The Protoss Help Me Thread

Teoita wrote
9 scout, use this http://www.teamliquid.net/forum/viewmessage.php?topic_id=296273 if he's Zerg. Against Terran or Protoss, follow your preferred build order as much as possible.

Teoita wroteYou can still have good enough scouting info and reads on a Terran with 13 scout on most reasonable maps

rsvp wrote:
You're not way behind if you just play standard, especially if you plant your nexus down before any gateway units.

TheDwf wrote:
Yes, other builds don't have spare minerals and/or there would be no point in the EB block.

TheDwf wrote:
Slightly delayed Barracks/gas(es)/Bunker, wounded or dead SCV limiting rescouting/forcing another scouting SCV to check if Protoss did expand (else you risk losing to some delayed 4g or even 4g prism).

Teoita wrote:
Other than making sure it doesn't happen, getting a zealot before stalker always helps. You more or less just have to accept your nexus will be delayed, but the terran will have invested 100 minerals into it. There really isn't much to it or an optimal response, it's just an annoying thing to have happen.

Teoita wrote:
The key to holding off this rush is having precise Stalker micro, getting shots off on the Marines and Scv's as they cross the map while staying out of Marine range. Once the fighting starts at your base, it's important to keep bunkers from being completed, focusing building scv's with chronoboosted Stalkers and pulling probes if necessary.

The easiest way to hold off this all-in is to build your third pylon next to your main's main ramp, and walling off with two extra gateways when you see the rush coming (either you scout hidden barracks or you see the scv's being pulled). This allows your Stalkers to get shots off on the Marines, and more importantly buys extra time for more units to come out. Note that this method isn't used very commonly in professional games as it forces the Protoss to have a pylon in position beforehand.

monk wrote:
Firstly, you want to identify all the builds your opponent can do with a gas. The prominent ones include 1 rax fe, 2 rax with reactor/tech lab, 3 rax, or a factory/tech build. These factory/tech builds could include cloaked banshee, marine tank allin, thor allin, or some type of hellion build. I'd say these are the primary builds you want to look out for. Now with every little piece of information you get, you want to cross off possibilities based on this information. For example, if you see a maruder, you will most likely want to consider tech builds as much less likely.

Now for actual scouting tips. The first thing you want to look for is whether he builds a rax or a depot after his first marine. The standard 2 rax build builds a rax before a 2nd depot, so if you see a depot being built, 2 rax will be less likely. Keep in mind, however, that your opponent can still cancel the depot. Next, the most important part is that you want to keep your initial scouting probe alive as he can get you a lot of information. Poke around his ramp and see what's guarding the bottom of the ramp. If you see a marauder, then you know it's a tech lab build, most likely 1 rax fe, but could be 3 rax or a tech lab 2 rax push. If you see 2 marines early, you know it's a tech build, because every other build adds an addon immediately. This also tells you that you can poke around with your initial stalker and harass him. If you see just one marine, you can choose to run past the marine and either see what addon he has on his barracks or what other units he has on the top of his ramp. If he has no units at the bottom of his ramp, feel free to poke up and have a look and make a judgement call off of that. Even if you see just marines, you can get a lot of information from that, because tech builds go for factory and then reactor later on the rax while a standard 2 rax gets the reactor right after the first marine. Thus, a tech build will have a gradual build up of marines while a 2 rax will just have one early and then 2 at a time.

Teoita wrote:
I try to return with my probe when i know there should be a low ground CC started, and only one Marine.

His first Marine should pop out at about 3:07, while the second should be done at 3:32. Poking in between those two timings should show you wether there's a low ground cc, and most importantly wether he's starting and add-on on his rax.

If you can't find anything with your probe, poke with your first zealot and stalker to get more info (regarding the specific info you are looking for, there's plenty of threads on TL).

Teoita wrote:
Marauders; once they are down cleaning up the marines is really simple.

kcdc wrote
I'm always really tempted to go for the counter-attack when I clean up those 2-rax plays, but I think the higher win percentage play is to stay patient and play your usual style, knowing you're playing from ahead. I find that the Terrans usually expect counter-pressure to come from Toss, and they wind up defending it pretty comfortably.

A 2-base colossus push can be game-winning of course, but one thing to bear in mind is that the 2-rax opening weakens his economy and his long-term unit count, but it doesn't delay his tech. In fact, you'll often see the 2-rax pressure opening into expansion get a reactored starport qucker than the gasless CC opening will get one. Since the colossus push is laregly about hitting Terran before they have a decent viking/medivac count, the push actually isn't much stronger in this situation than it would be if Terran had used a standard passive opening.

I'd recommend using your economic lead to take a series of smaller victories instead of going for the throat right away. Get an upgrade advantage, grab a faster third, delay his third, etc. There are a lot of good options, but busting his natural right away isn't the best one IMO.

Alejandrisha wrote:
The most important thing to account for with observer scouting is considering your opener and how quickly it permits you to train your first observer. Take four staple Robo builds:

1 Gate FE → Gate Gate Robo
1 Gate FE → Robo Gate Gate
2+ Gate Robo
1 Gate Robo → Gate Nexus

Remember this when changing up builds often or you might miss the things you are used to seeing when using a riskier build.

Depending on your probe/GW unit scout, your observer will be used for different things.

CC spotted

The things to look for are: Non standard play → 3rd CC → 2 ebay → 1 or 2 tech labs

If you see a 3rd CC, you’re free to take a 3rd. If there is no 3rd CC but double upgrades, taking your 3rd is a tough call because 2 base Terran with 2/2 and a 2nd starport will crush you if you are defending with Colossus. If there is just 1 ebay, look for the additional rax, as Terran will most likely go up to 5 rax and 1 starport from here. Terran will usually just get the 1 tech lab for upgrades with their medivac timing. If they get a second one, consider adding an additional Immortal.

Non standadrd play refers to tech labs on factories or ports. Immortal/Colossus/Achon/Stalker does well against these compositions because your army is too mobile to grant Terran a free siege-up so you can get a strong composition in time. Immortals are obvoiusly great against Tanks and Thors, and Storm (if you can get it in time) is extremely powerful as well as there will be less Medivacs and the Marines will not be well upgraded.

No CC spotted

If your probe scout doesn’t see a CC, but sees a bunker, you need to confirm it with a gateway unit poke. The number of Marines in the bunker can be indicative of whether or not the barracks is being used for Marine production (per standard FE) or building addons. A good time to keep in your head is 3:07, when the first Marine pops from a 12 barracks. Every 30 seconds from here should mean 1 more Marine in that bunker.

If there is no CC on the low ground, the Terran is most likely opening Factory. Your first observer needs to see what and where the addons are. If you see a port, but no tech labs on the factory, look for a 2nd port or a CC.

Teoita wrote:
Against any terran one base all-in, the most common and popular way to hold is to do some sort of 1gate expand into 4-5gate+robo defense with immortals. It's extremely hard to tech up against 1basing terrans and if you do the only viable choice is to go colossus, to deal with any delayed all-in that waits for too many marines for your gateway units to handle alone.

That advice is generally regarding 111, but it also works vs bio allins which are actually very weak.

kcdc wrote:
The answer is still good FF's and gateway units. You'll also need to do some zealot/stalker micro using zealots to tank, stalkers to focus fire, and keep reinforcing. Terran winds up doing a lot of damage stimming himself to death too.

monk wrote:
In order to defend this attack, it is essential that you scout it out. There is no other way. If you don't scout it or kill it very early, there is no hope of beating it. Ways to scout it include gateway pressure, stargate, pressure, but most commonly an observer. Any build that does not incorporate any of these is close to an autoloss versus the 2 port banshee build.

Now, once you scout the 2 port banshee, this is what you do:

Immediately throw up a stargate and a forge. Make only enough stalkers to fend off his initial banshee harass and then cut stalkers entirely. Constantly chornoboost phoenix out of your stargate and throw up ~5 cannons at your front. Make only zealots and sentries from now on, but don't overdo the sentries. ~3 is perfectly fine. Cut probes accordingly around 35 if you don't scout a command center.

This is the only way to fend off the 2 port banshee allin if you're on 2 base.

monk wrote:

First let's talk about the specifics about the push; this push can hit any time after 9:30 depending on the map and how greedy the Terran was in his build. Make sure you check your opponent's add-ons. Two reactors and one tech lab on his barracks is indicative of this classic marine-heavy build while two tech labs and one reactor on barracks is something slightly different, a push that usually won't be as strong but has other advantages. With the marine heavy build, the first wave usually consists of 2 medivacs, 2-3 maraurders, and about 25 marines with constant reinforcements. There are two key points to defending the standard 10 minute medivac push.

1. Being able to directly fight off the initial army
   Of course the first thing you have to worry about is being able to survive if the Terran a-moves up your ramp. In order to survive you need either a certain amount of gateway units or a certain type of tech. Let's look at some popular PvT builds to get some ideas:
   
   • Fast colossi build: This build gets one colossi by the time the Terran push can hit. This colossi is especially effective, because your opponent's push will be extremely marine heavy. But even though a frontal push won't work against this type of build, colossi builds are particularly susceptible to drops, so you have to especially worry about those.
     
   • Startale templar build: This build relies on a large amount of gateway units along with one archon to hold off this push. The archon provides essential splash damage and is often targeted down. Another thing to take away from the Startale build is the timing of the addition of gateways. Instead of a traditional 3->6 addition, the Startale build goes from 3->5->7. With this order, the 4th and 5th gateways finish just in time to defend the push.
     
   • Rain templar build: This build gets up charge before the push can hit. With a large amount of chargelots and good forcefields, this push can also be stopped.
     
   • Fast templar builds: These are uncommon, but it's possible to use templar, in combination with gateway units, in order to stop these marine-heavy pushes.
   
   So in summary, you need either: a large gateway force, a colossi, chargelots, or templar with storm. Good control is always required as the marine do damage so fast that one mistake could easily cost you the game.
   
   
2. Being able to defend the initial drops
   
   • The most common way to stop drops in the early game is with a six stalker force sitting in your main while your main army rests in your natural. Six stalkers is the magic number that will allow you to two-shot a medivac force. The problem with using stalkers to defend drops, however, is that if a drop does land in your main, you'll have a hard time removing it from your base. Blink is valuable addition and can net you medivac kills you normally wouldn't be able to get, but sometimes it is out of the way to research.
     
   • Generally, your other option to deal with initial drops is with chargelots; this is the tactic employed with the Rain templar build. Chargelots are great at stopping drops once they're on the ground, but they have trouble with doom drops if you didn't split your army properly.
   
   As always, vision is essential to stopping drops, so get observers in key places. Also keep in mind that when you're defending drops with units in your main, your army in your natural will be weaker. A high level Protoss should be able to tell how many medivacs and bio units a Terran should have at any given time, so if he spots the Terran army, he'll know how many rouge drops are on the map. He can then split his army accordingly.

Teoita wrote:
For having idle gateways and getting units when you need to and tech/infrastructure when you can, i suggest looking at this thread (yay for self promotion) http://www.teamliquid.net/forum/viewmessage.php?topic_id=364130; i talk a bit about how production is strucutred with macro pvt builds and why you want your buildings/gateways at specific timings. This is for 2base play, but 3base is similar. If you note, 3nex cuts units and ramps up to 8-9 gates worth of producion, and it gets a huge spike in production when a standard medivac timing would hit.

Teoita wrote:
If both sides fast expand it's a stylistic choice. A couple of good templar builds are http://www.teamliquid.net/forum/viewmessage.php?topic_id=319339 and http://www.teamliquid.net/forum/viewmessage.php?topic_id=364130.

If he's opening gas it's harder to open templar because while you can get colossi in time vs some one base allins, in some situations, it's basically impossible to tech to templar in that spot, and by the time you can confirm expand or allin (around 9 minutes or so if you go for a safe 3 gate into robo) it's too late to choose what kind of midgame composition you want to go for.

Teoita wrote:
To defend drops, probably a single forge fast blink opener like what PartinG does here:
http://www.gomtv.net/2012wcskorea/vod/70396/?set=2&lang=

You can go into either Templar or Colossus off of it; for Templar copy what PartinG does exactly, for Colossus here's a replay of mine: http://drop.sc/268703.

kcdc wrote:
Beyond the obvious (make observers, control watch towers), keeping your army moving helps. When you're moving your army, you'll naturally be thinking about where he is and without even realizing, you'll be making decisions about how to better keep tabs on his movements. If you watch pro games, you'll see their armies are constantly shuffling back and forth even when it seems like it's just wasted clicks.

And if your problem is not map vision but rather actually noticing what's going on in your minimap, I'd suggest using windowed mode and making the game window small. My monitor is simply too big and too close to my chair (shallow desk) for me to see my minimap well with my peripheral vision. The smaller window keeps the minimap close to the center of my focus all the time. It wound up making a huge difference in my minimap awareness, and it game me a small APM and click accuracy boost too (didn't have to move the mouse pointer as far on the screen).

Teoita wrote:
If they are going for 10 min medivacs, go up to 6 gates and establish a decent army and take the third when they do, and playing against fast 3 cc take the third off of 3 gateways

Teoita wrote:
Either 3 with range or one with no range. One with no range to hold medivac pushes is pretty much a two base templar build (you get your twilight council when +1 armor is halfway done, then start charge/+2/Archives and take a third); true colossus builds often cut at three and they are getting less common in the pro scene due to the 2/2 viking timing you talk about, as it can hit before storm even completes if done well.
I have seen some Kespa Protoss doing lots of double forge and getting as many as 4 colossi while maxing on archon/stlaker/zealot/colossus and delaying storm (off the top of my head, Jangbi, Terminator and Rain all did it in Proleague), but none of them has ever been up against that viking timing (i think) so i don't know how they deal with it.

monk wrote:
The key is vision. That is, observers, spotting pylons, and watch towers. The obvious benefit of this is that you can spot drops incoming. However, if you have vision of his army, you can correctly position your troops to accordingly defend both his army and any potential drops. For example, if you see his army far away from your base, you don’t need to keep your entire army at your front. Instead, put it in probable drop locations.

Another option is pressure. A believable fake allin will sometimes keep your opponent contained in his base. He will be afraid to send anymore than 1 medivac out to harass, which is a manageable amount.

Notice how I didn’t mention anything about units. Blink stalkers, chargelots, templar(with or without storm), and dts all do fairly well versus drops and almost any unit composition you go will be able to defend against drops.

Teoita wrote:
You generally need a 3base economy to get both Colossi and High Templar.
With a Colossus opening, you will usually get 3 Colossi, blink and a decent stalker count (10 or so), before reseaching Charge and starting a Templar Archives.
With a Templar opening (including getting a single colossus without range and then going for Templar tech on two bases), the timing is less set in stone and it comes down to the preference of the players. A good general idea is to aim at having a good High Templar count (about 6-7 spread between your 3 bases) to be able to defend while investing into a support bay and a second robo.

Teoita wrote:
kcdc wrote an excellent guide on lategame PvT:

http://www.teamliquid.net/forum/viewmessage.php?topic_id=370106

rsvp wrote:
I'm really glad you wrote this guide. As you noted, many commentators and players alike focus too much on build orders and timings, which become completely irrelevant when you get to mid-late game. The "management" part of mid-late game is so crucial, but often ignored. I'd like to add some additional tips I utilize among your 4 pillars:

Map vision: Observers. Make them. A lot of them. If you're a player who typically only makes 1-2 observers each game, and doesn't bother to remake them after the terran scans and kills them, try making 3-4. I regularly chrono out 3 observers every game as my first 3 robotics units. And then when I lose them, I prioritize remaking them over colossus. Put them along attack routes, drop routes, everywhere. The game will become so much easier for you when you have a virtual maphack.

Mobility: In addition to pylons all over the map, get in the habit of overmaking gateways so that you can instantly create an army anywhere on the map. This can also help you set up some nice flanks, or allow you to just drop 20 zealots into his main and wipe out his production (barracks/depots) even when it's mined out and you can't do any economic damage.

Also, I want to stress kcdc's point about multiple battles at once. Anytime you're about to attack, or an engagement is about to occur, you should be warping in zealots/dts somewhere else on the map, whether it's next to a pylon near his 3rd/4th, or via a warp prism into his main. It literally only takes 1 second, but has the potential to do so much damage - shift clickclickclickclick, then drag select and shift click into the mineral line.

Positioning: This is difficult especially in the heat of battle, but it's extremely important. As your armies are dancing around each other, keep your stalkers and colossus near each other, and zealots separately so that zealots don't get in the way of stalkers chasing vikings and stalkers don't get in the way of zealots charging into the terran army. Always know where the enemy vikings are, and position stalkers in between your colossus and the vikings. Keep your colossus away from cliffs and water where it's more difficult for stalkers to protect them. This means that when you're walking between your natural and 3rd on daybreak or cloud kingdom, make sure they're not hugging the cliff/edge of the path. When your armies engage, DON'T immediately engage with your zealots. Wait for the MMM ball to get in range of your colossus before engaging with zealots. The terran wants to kite your zealots while being out of range of your colossus, and if you just a-move your zealots into the terran and forget about them, the terran will end up killing all your zealots while avoiding colossus fire while your colossus are dying to vikings. If terran runs out of range of your colossus, PULL your zealots back and wait for him to get in range of your colossus again before engaging with zealots again. Meanwhile, always target fire vikings with your stalkers and don't let them just a-move. So pretty much when the engagement happens you want to quickly shift-click target a bunch of vikings with your stalkers, a-move with the colossus, and then focus on controlling/dancing your zealots with the MMM ball. Use blink to reposition your stalkers so that they're not in the way of your zealots and/or so that they're not directly being killed by MMM.

Another note, when the terran gets a million ghosts and you can't land any storms, you should be colossus heavy (4-6+ colossus), and focus on storming the vikings. You'll never get a storm off the main MMM ball, but it's much easier to get storms off the vikings and they will actually die pretty quickly (especially if you also have archons) and then at least you'll still have colossus against MMM.

Space control: HT all over the map. You never want more than 3-4 templar in a single group, because it just takes too much effort to keep re-spreading them out. A lot of times in late game I don't even have any templar hotkeyed. Everytime the terran pushes forward, he should have to go through 1 templar at a time, eating storms or feedbacks if he's not careful. If he is careful, you only lose 1 HT at a time. Be active with your group of blink stalkers so you can snipe ghosts who stray too far forward to snipe your HT, or small groups of MM that try to snipe your HT.

Also related, while it's not about space control but rather how to use HT, always try to "attack" the terran army like kcdc mentioned with 1 or 2 HTs at a time. It's easy for terran to deal with stationary HT after he scan where they are, but oftentimes you can take the terran by surprise by just attacking with 1 or 2 storms, by themselves, at a time.

Teoita wrote:
I keep 4 ish in my main army, one in my main, one in my third, one for each outlying expansion, and like another 3-4 or so around the map gathering energy.

monk wrote:
Nexus first is more economical. 14 Forge 17 Nexus is always safe while blind 17 Nexus is not safe. If you scout 14 pool or later on all FFE-able maps, you can safely go Nexus first. Depending on the map and the timing of the pool, 17 nexus may or may not be safe versus pools earlier than 14 pool.

monk wrote:
15 hatch is actually somewhat safe versus Protoss. The 2 reasons high level Zergs don't do it often are as follows:
1. You normally can't get the 15 hatch down before a probe comes to your natural and tries to block it.
2. Cannon rushes can be very deadly versus 15 hatch. You usually need to commit a drone scout or drone block to be somewhat safe versus cannon rushes.

With gateway first versus 15 hatch and a late gas, you can chornoboost stalkers out, gain map control, and force lings or spines. With forge first, if your opponent isn’t blocking his ramp properly or isn’t watching your probe well, you can attempt a cannon rush. However, both of these options are dependent on your opponent, so in reality, there’s no reliably consistent way to punish a 15 hatch.

Teoita wrote:
Assuming you open FFE:
Scout for a third at around 4.30-5.00 if he opened 15p/16h, or for a nat at 4.40 and a third at around 6.30 if he opens speedlings (im stretching these timings a bit, it should be faster but no silver player will have a properly timed opening). If he doesn't take it, get a fast sentry and an extra cannon because he might be allining you. If he doesn't even take his nat, he's probably going for some 1base roach or baneling allin (really really common in low leagues afaik). Also if you see no third, patrol a couple of probes around your base to check for nydus stuff.
If it ends up being 3base vs 2base, do whatever your go to build is, whether you take a 3rd or just kill him on 2 bases. The big sign to look for is whether zerg is taking 4 or 6 gasses, if he only has 4 he's probably going for a roach max while if he goes for 6 he is more likely to tech.

Monk wrote:
The best way to prevent this is to estimate when the queen will pop out and surround it with probes. If you have access to GSL vods, MC vs Leenock shows a perfect response.

kcdc wrote:
(1)See his drone coming, and hatch block your natural with a probe. There's no good reason for Z to send an early drone scout, so if they do, you should expect that they're up to something.

(2) Kill his hatchery before it finishes. You really only want to make 1 cannon, so it's good practice to have your building layout allow a single cannon to cover both your wall and the spot where your nexus would go. If you start the cannon early, you can kill the hatch before it finishes just by pulling a couple probes early on to supplement the cannon's DPS.

Teoita wrote:
If they went 14-15 pool, your cannon should barely be done in time on most maps, if he doesn't send the lings straight across (which is risky since he will ignore your probe and risk getting a hatch blocked). If you want to play it safe, have a probe on hold position in the gap in your wall, ready to throw down a pylon to block the entrance fully as the cannon finishes.

CecilSunkure wrote:
That's a super minor detail that doesn't really matter ever. I always wall with my Cyber core. What really matters a whole lot more is the location and placement of the wall buildings, rather than specifically whether or not it's a gateway or core.

rsvp wrote:
I chrono out 2 zealots, which will beat 4-6 lings. If he makes more lings than that, add 2-3 extra cannons, chrono sentries.

monk wrote:
There are 4 ways to defend a roach ling allin with a 1 gate expand:
1. Fast Void Rays: put up a Stargate right after nexus.
2. Fast Dark Templar: put up a twilight council right after your nexus. You will take damage from the roach ling attack. You also probably won't survive the earliest most allin versions. However, as long as you survive long enough, you should be able to do a ton of damage to your opponent, because he won't have an evo or lair.
3. Fast Forge into preemptive cannons
4. Preemptive chornoboosted stalker/sentry off of 4+ gateways

monk wrote:
3 gate expand is still viable. Here are a few tips to help you survive these roach ling attacks.
1. Sim city. Use your buildings to protect you so that you have as few paths to your army as possible. This is usually with 2 pylons and a forge or 2 pylons and a 4th gateway.
2. Do not overly rely on cannons. Cannons are good, but the problem is timing. Roach ling attacks can hit anywhere between 7:00 and 7:50. Even if hits at 7:50, the latest time possible, your cannon with a standard 3 gate expand will just be finishing.
3. Perfect forcefields. Pretty obvious advice, but I'm just reinforcing it.
4. Constant unit production. Do not skimp on units. Keep pumping stalker/sentry as much as you can. This is probably the biggest problem with the roach ling attack. Before it became popular, you could transition to stargate or get a quick robotics safely. While you still somewhat can, it leaves you very open to roach ling allins.
5. Reinforce with stalkers. Once you see the attack coming, making nothing but stalkers. Sentries don't do enough dps and they come with only one forcefield. On the other hand, zealots are great versus lings but you want to rely on sim city and forcefields versus lings instead of zealots. Of course, if you have only few minerals and tons of gas, a sentry or 2 doesn't hurt.
6. Do not cut probes if you don't have to. That is, you want to focus on making units, but with extra minerals keep pumping probes. You can judge how allin the roach ling attack is by the timing. The later attacks are simply pressure while the early attacks are more allin. Even if you defend a 7:50 roach ling attack, you can still fall behind, because you cut so many probes during the attack.
7. Chornoboost gateways. Again, pretty obvious one.

Teoita wrote:
If he keeps mining gas you have to worry about potential busts and 2 base tech play. Move and poke around with your probe to check for a (very unlikely) third. Get a second cannon. Get a sentry as a second unit if you want to be 100% safe, or a stalker. vs such a fast speed it's hard to get in to see what he's doing, so do your best to at least hold your xel'naga tower and see if there are any attacks coming. After your first sentry is started get a stargate both for roach defense and scouting.

Please see this thread: http://www.teamliquid.net/forum/viewmessage.php?topic_id=381719

monk wrote:
If you want to play a macro game versus a Zerg, you must scout him. Scout his tech, his drone count, his base count, his unit composition, his army size and his available tech. There are a few ways to accomplish this.

1. Robotics for Observers: In my opinion, this is the worst option, as even if you scout certain builds, you're either too late to respond to them or cannot aggressively punish greedy builds.
2. Hallucination: hallucinated Phoenix are a great way to constantly scout your opponent.
3. Stargate: This is a form of light pressure. With either Void Ray or Phoenix, you are both pressuring and scouting at the same time.
4. Pressure: This technique scouts your opponent by pressuring his base and seeing what he reacts with. You may not get to see everything, but as you get better at the game and develop better game sense, you will be able to make more and more reads from pressure scouting.

Alejandrisha wrote:
If you are planning on taking a 3rd, your first observer has to go straight for the zerg main. Look for the number of evo chambers and tech. 1 evo, a roach warren and no tech is indicative of early roach max, 2 evo and a roach warren and no tech is indicative of passive roach into infestor and a 4th base, so this observer should rest at the Zerg’s natural 4th after getting the Lair tech scout.

If you suspect a roach max, or catch wind of some kind of hydra/ling bust, the 1st observer should scout the zerg's 4th base and transition while a 2nd observer should be positioned to watch one of the attack paths. Once the 4th timing is spotted, this observer goes back to the zerg’s main and/or nat to watch for the hive transition or muta transition.

It's also important to scout the timing of the Hive, as this will tell you how long you have until broodlords come out:

Infestation Pit build time: 50 seconds, Time till Brood Lords: 4 minutes 44 seconds
Hive build time: 100 seconds, Time till Brood Lords: 3 minutes 54 seconds
Greater Spire build time: 100 seconds, Time till Brood Lords: 2 minutes 14 seconds
Brood Lord morph time: 34 second

kcdc wrote:
At Masters+, the standard way to get mutas is to take a fast third, tech lair and drop a roach warren for safety, and then start a spire after lair completes. If they're taking a fast third, then seeing a roach warren doesn't tell you a whole lot because they'll get the roach warren whether they're going roach bust, mutas or infestors.

Teoita wrote:
If you don't already have phoenixes don't invest into them. Otherwise, fleet beacon and 8-10 phoenixes or so (from a starting point of 4-6) works well. You should still invest into some other tech (colossi work great here) that allows you both to kill lings and other ground based armies, as well as to push pre hive; you should be able to afford both on 3 bases.

If you don't go phoenix, go blink before storm; storm is absolutely necessary when the muta flock grows.

Archons alone aren't enough as you need storm eventually, so ht archives should go down before the dt shrine. Going dt helps to keep in check his base count, as he will likely get a ton of expandsions behind the mutas. Archons are fine if you get them from templar with no energy, but it's not worth it to rush for them.

rsvp wrote:
There's not really an answer to this besides better micro and army control. The only answer against banelings no matter what units you have is to spread out your own units well (i.e. don't let all your zealots stand there in front and spread them out everywhere, blink stalkers back 1 or a few at a time), and try storming banelings. Morphing some archons might help soak up some of the damage as well.

Teoita wrote
Templar are really good, especially with a carrier transition, at least to keep corruptors at bay and kill infested terrans. If they aren't getting storms off it's probably because you are sending them in first, which eh...kinda silly. Having storm also allows you to storm drop which is really helpful in lategame pvz.
Storm is kinda bad vs roaches, but against broodlords i feel like it really helps, and it also protects your colossi/mothership.

kcdc wrote:
On Antiga, it's really hard for any race to take and hold a fourth base. If you can hold a fourth base in PvZ, it means you control the center of the map, Zerg has no more than 3 bases, and they can't push you out of the middle. In other words, you'd already won before you took that fourth base.

This is going to sound unsatisfying, but on Antiga, I'd recommend focusing on 2-base and 3-base strategies and doing everything you can to prevent Zerg from taking a fourth base. I think playing to limit Zerg to 3 bases is a much better strategy than trying to take a fourth base yourself on Antiga.

Compiler's Note:this advice applies to all matchups not just PvZ

monk wrote:
You must scout this in time, preferably as his first proxy pylon is building. Immediately throw up your own forge. His first cannon will finish and there is usually nothing you can do about it as long as the cannon is not in range of your probes/buildings. Throw up a cannon between your buildings and his first cannon but not in range of the cannon. If he tries to throw up more than 1 cannon, keep throwing up cannons, either matching his cannon count or going over by 1. The key is to make sure your cannons win this cannon vs cannon fight. Whoever wins this will usually win this game. If you think it will be close, you can even pull probes to attack the cannons just so you get a small edge on cannon hp, even if you lose probes. If you think you can decisively win this cannon war without additional cannons, cancel your other cannons in production.

monk wrote:
Ideally, if you scout it in time, throw up a 2nd gateway yourself and chornoboost out zealots. Your better economy, combined with the rush distance to your base will allow you to defend..Pull enough probes so that you have a dps advantage over him. For example, if he brings one probe, pull 2 probes to fight him. If you don’t scout it in time, throw up a forge for a defensive cannon which covers your entire base. Then, transition into a 4 gate.

Teoita
Get a forge and position 1-2 cannons to defend your nexus/pylons/gateway from both his zealots and cannons.

You should come out ahead as his probe count will be lower than yours. Be active with your scouting probe to make sure he doesn't take gas/gets warpgate faster than you though.

monk wrote:
Although there are additional ways, the following two guides represent two of the easiest ways to defend such an attack:

• 2 Gate Defense: 3 Stalker Rush
  
• 1 Gate Defense

Teoita wrote:
http://www.teamliquid.net/forum/viewmessage.php?topic_id=233968 works both in pvp and pvt. It's just one variation though, there's more out there.

Alejandrisha wrote:
Maps that have locations where the neutral low-ground and a main base share a cliff allow a blink-obs player to pressure an expanding robo-player. On these maps, it is known that blink is strong, so SG play is very risky, and DT play can be a strong calculated risk. Robo is the go-to safe option, and blink-obs play is extremely common either as a delayed expansion play or for dedicated pressure. See: Cloud Kingdom

On maps where a blink-obs player cannot blink into the main without engaging the natural, blink play is gimmicky at best and usually relies on a blink-in timing as soon as 7-10 stalkers are out and blink is finished. Robo play is very common here, so DT builds are rare and cheesy. Stargate play is therefore common. See: Daybreak

Teoita wrote:
By walling off either your main ramp or natural with gateways.

Alejandrisha wrote:
The most important thing to account for with observer scouting is considering your opener and how quickly it permits you to train your first observer. Take four staple Robo builds:

1 Gate FE → Gate Gate Robo
1 Gate FE → Robo Gate Gate
2+ Gate Robo
1 Gate Robo → Gate Nexus

Remember this when changing up builds often or you might miss the things you are used to seeing when using a riskier build.

Assuming 1 gate Robo opener and no delay of Observer, you see:

Units on middle of map

3 Stalkers or less - no read
4 or more Stalkers - Blink

Expansion

Expand behind your opponent and cut units in favor of economy in tech. It’s not too bad to expand behind your opponent with a fast Robo build, as FE builds in PvP cut more probes than those in other match-ups. Chronoboost out probes, and do you best to deny him scouting of your natural. Alternatively, you can chronoboost out a Warp Prism and try to kill him off the back of 4 Warpgates.

If you expand in turn, the only thing you need to worry about in this case is a Cybernetics Core → Nexus FE that expands very quickly, yet retains the aggressive potential of 5-6 units at a pretty early WG timing. As mentioned above, your primary goal is now to deny scouting of your expansion. Unless your opponent played very greedily, your Observer scout should be much quicker than his. You can even chronoboost out a 2nd observer and try to kill his on its way to your natural if you’re crafty enough.

The reason you want to deny scouting is because you can expand him, yet catch up and surpass him by spending all of your chronoboosts on probes whilst he is still in the dark and must account for 4gate, blink and WP-4gate rushes depending on his probe scout.

Once you scout your opponent’s expansion and tech route, the next important cue to look for is the gas at his natural expansion. Standard builds delay these in this match-up until an expansion is scouted (and probes aren’t cut) because probe counts are often lower in the 2 base phase, but staying on 2 gases for for too long (a good barometer is you are mining from yours and he hasn’t started building his assimilators) indicates some kind 2 base WG attack, whether or not a Warp Prism, Immortals or Blink are included, or some other kind of timing like 6Gate with one Colossus and no range. However, almost always an expansion indicates a macro Colossus follow up and should prompt you to add gases 3 and 4.

Twilight and Robo
It could be either strong aggression or an expand build. Stick around until his obs should be popping and park your first observer on a ledge so that it barely clips where his natural Nexus would be to avoid getting picked off. In this situation, it’s better to expand only if he expands first because you have a read on when he expands so you can get more probes as mentioned in the Expansion section. In the meantime, chronoboost out immortals. If he does in fact go for a blink + obs pressure, you should be able to fend it off with continuous chronoboosted immortal production and whatever gateway units you can muster. After you smash it, get your own blink to set up an attack where you will have more immortals in a blink vs blink situation as soon as your blink finishes. The higher immortal count should give you an easy win.

Twilight
If you see Twilight and no Robo, your opponent is likely going for a blink rush. CB immortals and hold the ramp with warped in zealot-stalker. Zealots are actually great in this situation because they do very well against stalkers when stalkers have to penetrate a ramp, even with blink. feel free to drop a forcefield at the bottom of your ramp before the stalkers arrive so that he can’t blink 1 stalker up the ramp suddenly so the rest can also blink on top of it.
Alternatively your opponent might be going for some kind of DT opening, wether it's an all-in or an expand build. Either way, you want to make sure his dt's do as little damage as possible, by holding your ramp with forcefields, units on hold position or a 3 pylon wall while chronoboosting a second observer. Keep your observer active and scout his followup; often times DT builds go into a Zealot/Archon bust off of either one or two bases.
Against the faster one base variant cancel your Nexus and defend your ramp with a building wall off and carefully placed forcefields: make sure they are as far away as possible from his Archon(s), which will allow your units to get shots off while he micros to smash the FF. Against the 2base variant tech to Colossus, simcity your natural with gateways like you would in PvZ and micro similarly; you have the advantage of stronger tech so as long as you take little damage from his attack you will come out ahead. Additionally, be wary of DT's being dropped in your main.

Stargate

Your response is predicated on what accompanies it. If there is no Robo, it means he’s going to try to kill you with 3 or 4 gates of units and phoenixes. Deal with this by creating obstructions to prevent his zealots from closing in on your units, probes and structures; phoenixes nullify sentries so you can't rely on forcefields. Focus on getting stalkers out, as you can fully use your economy on this composition, and these units are hardiest against phoenix lifts. On the engagement, you need to focus on killing his phoenixes; killing his gateway units and leaving those phoenixes to recharge energy is not good enough. When you allow this, you sometimes reach a fallout situation in which his phoenix count becomes so high he is able to lift entire probe lines and stalkers during flybys. Once the phoenixes are dead, you should have enough stalkers to beat his zealots.

Alternatively, scouting a robo indicates that he more likely wants to use the SG as part of a solid opener that accounts for DT and Blink builds. Just keep an eye on that robotics. Since you’re not opening blink, if immortals start coming out of that thing, he's going for a strong composition of phoenix/immortal/zealot. Expand behind him as mentioned above, and count on making his gas investment into SG and phoenixes a waste compared to your gas advantage in robotics tech/composition. If he expands, you might want to consider adding blink just to deal with the phoenixes more effectively, as losing even 2 or 3 here and there is a real pain. No more than just 8 stalkers, though, as they quickly become a liability in 2 base games.

Obs or Robo
Mirror builds. You can start teching up to Colossus. Make sure scout wether he's expanding or going one base Colossus.

Teoita wrote:
You need to buy time for your detection to complete. This can be done by holding your ramp with forcefields in front of the DT's blur, units being left on hold position on the ramp or a 3 pylon wall while chronoboosting an observer. If he does get up the ramp, make sure to pull your probes to limit the damage taken, and if he tries to kill your Robo (or Forge) forcefield around it to keep the DT's from getting in range.

Teoita wrote:
There's two ways to beat a one base colossus all-in: either use sentries to ff your natural ramp, and when the colossus comes up to break the ff focus it with immortal (which is what kcdc does), or go zealot/immortal heavy, engage in the open and try to get a flank off.

I think if you read it perfectly you also -might- get colossus up in time (i think i've seen Cecil do this on his stream but don't quote me on that), but i'm not completely sure on that. Personally i haven't played vs 1base Colossus in ages, and when i did i usually held somewhat comfortably if i didn't fuck up.

kcdc wrote:
If you go for a quick expansion with robo follow-up and you peg your opponent on the same, you can do a 1 colossus rush which is very powerful.

Basically, you skip your obs, rush out the support bay as soon as the robo finishes, stay on 2 gas, add gates until you're at 6, rally your colossus across the map and attack. It works because they're still producing probes, going to 4 gas, researching colossus range, adding a forge and +1 weapons, and a lot of times they'll have gone to double-robo. Once you have enough gas for your first colossus, I recommend going to 4 probes on gas from that point forward and building only zealots and colossi.

It's not 100% all-in because at least you have colossi, but if it fails, you don't have range or upgrades and you've cut probes. That said, I think I'm undefeated with the push (the scenario doesn't come up much).

Teoita wrote:
The most common timings used off the top of my head are:

1) Some kind of 5-7 gate blink stalker timing, sometimes with immortals.
2) A 2base chargelot/archon bust, which is not very popular since it's somewhat easy to react to it and wall off in time
3) 6gate with a single Colossus with no range (kcdc detailes it much better below).
4) Very rarely used, 2stargate phoenix after expanding. If you can deny scouting and kill the observer, this can be deadly against someone cutting units to tech to Speed Prism/Immortal or Colossus as fast as possible.

Generally 2base timings in PvP aren't very figured out because the early game is so volatile. Mostly you see them when someone gets an advantage from the early game, like shutting down gateway pressure, DT's or Phoenixes, having a significantly faster nexus that goes unpunished, or after doing damage with blink/obs harassment.

Teoita wrote:
You can throw down a twilight council when you see his phoenixes, expand when he does, and then go for a chargelot/blink/archon timing to bust him. If he doesn't do enough damage with his phoenixes and/or over invests in immortals, your chargelot/archon army should wreck him.

It's also possible to go for a faster bust off 6-7gate blink. Also, never commit to a chargelot/archon bust if he's walled off. Instead, take a fast third, tech to colossi and keep map control, delaying his third. If he moves out on 2bases, engage in the open field as his higher colossus count will melt all your zealot/archon in a choked up area.

Teoita wrote:
As far as macro PvP goes, here's a few notes:
1) Tons of Colossi is what you want in the midgame, and most builds go double robo off two bases.
2) Getting 2 immortals and a speed warp prism before the first Colossus is extremely good against fast expand/robo builds, which will cut stalkers heavily and open up room for your warp prism to harass.
3) In my experience (which might be wrong) you can take a third when you confirm that no dedicated 2 base all-in is coming, ie you see a relatively low gateway count (up to like 4) and double robo. If you see he doesn't respond to your third with his own nexus just cut probes and a bit of tech and add gates and units as you can afford it.
4) Doing a delayed all-in with two robos isn't very effective because having the defender's advantage with the Colossus reinforcement is so strong, so it's not done at all in pro games.
5) Attack upgrades are key, as being the first to have +3 attack opens up very strong timings. Armor upgrades on the other hand are very weak.
6) Lategame army compositions have changed recently, as people have figured out that it's not only about having a high colosuss count; it's more important to have several layers in your army. You want one row of zealots to tank the first volleys of Colossi, a second row of Archons, a third of 5-6 Immortals (edit: monk tells me modern lategame armies go up to 10-12 wtf :O), and finally about 8 Colossi. You should switch into Immortals after reaching a high Colossus count.
7) 4th bases are taken after maxing out, which is when people also start teching to Mothership. Trying to get the mothership before maxing is very dangerous as it opens a huge window of vulnerability.

kcdc wrote:
u do want 3 mining bases in PvP, but you don't necessarily want them fully saturated. In PvZ or PvT, you want between 72 and 80 probes, but in PvP, a few factors combine to put the optimal probe count around 60. These factors include:

-Slower second, third and fourth bases, so your main will begin mining out before you take a fourth. Probes in excess of 60 don't provide much advantage because you don't have enough mineral patches to justify the investment.

-Strong defender's advantage once both sides have a decent colossus count. It's really hard to trade effectively with attacks, and if you can't force trades, the superior reinforcements that come with a higher income don't mean much.

-Slow colossus build time places heavy emphasis on winning one big battle. You can't win fights without a good colossus count, and it takes forever to rebuild that colossus core if you ever lose a battle. This means that whoever wins a decisive battle almost always wins the game. A maxed army with 60 probes is more likely to win a decisive battle than a maxed army with 70 probes, and the latter player's reinforcement advantage doesn't matter because they can't rebuild their colossi fast enough.

Teoita wrote:
Start the Forge shortly after your support bay and before the second robo. Prioritize attack upgrades, armor is very useful. This also means that you should never go double forge.

Teoita wrote:
When you know no strong, committed 2base timing is coming your way, and you have set up your infrastructure. This means your opponent should have his natural gasses, a robo bay and a forge, and you should have the same as well as two robos. Once the game goes into the double robo stage, it's very easy to hold a third base because the defender will have a strong defender's advantage, as he will fight close to his robos.

Teoita wrote:
Stargate play is extremely weak to any kind of timing attack once it gets scouted. Both Phoenixes and Void rays are very weak until they reach a certain critical mass, which leaves a huge window of weakness for the opponent to exploit with a timing. Additionally, it's possible to react to Void Rays simply by matching the opponent's stargate count and chronoboosting phoenixes. All in all, it's a huge risk that isn't very much worth it. (san vs whoever in gstl vod)

Teoita wrote:
In PvP, it's really not worth it to spend 250/250 in the midgame on an Archon when for less than that (300/200) you could have a Colossus. When you take a third and your gas income goes up, adding archons from HT's does work out nicely though, especially if you cut colossi to add Immortals which are pretty mineral-intensive.

As monk pointed out in the PartinG-Rain OSL recap, the perfect pvp army has layers of units, ie a few Zealots in front, Archons behind, then Immortals, then Colossi, rather than mass mass Colossus with fewer support units. With the way timings and income work out, it's usually more efficient to add a Templar Archives rather than a Dark Shrine: cut Colossi at 8ish then spend minerals on Gateways/Zealots/Immortals/Warp Prisms/eventual cannons, and gas into HT Archons.

monk wrote:
Yea, another thing is that in PvP, your gas income to mineral income ratio is much higher than in other match-ups. You usually have only 55-60 probes with 6 gas instead of 70-80 probes with 6 gas as in other match-ups.

Archons are only really useful in a standard PvP in the late game. And again, as other have pointed out, a lot of late game stuff you want is actually high mineral/low gas, like zealots, cannons, gateways, pylons, nexi, warp prisms and immortals. In contrast, only archons and upgrades are gas heavy in the late game.

Alejandrisha wrote:
assume ~60 probes, giving us ~140 food to work with

8 colossus: 64 food
8 immortals: 32 food
archons: 3+, however many you can afford
zealot reincorcements: whatever supply is left

Basically, colossus will chew up all of the gateway units. Once the gateway buffer has evaporated, your Immortal and Colossus line work on your opponent's Immortals. This is the point in pvp engagements where some people lose their armies, because the Immortal back bone is actually more important than the # of Colossus in your arch. Archons are more food efficient than any other gateway mix, and they tank Colossus and Immortal shots quite well. Immortals do everything a Stalker does and better for the food cost; Stalkers on the other hand perform extremely poorly in lategame PvP.

monk wrote:

The optimal PvP army may consist of air units, but this composition is almost never able to be achieved these days in professional games because of the strength of timing pushes and of the large investment it takes to transition to air. The optimal realistic army consists of:

• 60 Probes:You generally don't need more than 60 probes, because PvP usually ends in one climatic battle. You want as much supply as possible in your maxed army, so more supply wasted in probes means more supply missing from your army.
  
• 8-10 Colossi, range 9: While it was once considered prudent to amass as many colossi as possible, we have now learned that it's more important to diversify damage among many layers of Protoss units. What this means is that we want many different types of units that fire at varied ranges to tank colossi shots and reduce splash damage as much as possible.
  Imagine a scenario of 12 colossi and a zealot line versus 10 colossi, 3 immortals, and a zealot line. The 12 colossi will eradicate a zealot line in about the same amount of time as 10 colossi plus three immortals. However, once the zealots lines of both armies are eradicated, the 12 colossi will have to waste shots on the 6 range immortals while the 10 colossi are free to work on the extremely clumped bunch of colossi who are all firing at 9 range.
  
• 8+ Immortals, range 6: One of the most recent innovations in PvP is the use of the immortal. Immortals have a funny place in PvP; they're great in the early game against blink stalkers, terrible in mid game when zealots and colossi are the biggest factors, and essential in late game to tank colossi shots. Thus, you'll often see heavy immortal production in the beginning of a PvP, and then a switch from colossi back to immortals to finish off the last 8 supply of a Protoss army.
  
• Archons, range 3(as many as you have the gas for): Big soft cotton balls. This both perfectly describes their appearance and their role in a PvP. As archons have many shields, large collision size, and an unique range, they're perfect as one of the first lines of defense against colossi beams.
  
• Filler Zealots, range 1: These are the filler of the Protoss army when you don't have the gas to build anything else. While zealots are great for dps and providing the first layer of defense against colossi beams, they are best replaced by archons as soon as you have the gas available. Zealots, unfortunately, still have much smaller collision size and many fewer hit points compared to any alternatives; thus they're the perfect targets for colossi beams. Still, you'll always see zealots reinforcing Protoss armies in the middle of any big fight.
  
• 1 Sentry: One is all you need for guardian shield and even this use is questionable.

A great example of a game where superior composition decides a PvP battle is Rain vs Trap from Proleague. Trap obtains an early game lead, but Rain's defense and his superior end-game composition decides this game.

Teoita wrote:
In competitive play it's only teched to after playesr first max out; teching to it faster leaves you extremely weak to a +3 Colossus/Immortal/Archon/Zealot maxed attack.