Blizzard's matchmaking system, you're going to get a bunch of varying maps. If you played on iCCup, you could just play the same map over and over again ad nauseam. You can still do this if you just open a 1v1 game on melee. That way, you can practice a certain map and don't have to worry about ladder points.
Now that that's out of the way, let's put unit compositions into a couple of very generalized categories, which are: mobile and area-control.
Slow ground units with high DPS can be considered for area-control, such as Lurkers, Siege Tanks, Spider Mines, Reavers, High Templar.
A mobile army is one that quickly move from one area of the map to another area of the map to respond rapidly to enemy movements.
In TvZ, Terran can go marine/medic/firebat for a fairly mobile army that can move out of range of lurkers before they can burrow. The medic/marine force can move around the map and hit the Zerg in many different locations via dropships. Meanwhile, the Zerg has to control ramps and entrances with lurkers, which they have to pull up and burrow constantly in order to move them around. Essentially, the Zerg is playing area-control until they can afford to get ultralisks. If the Zerg tries to move out onto an open map, they will get flanked by the medic/marine force and take heavy damage. The only way for the Zerg to win with lurkers versus medic/marine is for the Terran to mess up.
When it comes to flanking, the more mobile army always has the ability to hit the less mobile army from many different angles. In the example above, the Zerg moves lurkers across a bridge, and they get intercepted by medic/marine forces and slaughtered.
If the marines cross the bridge, they will likely get killed because the bridge funnels all of them into a small space where the lurkers can easily deal with them. The best place for the more mobile army is an open space, and the worst place for them is in a choke point where they get herded into heavy dps or AoE attacks.
Let's look at Fighting Spirit! For the following map, the black areas are impassible by ground, grey represents high ground, white is low ground, and blue is resources:
So what can we glean from this? Well, you can see that around the edges of the map, there is more high ground and barriers, and in the middle of the map, the area is fairly open. High ground and barriers are going to favor area-control types of forces, and wide open areas are going to favor more mobile forces. I'm going to show examples from three match-ups to show how this map can be generally played.
First, Zerg versus Protoss.
+ Show Spoiler +
In terms of movement speed as it relates to ground mobility:
High Templar: 2.4
As you can see, with Protoss units, the average speed is slower than Zerg's, but also because high templar and archons tend to lag behind, the Protoss army constantly has to stop and regroup to wait for the high templar to catch up. This means that the Protoss force is traveling at an effective speed* of 2.4, whereas the slowest Zerg unit from this list would be moving at a speed of 4.1. This means that Zerg is going to have a mobility advantage, and so on a map with a lot of open space, the Zerg is going to be able to respond to threats much quicker than the Protoss can field them with ground forces. This is why in PvZ, the Protoss must be able to use corsairs, shuttles, and timing attacks in order to try and prevent the Zerg from having a large mobile force.
*When I say "effective speed", I mean that the army has to wait for the primary damage-dealers. For example, if Terran has a vulture/tank army, the tanks are going to deal the most damage, and so the vultures have to wait for the tanks to move into the position, meaning that the army will be effectively moving at the speed the tanks can move.
When we look at this match-up from the Zerg perspective, going into the mid-game is usually going to cause the Zerg to pick a tech path: either going primarily hydralisk or going lurker.
Let's say both players spawn in the top positions:
Because the Zerg player has the more mobile force and can control more area, the Zerg can take the natural of the bottom left base as their third, whereas the Protoss is better off taking the closer third base.
Notice that on the map, I put a red line on the Zerg side, and a green line on the Protoss side. These are effectively the areas that each play is trying to control in the mid-game. Of course, the Zerg wants to deny the Protoss control over the green area as much as possible, and the Protoss wants to deny the Zerg control over the red area as much as possible. However, if the Protoss does a timing attack, and it doesn't yield much damage, and the Zerg gets a big hydralisk army, it's going to change the game a lot.
The Zerg is going to have access to a fourth and fifth base, and in order to expand, the Protoss has to beat the Zerg on the open field, or else the Zerg will be able to continuously deny the Protoss from taking a base in the bottom right. If the Zerg gets to hive before the Protoss can take that fourth base, the odds of Protoss winning drastically fall. The Zerg will able to cost-effectively hit anywhere on the map with overlord drops, or zerglings, or even with their main army. So if the Protoss army leaves their area in the top right to secure the bottom right, the Zerg can hit the top right or vice-versa.
The question for Protoss then becomes: How do I win?
The answer sounds simple, but it takes practice: If the Protoss has good timings and he or she can disrupt the Zerg economy early, the mid-game hydralisk army simply won't be big enough to fight the Protoss army head-on. The Zerg will have to wait for Hive to be able to take on the Protoss force, and by then, the fourth base should be secure for Protoss.
Another factor is the Protoss' own advantage in mobility, which is the corsair. Since hydralisks can't be everywhere to defend everything, corsairs can knock out scouting overlords and scourge, which will allow the Protoss shuttle units into the Zerg's bases to do damage.
But for Zerg, if you're going into the mid-game with hydralisks, you're trying to control the middle of the map:
If the big area in the middle is yours as Zerg, then it's going to force the Protoss into a possibly unfavorable engagement, or have to do damage with drops. If they can't do damage with drops or engage, then you will have a big advantage, and once you get hive, can take your fifth and sixth bases, and just continuously deny the Protoss from the bottom right and/or drop on their main base, and do this until their economy is exhausted.
If your hydralisk army or hydra/lurker army is fairly wimpy and can't control the middle, then you're better off not trying to control the middle at all.
In some cases, you'll just turtle up with lurkers and go straight to hive. When you go hydra/lurker or hydra/muta in the mid game, it costs a lot of gas, but if you're not going to be making a huge 6-control-group force, you may as well spend the gas on going to hive. Obviously, if you turtle, you will forfeit map control to the Protoss. You will only have access to the chokepoints that you can control with lurkers, and a bit later, lurker/defiler/zergling.
If you're turtling, typically, you will have the minimal amount of defense possible, and dump all your extra money into upgrades for everything, lots of drones, and lots of hatcheries, so that once you saturate all your bases, you can crank out a very large number of hive units with good upgrades.
If you are Zerg and are losing in the early game or mid game, this is a good way to claw your way back into a position where you can win. Or maybe you just say "My micro sucks, I can't control mutalisks or hydralisks and they all just die." and playing more turtlesque is not a bad way to play for a newb Zerg player. So there are reasons to concede map control in ZvP to the Protoss, but it won't allow you to always deny their expansions from going up.
Terran versus Zerg (We're talking both bio and mech!)
+ Show Spoiler +
In terms of movement speed as it relates to ground mobility:
Yes, you read that correctly. A lurker is as fast as a stimmed unit. On paper, the Zerg army looks more mobile. A defiler moves at a speed almost equivalent to a normal Terran infantry unit. The Zerg army can traverse area-by-ground quicker, but there are a few factors to consider: The first is cost-effectiveness of melee units versus ranged units. In small numbers, melee units destroy ranged units. However, as the number of ranged units increases, the trajectory of their effectiveness slopes ever more upward. Also, due to the effectiveness of firebats and medics in supporting marines, the effectiveness of zerglings is essentially very low unless they are supported by bigger units, such as lurkers/mutalisks/defilers, or until the Zerg can upgrade adrenaline glands.
Another factor to consider in a fight between marines and lurkers is cost. Marines cost a lot less, and so trading units (both armies being mostly dead at the end of a fight, or losing units to kill units) hurts the Terran player less. This means that the Zerg needs a lot of lurkers and zerglings to win a huge fight, or has to manage their losses carefully. One last factor to consider in mobility is the simple fact that lurkers have to burrow to attack and unburrow to move. The slow speed by which lurkers burrow allows the Terran player to move units out of the way before the lurkers are burrowed, and even to attack and kill a lurker before it is burrowed.
Map control is very important in this match-up. In other match-ups like ZvZ, map control doesn't matter as much.
The general flow of the TvZ match-up has changed a lot over the years. In the late 2000s TvZ, the match-up would work something like this:
Both players would expand as quickly as possible, and the Zerg would tech up to mutalisks. The mutalisks would give the Zerg de facto map control because the Terran wouldn't be able to leave his/her base. The medic/marine forces would stay at home to fend off the mutalisks and waited for science vessel to move out.
Until about 9 minutes into the game, the map would look like the above map, with Zerg essentially having the whole map, and taking another main base. Because the science vessel's irradiate does splash damage, stacking mutalisks is no longer effective at that point, and the Terran would be free to move out onto the map.
That paradigm of the Terran being stuck in their first two bases changed when Terran players realized that they could delay their science vessel in favor of getting more marines. Therefore, if the Zerg committed their mutalisks to attack the Terran's base, the Terran could easily produce enough marines to break the Zerg's front before lurkers AND defend against mutalisks. This meant that the role of the mutalisk for Zerg changed from being an offensive weapon to a defensive weapon. The mobility offered by mutalisks and zerglings allows the Zerg to bounce back and forth across the map in the mid-game and head off the Terran forces in order to defend their natural and their far-third.
The first part of the game involves the Terran controlling the entire middle of the map with medic/marine, then adding in tanks and science vessels. However, because tanks are so slow, the Terran begins making vultures and controlling the center with spider mines.
There's a lot going on with this next map, but bear with me:
In this scenario, the Terran wants to take the top left base as their fourth base. Eventually, the Terran player will also take top left natural and 12 o'clock positions as well.
The standing Terran army will be using science vessels and tanks to put pressure on the Zerg, along with dropships. The more stuff the dropships and science vessels can kill, the longer it will take the Zerg to build a strong economy. The middle of the map is blanketed in mines, so if the Terran attacks one front, the Zerg won't be able to move out and flank from the other direction. The Zerg's only other option is to use overlords to drop units into the Terran expansions, so the Terran begins building a lot of turrets around their vulnerable bases.
Why does the Terran need so many bases? Because the ultimate goal is build a mech army out of factories, which is very difficult for a Zerg to deal with unless the Zerg has a big economy and lots of production.
So why not just go mech to start with?
Obviously, you can. But once Terran goes mech, the mobility advantage and map control advantage will go to the Zerg player. On Fighting Spirit, Terran is essentially confined to three bases when going mech, and has to step out into the open middle in order to take a fourth base. The same sort of dilemma exists in TvP, where Protoss controls the center in the middle game.
So let's look at another map:
This is Destination.
As you can see, there isn't one big area of open ground on this map. There are some big areas, but they're separated by bridges. This is where Destination is notorious for giving an area-control army an advantage. Armies that have a lot of mobility aren't as free to spread out and out-maneuver their opponents. They're all kind of funneled into a small area and small areas, chokepoints, and high-grounds are where area-control forces excel at being efficient. So while on Fighting Spirit, you need to out-maneuver your opponent, on Destination, you need to out-position your opponent.
So, Terran starts on top and Zerg on bottom. At the start of the game, the only caveat to going mech is being vulnerable due to having so few units. Some ways to deal with that are to harass the Zerg with vultures and/or wraiths, and scouting to see if the Zerg is going mutalisk or hydralisk first. If it's mutalisks, you start producing goliaths first, and if it's hydralisks, you start making tanks first. A good mech army is eventually going to have both goliaths and tanks, with about a 60/40 or 70/30 distribution of units, plus vultures, but it's important to know what the Zerg is doing by scouting. After that, it's just mines, tanks, turrets, and science vessels everywhere for days and days.
As you can see, Terran has very little map control. It's okay, as he or she will only need to defend that area and build up enough forces to creep forward and take more bases.
Eventually, Terran will take over the top of the map. I know the Zerg has a lot of bases in this scenario, but this is still perfectly winnable for Terran. I drew arrows showing the paths that Zerg can take to move into the Terran area, and all of those paths are across small bridges, which means the Zerg will have to go air, or send endless waves of units into suicidal attacks.
As long as the Terran stays busy killing things with science vessels, and laying mines, and replacing lost tanks and turrets, it will be difficult for Zerg to punch through unless the Terran has been massively delayed from expanding. Even if the Zerg breaks through, it's not a guaranteed death blow.
So how does Zerg deal with this? Part of the answer is to delay the Terran from expanding.
I would recommend starting off by trying to control the area across the bridges from the Terran's natural. The Zerg will not be able to hold this indefinitely, but by containing the Terran for a long time, it'll prevent the Terran from taking their third base right away. Even still, this can be a tricky map to play on versus Terran, because the chokepoints really make using ground units difficult, and the Terran can easily make a small fleet of valkyries and/or science vessels to deal with mutalisks/queens/guardians.
To summarize: If you're a Terran who is used to going bio, you also need to consider that there are some maps where you can go mech, and it will work really well for you. If you're Zerg, you need to have a good anti-mech strategy.
Protoss versus Terran
+ Show Spoiler +
In PvT, mobility is weird. In PvZ, as Protoss, your static defense is really good. In PvT, there's no such thing as static defense. Tanks massively out-range cannons, reavers, and high templar, so setting up a static defense to fight tanks is pretty useless. Protoss has to be completely mobile army in order to flank and destroy the Terran force, and it has to be done with arbiters following a huge number of low-tier units. However, Protoss has the option to go carriers.
Carriers are going to generally work better on maps where they can hit a target and then run away. If there is a high ground or some kind of barrier, that's where carriers are going to be the strongest. So on a map like Destination, Carriers will be more effective than on Fighting Spirit.
The first phase of the game for Protoss is going to be containing the Terran into a corner. The dragoons and zealots form an arc outside of the bridge in order to catch the Terran if he or she moves out onto the open center of the map. Technically, the Protoss has the more mobile force since tanks have to siege and unsiege in order to change positions. However, the main threat at this point of the game is not the Terran moving out with tanks, but sending out vultures. Vultures are the fastest unit in the game, and will outmaneuver any other kind of army, so it is necessary to place dragoons, pylons, and cannons in ways that become physical barriers to the vultures.
The next phase of the game for Protoss is to max out his or her army completely while teching up to arbiters, and begin building more gateways. In the map above I show the Protoss controlling two main bases, and having gateways in each one, and rallying those gateways out onto the map.
Also note I crossed out the 12 o'clock base as a possible expansion. I'll talk about that more in the next section, and why I put that there.
Once the Terran reaches a certain point (either 150 or 200 supply with double armories) they're going to move out onto the map. For Protoss, the next step is fairly straightforward: while the tanks are unsieged, send in units from as many directions as possible. Dragoons should go first, then zealots. If Protoss has shuttles, they'll then send them to drop units into the tanks. If the Terran screws up, the Terran army will be wiped from the map, but if the Terran engages well, they will push the Protoss back.
A good Terran engagement involves keeping vultures out in front, not sieging on spider mines, having six goliaths to shoot down the arbiter, and putting their EMP on either an arbiter, or into the bulk of the Protoss forces. Of course, not having all the tanks in one big clump helps as well.
If the Terran breaks through, the Protoss natural is probably going to fall. But it's alright. The Terran doesn't really have "map control". They control a weird-shaped path through the top of the map, and when spread out over that wide of an area, there are weaknesses. A good Protoss will not freak out over this, but in fact, should expect this. The Protoss will re-rally all of the gateways in their main to the top of their main base's ramp, and use the other gateways on the bottom to flank from the bottom of the map, or do a recall into the Terran's base.
If the Terran gets into this particular position, the best thing to do would be to contain the ramp to the Protoss main, and not try to go up into it. The Terran should send a group of two tanks to the Protoss' third base (in this case, the 9 o'clock position), and send the bulk of their army to block of the path of the Protoss reinforcements:
In this scenario, killing the 6 o'clock and 9 o'clock bases are going to be key for Terran. The bulk of the Terran force should be to block off reinforcements. Mines should be out on the map, and turrets should be up, because this is the time when a recall from the Protoss could screw everything up.
Terran's expansion options are limited. Expanding to the middle isn't a bad idea, but it isn't easy to defend, and there aren't very many minerals there. The 12 o'clock is close, but susceptible to a very close recall. The bottom right is also vulnerable, but isn't as close to the Protoss main.
If the 6 or 9 positions fall for Protoss, the only way to reasonably have a chance is to utilize spells effectively, and do damage with recalls.
+ Show Spoiler +
This section has a lot to do with planning where to expand and why certain spots can help you to defend yourself better, and also attack better. Usually, the rule of thumb is: Don't expand towards your opponent. There are many exceptions to this rule, which I'll discuss in a second, but first I want to explain why someone would want to avoid expanding towards their opponent. Let's look at PvT again:
The first factor is distance. If Terran spawns top right and Protoss spawns top left, then the 12 o'clock expansion is probably not the best choice for the Protoss. Rather than having to move deep into the open area where getting flanked is possible, the Terran can leap-frog right up to the 12 o'clock, not get flanked, and burn down the nexus there with relative ease.
The same would apply for a Protoss player expanding right next to Zerg, or Zerg expanding right next to a Protoss. It's not impossible, but there is very little time to scout an incoming attack and react to it appropriately. However, if you are building a lot of static defense, then yes, expanding towards your opponent is not only doable, but logical as well. If you're Zerg, and you expand somewhere close to the Protoss and put a bunch of lurkers/defilers/sunkens/spores there, the Protoss might end up throwing away a lot of units to try and remove you from that spot. However, if you cannot defend the base reasonably well, DON'T DO THIS.
You're Zerg, you're planning to take the close third. You spawn top left, and Protoss spawns top right.
The distance difference between the 12 position and the 9 position gives you more time to send your forces to where they're needed. In a similar scenario, if you and Protoss spawn in the same positions as shown here, and you want to take the far third, which base to take would be a no-brainer. You'd take the bottom left natural.
You're Zerg, you're planning to take a far third. You spawn bottom left, and Protoss spawns top right.
Where do you expand? This has a less definite answer, because if you're in opposite corners, then no matter where you expand, it's going to be towards your opponent, relatively speaking. So, the solution I would propose is to ask yourself, "If the Protoss does a 7-minute zealot timing attack, from which natural will it be easier to scout?" The answer may come down to distance:
If you assume that all attacks will come from the Protoss natural, then the distance probably won't differ much. I've heard the argument that "On fighting spirit, the Protoss always blocks off their third base ramp, so no attack will come through there."
It's an interesting proposition, for sure, but I would also say that most Protoss launch their timing attack while taking their third nexus, or slightly after, which means the ramp will be unobstructed, and the Protoss area will be difficult to scout with few units that early in the game. From the example above, Option A, to me, seems more ideal because it is away from the Protoss third base, as opposed to Option B, which is right next to the Protoss third base.
There's actually no "wrong" answer here, but I hope that the logic behind my reason makes sense. I'll add another argument for expanding to Option A in a moment when I talk about drop paths, but first, I want to give one more example from ZvP about expanding.
You're Zerg playing against Protoss on Destination, and you're in the hive-stage of the game, and have all your tech. So which base do you take next after the first four? There isn't really a correct answer for this, but I wanted to propose a scenario in which you expand towards your opponent:
Rather than taking the 10 o'clock position, a Zerg with defiler/lurker can easily take the base on the middle left, which is somewhere between 3 and 4, so I labelled it "3:30". The benefit to taking such a base is that because of the high ground, and lurker/defiler, this base requires very little resources to effectively take and defend. Most Protoss players like to take the 4 o'clock position, because it is easy to defend with cannons and templar. As an interesting side note, the bridge down there is bugged so a scarab cannot shoot across the bridge.
Having the 3:30 base makes assaulting the 4 position very easy to do with guardians and such, and will pressure the Protoss to engage the 3:30 base. This can be quite good for Zerg because the Protoss will be fighting uphill into a bottleneck.
I don't know if there is an existing term for this, so I don't want to take credit for having coined this phrase. However, I rarely hear anyone talk about drop paths or anything like it, so I'm not entirely sure. So, please forgive my ignorance on this matter.
Simply put, a drop path is an area that you can fly over that is safe because you have presence there and your opponent does not. Typically, when a player does a drop, either with overlords, dropships, shuttles, or with an arbiter, they tend to fly where they think their ship will not be intercepted or destroyed. This means that flying over the middle of the map into a dark fog-of-war is a bad idea, generally speaking. You don't want to fly over anything that will see your drop coming, because then once the drop arrives, there will enemies waiting to neutralize it.
You're a Protoss and want to recall into the Terran main.
The Xs are turrets. You can probably see the problem here already. There is no obvious drop path to the Terran main. There is no drop path to anywhere, really. Every route you could take with the arbiter will get spotted by mines, and the arbiter will have to fly over turrets somewhere, and probably will be crushed by ground forces even if the recall is successful.
Edit: Just to clarify, if Terran is turtling really hard, Protoss can take more bases, and afford carriers. I think it's easier to play carriers on Destination, but if you're Protoss and want to recall, try this: hallucinate your arbiters, and keep the hallucinated arbiters in front, and let them take the turret hits.
Kind of the same scenario as example 1. You want to drop on the Terran main and you're Zerg.
This scenario is different because the Terran expanded towards the left side of the map. Because you, as Zerg, control the 3:30 position, you can fly your overlords over that base, which means that the time the Terran will have to react will be very short. Even though there is a turret ring, the best place to drop will be around the back of the Terran's main.
Keep in mind that I've not saying the drop will be a success or a failure. I'm just saying that in order to hit your opponent with a drop, it is best to fly over area that you control. If at all possible, you should think of this in advance. On some maps, you won't even have to. If you play ZvP on Tau Cross, and you spawn bottom and Protoss spawns left, your natural and third create a very convenient drop path right into the Protoss' main. Maps like Python, Gaia, Paranoid Android, and other maps with 'close' spawns have fallen out of favor because the drop paths can be so short that it is almost impossible to scout them or react to them in time.
You're Protoss and you want to drop a combination of high templar and dark templar into the Zerg.
I brought this up earlier about expanding close to the Protoss third, and here's one more reason why I think it can be less-than-ideal. If you, as Protoss, spawn top right, and the Zerg spawns bottom left, but takes the bottom right as their third and fourth bases, then you have a really convenient drop path to the Zerg's third. All you need to do is clear the area with corsairs, and then psi-storm the drones.
If the Zerg takes the top left instead, you have a much longer drop path across the top of the map. If you want to cross the center of the map, it can be extremely risky unless your shuttle is following your army through the middle. Generally speaking, the more distance the shuttle has to cover over unknown or enemy space, the higher the risk.
Creating a drop path can vary in difficulty depending on the situation. But basically, as long as your stuff is there to defend the area, you can generally feel safer flying there.
There are many different drop path scenarios I could talk about, but for now, I want to focus on the two key reasons why you should know where drop paths are: The first is to know where to send your transports so they don't die and you don't have to lose that investment. The second is to be able to defend against drops. The end.
What do you think? Let me know in the comments. We'll talk it out and have a good discussion!