3: Race Specific: Terran A:
IntroductionB:
Openings/CountersC:
Unit ChoiceD: Late Game [ Never Posted AFAIK -- Knickknack]
Introduction: Terran can be a pretty nasty race to play against but in general they are a more forgiving race then zerg. Most terran players can't click fast enough to get the most out of their race and that makes things really easy on you. On the other hand, for those terran players that actually can click fast, you could be in some degree of trouble. Terran, in order to be at their most annoying, must use dropships well, must macro near perfectly, and must have very strong timing. Their goal is almost always to harass protoss constantly and then catch them overpowering. Depending on start locations, Terran may also employ a center control strategy, gradually pushing into the center and counting on his unit's ground superiority to trump your macro advantage. Protoss's goal, conversely, is naturally to get away with overpowering, or with teching to high level units like carriers and arbiters in order to regain parity in the relative cost effectiveness of his army. You do need to overpower terran at some point because it's almost impossible to launch a successful attack without a far bigger army against terran, unlike against zerg. Terran are just naturally way stronger on defense.
Remember that much of terran's goal, and much of his advantages, rely on him being able to dictate the flow of the game to you. Never was this more apparent then Nada's 3-0 romp over the very strong Reach in kpga finals. By Nada getting off to an excellent start in all three games, he was able to dictate the play and force the battles to be fought on the ground of his choosing at all times. Terran can always come back from a poor start against protoss. Not so the other way around.
Upgrades are not as important against terran. In fact I rarely bother to get more than attack at all. They simply don't make much difference, especially armour and shields, because terran units are doing damages like 110, 70, 20, 30 to you, so doing 1 point less means almost nothing. Attack upgrades are a little helpful though. In general, if the terran is upgrading vigorously I'd upgrade attack with 1 forge along with him. The bad news is that upgrades do help terran more then they help you, so 3-3 terran is quite a bit better then even 3-3-3 toss. Blizzard didn't take terran metal into account when they balanced protoss vs terran upgrades, because they thought terran would use infantry mainly with metal only as support. Where upgrades are really necessary is your carriers. If you're making carriers at all, or if you ever think you might want to, start upgrading air attack at once. Since interceptors do 6 damage each, getting that +1 every time actually is a pretty big difference. It's kind of like getting +8 per carrier. Also you just need to get it to do anything against heavily upgraded terran ground, since you'll do only 2 damage to 3-3 goliaths without any upgrades. At that rate, 1 carrier can barely take out even a single goliath.
Psi storm and micro in general is a lot less important against terran. You do need to micro any early pushes he does well but aside from that micro is not nearly as important as having good strategy. Mostly you'll just be throwing units in his general direction and hurriedly rebuilding them while they fight. Psi Storm can be helpful, but only if it's a stagnant sort of game. If terran is one of those types that thinks he's got enough power and is just producing units to eventually push out and kill you with one massive strike, making templar is a good idea. They'll be excellent to use in one big battle, after they've had lots of time to build up mana. On the other hand, if terran is harassing you constantly and attacking you in different places constantly, your templar will not often get a chance to build up mana and you will not often get a chance to micro them. Luckily for you, that is also in your favour. Unmicroed toss vs unmicroed terran is toss's advantage. Furthermore, microed toss vs microed terran can also be toss's advantage, provided you have the right units, especially templar. For more on this, see section 1i.
Macro, on the other hand, is absolutely critical. It's usually only through sheer brute force that you're going to beat a terran in the middle of the map. So make damn sure you have enough gates. This can be more difficult against terran because in general you're going to be expanding a lot more. Just remember to always have 4 gates per mining base (3 is ok for a mineral only) and you'll be fine. Also mineral onlies are a fine expansion to take against terran because they are actually easier to defend then mains and naturals (due to being in the open, being closer to your base, and being less threatened by drops) and because you don't need as much gas, just plain zealots and dragoons can do fine. Finally if you ever find yourself in a bind where you just can't spend your money because you have too much gas and not enough gates, mass dark templar. Many terran players just can't believe how many dark templar you can have if the money starts to build up. Many is the game that ended for me when I attacked terran with 20 or more. Build a couple spare shuttles and load them up, drop as many dts into his main as you possibly can. Because of their extremely high damage you can kill a lot of his buildings real fast, especially his turrets and scanner, leaving him pretty much high and dry. Then you can attack after a few seconds into the front of his base with the rest of your dts and other units and do still more damage.
Which leads to my next point; a shuttle is never a waste of money against terran. In fact, not having shuttles can lose you a lot of games. Once terran is already on your cliff, it's too late. You need to be able to hold your cliffs all the time--not just once he's already entrenched up there. The reality is that by dropping goliaths or making turrets, he can turn your natural cliff into an unnassaultable position. Having shuttles gives you options, and it takes them away from terran. I would advise in general that in any long game, you should research shuttle speed at some point. You should have at least 2 and as many as 4 shuttles even if you can't see an immediate use for them. The problem with shuttles is that they are expensive and they take a long time to build. If you suddenly see an opportunity to drop terran, or you suddenly need to hold a cliff, or you suddenly notice he's taking the islands, if you don't already have shuttles, you won't have them in time. Unless you are in a game where there is constant fighting and you need every single mineral to build troops to hold onto the middle of the map, making a shuttle is a good investment. One thing that is sometimes overlooked as well is that shuttles are good in a center fight. 4 shuttles loaded with zealots and dts coupled with a ground force of mainly dragoons is an excellent army in a variety of circumstances. It has several key properties: 1, it won't be vulnerable to goliaths because the goliath ai will priority target the dragoons on the ground, so long as you send in the goons first. 2, it somewhat disguises your true strength, as terran can never be sure just what you have in the shuttles. 3, it makes your melee troops far more effective, as they can get into melee combat without having to trudge through tank fire first. 4, it makes his mines a death trap to himself. 5, it is far more compact and mobile, allowing you to project force onto any place in the map any time you need to. The only time I would not recommend shuttles is when dropping onto a large number of turrets, since turrets, unlike goliaths, cannot be distracted by land forces.
For an example of how deadly and how useful large amounts of shuttles can be, try watching Liquid Nazgul's games, he uses them quite liberally against terran.
Openings and Counters: Here is a list of openings and counters against terran. The protoss vs terran match is very much like the zerg vs protoss match. Your goal will always be to sieze the initiative early and play aggressively. Try to keep the terran conservative beyond his opening, try to keep him from attacking you and try to keep control of the center of the map. Your goal is not so much to win the game, as to keep the terran from winning. Do not be primarily concerned with landing a finishing blow, be concerned with keeping terran from expanding, and with taking all of the resources in the map for yourself. So once you detect a terran opening, switch to aggressive play right away. Against zerg you must stay conservative until you are absolutely sure you can land a finishing blow. The opposite is true against terran, you must take risks or you hand him the game. If both players stay conservative, terran will end up winning. The same is true of zerg against protoss--zerg must play aggressively there too. The difference is that it's a little easier on zerg. For protoss, the aggression must be very smartly done, because any mistake will get punished.
Fast Dropships: This strategy can go a number of ways. He can load up siege tanks, vultures, or a combination with his first drop ship. One thing is the same though--terran should always at least try to attack with that first ship. I rarely see a terran wait for a second ship, he will almost always attack with the first one. All you have to do is be ready for it with a few goons. To defeat this strategy, knowing it is coming helps enormously. I will go over what steps you should take for that later on. As for what you should do after the drop, if you lose more then 8 probes, you should probably leave the game. Assuming that you kill his dropship and tanks and lose no probes, you can safely expand, and if he is expanding too, you can safely expand twice, to your natural and to a second main. Assuming that neither of you lose anything much (the most likely arrangement in my opinion) then you should expand right away and queue up a shuttle. Continue to produce shuttles and observers constantly. Your first emphasis should be on shuttles though, so scout the map without observers at first, using probes or zealots. One trick I occasionally employ, depending on the start locations, is to build pylons along likely routes for his dropships. Pylons are cheap, tough, have good vision range, and are needed anyways. Furthermore, since he's going drop ships it will be a long time before he can spare the ground forces to destroy them. Terran has a number of options after that. He will most likely expand, and use the two tanks in his ship to safely defend it. He has a choice between staying with dropships, or switching to a mainly land force. Try to use an observer to get a sense for which his intention is. If you feel he will be staying with drop ships, keep producing shuttles, trying to have as many or more shuttles as he has dropships, while also teching to high templar, the most effective unit to defend against drops, and speedy zealots and dts, the most effective units to drop. Don't be worried that you are sacrificing ground units too much, because those ground units won't be too much use anyways as he is dropping your cliff. If you suspect that the game will be about controlling cliffs, be sure and get a robo fac for shuttle speed, or, if you think you can afford it, start teching to carriers right away.
An alternate way to go from here, if you scout that he expanded very quickly after failing his drop, is to build up a decent land army with a single shuttle and attack him. Many is the terran who will be trying to hold his exp with 3 or 4 tanks when you can attack with a dozen goons. With proper micro you can slap him down smartly.
If he is switching to mainly ground forces though, hold off on the templar and don't make too many shuttles, two or three is enough. Mainly you want to concentrate on producing enough ground units to hold the middle of the map while you stay ahead in expansions. If you go high templars and mass shuttles against his mainly ground force, you won't be able to stop him from marching to your front door and seiging you, and doing so with the aid of enough turrets to stop your shuttles from being much use.
Early Push: If he is early pushing you, there are a few steps to take to minimise the damage. First you should not expand right away. Don't be concerned that you are not getting ahead of him at first because you should still be able to expand ahead of him. If he is pushing you, he isn't expanding himself. You just have to defeat his push before you expand. You should try to make him waste mines in the middle of the map with your goons if at all possible, without losing any goons of course. Losing a goon at that stage of the game can be fatal, because unless terran screws up and similarily wastes a unit, he will be ahead of you. Slap down additional gateways right away, and make a shuttle if you have a mineral build up. Produce some zealots if he doesn't have many vultures, even if you don't have speed yet, because in small numbers of units, zealots can be reasonably effective. Also if you are making a shuttle, you obviously want zealots to put in it. Retreat your initial goons away from your main and try to use them to cut off reinforcements. Once the terran sets up his push at your min only, move your initial goons to the center of the map to ambush reinforcements, or use them to flank the push. Loading zealots into a shuttle takes away their slow speed disadvantage and helps out a lot vs tanks and mines. In general I would say that the protoss player should be able to defeat the push and expand to his natural first if he doesn't make any big mistakes. There are a couple of different flavours of early push.
Bamboo Rushing would be to bring his first tank and vulture along with 2-4 scvs and 4 marines, and try to build a bunker in your natural, then build an ebay to float to your ramp, then start producing turrets. This push will fuck you up if his bunker completes. If you pursued a robotics before second gateway build, as I usually do, you are at a definate disadvantage. If you do not have the units to stop him from getting that bunker done, don't waste them. If at all possible, keep them somewhere safe but still on low ground. Have at least 4 gateways producing dragoons, and zealots if you can afford a shuttle or if he doesn't have vultures. Also be working towards getting speed. Don't attack his push until you have to. In order to make his push work terran will probably have sacrificed scv production, so you'll be ok economically. If you sense that you can kill his push at any time with your sight range upgraded observers hiding on top of cliffs, do it because it means he expanded. If he keeps putting shit there, take an island with your shuttle. When his tanks get to within range of your units on the ground, you have no choice but to attack him all out. Depending on how soon this happens you should have anywhere from 4-6 gateways producing non stop speed zealots. You'll have a big gas build up, use it up on DTs for your next waves. Hopefully the first attack eliminated some of his turrets, but in any case once the units on both sides are reduced, the effectiveness of dts skyrockets. Also you'll be able to easily contain him assuming you win the fight, since he'll have to slow push out with turrets once your dts are in play. Depending on your relative economies at the time you'll want to take your natural and possibly your mineral only as well. If you break out of a bamboo rush, odds are good you'll win the game. Too bad that's so hard to do.
He can also go for a straight metal push of tanks and vultures, with scv backup to repair and produce turrets. This generally comes a little slower than the bamboo. Just pick your units according to his; more goons if he has vultures and more zealots if he has tanks. This rush is a lot easier to beat if you went robo before second gate than the bamboo rush. From terran's perspective though, it's not as risky, because many toss's will 2 gate and sometimes even range upgrade before robotics, which eats bamboo for breakfast. This rush will quickly turn into an expansion if terran senses you have the units to hold him, but your own expansion will still, by necessity, be delayed, since you must ensure that you can hold his push before you make a nexus. So be careful of a terran showing you an early push then turning it into a macro build while you're digging in. Why won't a terran turn a bamboo into a macro build? Because he will already have sacrificed a lot of his economy by bringing a significant number of his scvs out while simultaneously stopping or slowing scv production, so trying to then outmacro you would be foolish.
Sometimes terran will come at you with only vultures, hoping to get into your mineral line. All you need to do is make sure you have a goon or two at your ramp holding position, so he can't climb it after running around your first set of goons. Then have an observer ready to spot his mines and catch him in a pincer. Then make a forge and expand right away. As your nexus is finishing up you should have a couple of almost completed cannons. You need 2 cannons minimum to hold vultures when goons are not present, keep that in mind. If he expands himself, check with obs to see if he can hold. If he can, you can expand again. If not, hurt him with your goons and then expand again later. Bottom line is that if his first vultures don't kill a lot of your goons or get into your mineral line, you are really far ahead. If he does get to your probe line, hopefully he only has a couple of hurt vultures left. Bring your probes off of mining to attack them while your goons try to catch up. He'll lay mines to try to make your goons blow your probes up, but if your probes are attacking, they will destroy the mines before the mines blow up almost every time. It only takes four probe hits to kill a mine, so you'll be ok 95% of the time.
Early Expansion: This build had largely been replaced by the early dropship and expansion combination in early 2003, but it is done more commonly now. It gets the terran's natural up and running long before your own unless you did some kind of bizarre nexus first build. There are two counters to this build. The first is to try to hurt terran right away before expanding yourself. This can be done with a dt or reaver drop, or with a very well executed elevator. Your hope is to damage the terran so severely that he either leaves the game, or in the confusion you can safely expand twice yourself and get ahead of him. The other option is to expand twice right away, foregoing any offence. There are dangers to both, obviously. If terran defeats your drop without getting hurt much, you're dead. If terran times his attack well enough, he'll kill you before your two expansions can create enough mass for you. This terran strategy is unique in that it doesn't rely on deception or surprise to succeed, it relies almost only on the basics. Whoever's macro, micro, just their solid skills overall is better will probably win, so you know that if a terran is going for a plain old fast expand build, he could telling you that he thinks he's the better player. There is a third possibility that is gaining a lot of momentum with terran's today as well. This would be to stop scv production relatively early on in order to push faster. This catches a lot of toss's with their pants down. The best way to counter this threat is to not take your min only right away, but to just match him in expansions and rely on your superiour probe count to stay ahead. When the attack comes, you should have more than enough units to smack it down and then safely expand while containing him. The bad news is that if he has not stopped his scv production than you will not be far enough ahead of him in macro to win a slow push macro based center fight, forcing you to try to tech to carriers and arbiters to get back into the game. The only real thing you can do is try to get range upgraded observers to give you an idea of whether or not he is stopping scv production to push soon or whether you need to expand a second time. It's not an easy call to make even assuming you do have good scouting info. Sometimes Starcraft can be a bitch.
There are two main variations on the standard goon/obs pvt opening. They are gate-core-robo-gate and gate-core-gate-robo. The difference is subtle but important. If you go second gate before robo you are going to have many goons and range upgrade, but late observers. If you go robo before second gate you will have excellent scouting but not enough goons to easily stop an early push. There's no way to know what you should open with if your scouting probe isn't perfect. Most of the time you'll have to rely on intuition and on past experience with that player to know how you want to open. If you go robo before second gate you will be able to scout a drop before it happens or a fast expand before it's too late to do anything about. If you go second gate before robo you'll be able to easily stop any early push, but face big risk in getting screwed over by dropships or a fast expand. In general it's usually better to robo before second gate if the spots are long or if your scouting probe can't get into his base or dies early. It's usually better to second gate before robo if the spots are close (and you're not going for a trick strat) or if you have good reason to suspect an early push, or if you are playing on Gaema Gowon or Bifrost (in which case you can climb a ramp to get to your cliffs if he is going drop ships, rather than having to wait for shuttles)
Here are a few alternative openings designed mainly to catch a terran player off guard, or if the spots force you to play desperately (ie 12/3 temple).
Dt Drop (6/9 or 12p/3t): This build sacrifices an expansion to launch an early offensive on terran. It can be decently effective against each of the terran openings, but it can also go bad very easily. Basically you get a robotics before making a goon, then start your dragoon, then a citadel. You make a shuttle when the robotics finishes, hold your ramp with your goon, and make an archives when the citadel finishes. Add another gate as well. Your archives and second gate should finish together, start 2 dts, and load them into the shuttle. If he is dropping or pushing you, use the dts to defeat the attack and expand. If he is expanding himself, attack with the dts and expand. Your goal should be at a minimum to kill some depots and force him to make many turrets, but it can kill a few of his units and scvs and stop him from mining for a while if you're lucky. This build will also protect you from any cliff drops very well. Caveat: if terran has an scv in your base build a goon to get rid of it to the exclusion of all else, and if the scv sees your citadel and your robotics, cancel the citadel as soon as the scv is gone, then expand. Hopefully his main is turreted and defended up the wazoo while your natural is coming on line nicely.
Hidden Robotics (12/6 or 3/9): This build aims to decieve the terran completely. You should first display to the terran an intent to expand with dragoons by moving 3 or so to his ramp in a containing posture. Meanwhile a probe is building a pylon just outside of his base in a hidden location, and then progressing to reaver tech from that pylon. By displaying an intention to expand to him, you disguise your actual intention of dropping him. In all likelihood he will leave his base to attack you somehow and neglect his own defenses since such an early expansion would leave you very vulnerable and unlikely to mount any counter attack. Just as his push should have chased your goons to your natural, you should be ready to drop a reaver on him with your hidden robotics tech. With luck, his attention will be focused on his push, making your drop that much more effective.
Double Robotics (12p/3t): This build aims to overwhelm the terran's defenses with a drop of unnexpected power. You build a goon to block your ramp and a robotics. As the robotics nears completion, you should be able to start your second one. Make a shuttle when the first robotics finishes, then make a support bay. Make sure you have 43 supply ready for when your shuttle finishes completion even though you'll probably only be in the high 20's. If you only have 35 supply you won't be able to build both reavers at once. The support bay has a low build time and should finish when the second robotics and the shuttle do. Then build 2 reavers at the same time. Make sure to separate the two robotics from each other so that if for some reason he scans you he will only see 1 robotics. Make sure to hide the shuttle in a corner so that he won't see your shuttle if he scans you and assume that you are going observers as always. When the two reavers finish, pick them up and proceed with your drop. Most terrans at this point will be ready for a single reaver, but few are ready for both. Therein lay your chance to cripple him. Microing those reavers is critical. Unless he does not have any turrets at all, do not try to drop in his mineral line. If you see a turret, go somewhere else or drop out of its range and kill it. Your aim here is not to kill his scvs, as you might think, because that would be way too greedy. Instead you want to kill his tanks and vultures and outlying structures, which is not only easy, but just as effective. Drop out of range of his sieged tanks and turrets and kill whatever is available. Terran will unsiege his tanks, trundle their slow asses up to where you are, and siege them up again. Just chuck 2 scarabs at him while he does that and there goes a tank. A lot of times terran will come at you with vultures and try to lay mines next to your reavers, never realising that a scarab can kill a vulture and a mine, and any other vultures and mines that are nearby, all at once. Even if he has 2 dozen vultures theres no way he can kill your reavers by coming up and laying mines beside them, so you can happily sit there and kill vultures to your hearts content. If a siege tank safely sieges up within its range of your reavers, but out of your reaver range, simply pick them up and head elsewhere, repeating the process and snagging yourself a few more free kills. This practice slows down terran loads, forces him to erect turrets everywhere, and almost always allows you to safely expand to another main (which you want to do if it's 12/3, which is when you'd normally use this build)
Speed Zealot Elevator (6/9 or 12t): This is one counter to a fast expand build. Basically you progress as normal but delay your own expansion a bit to research speedy zealots and get a shuttle. You also switch to zealot production after your second or third goon. This build works because most terrans, when playing defensively (as in to hold their expansion) rely on tanks in siege mode, not vultures. For whatever reason the tank just seems more defensive. So your zealots are a good choice. As soon as you have a shuttle and zealot speed researched you should have as many as 8 zealots. Use the shuttle to elevator the zealots into his main. To do this, position all 8 zealots and the shuttle below his main cliff in a spot where he won't notice you, and load up the first four zealots, drop them onto his main plateau, but keep them out of his sight. Only when the second group of zealots are on his cliff should you attack, hitting him with all 8 at once. Rally your reinforcements to the front of his base, and hit his expansion after a couple of seconds have gone by from the start of your zealots attacking his main. That should give him time to move his troops to the back of his base leaving the front open. If you see a chance, make an archives and switch to dt production. It may be possible to kill all of his turrets with your zealots, and his scanner too, leaving him very open to a dt drop follow up. One variation of this opening specifically applies to that, calling on you to wait for 4 dt's before dropping in the 8 zealots and using the 8 zealots to eliminate all detection, largely ignoring the other defenses. You can then drop in the 4 dts and do incredible amounts of damage.
Double Elevator (6/9): This build is one that I use about 80% of the time if the starts are 6/9 or 9/6. It is extremely easy to pull off from there, and as a result I win about 95% of my games with this strat, even against players that I know for a fact are better then me. Basically you proceed as normal but make 2 shuttles and 3 gateways before expanding. You should also tech to speed zealots quickly and add dts later on. Use the two shuttles to ferry all of your troops, as soon as they are made, into his base. This build defeats dropships, early expand, and early pushes. The only thing that beats it is if the terran plays totally defensively, or if he is building towards a gully push. In those cases his concentration on building turrets and mining drop zones will probably keep your drops from doing damage. The key here is not to expand until the last second before your main minerals are mined out. You will be able to keep the terran reeling until you expand, but if you do it will give the terran player a chance to make enough turrets and mine enough to block further drops. If he is still alive at that point (which is rare) you have to switch to a standard strategy and should probably use his fear of your drops to cover an extra expansion and start powering like crazy.
DT Rush (3p/12t): This build is a straight up oldschool double gate dt rush. Just go gate, core, citadel, gate, temple, and make only a few dragoons. As your temple nears completion, have both gates and enough gas and supply ready to produce 2 dts at once. Send the first one in with the second trailing behind a little. If the first dies to mines before getting to his main, hold the second one back. If the first gets to his main but there is a turret there, hold the second back. If the first one is beating on his depot without dying, send the second one in and the rest of your forces too to try to get through his wall. If you do, a lot of times that's it. If not, no biggy. While you were doing this, a probe was already expanding at another main and building a forge. You sent a goon with that probe since you'll need it if the terran has a marine, scv, vulture, or mine there. If he does have a mine blocking your nexus, keep the probe a safe distance back and walk into it with your goon so it explodes, then block the ramp with your goon so your probe can build stuff without getting killed by a single patrolling vulture. By the time he's ready to send a vulture hit squad, hopefully you have cannons up.
Reasoning behind the posititioning choices: Double robo is only effective in the 12p/3t or the 6/9 starts because it's critical that your shuttle have a short distance to go to pick up more reavers to keep up the pressure before he can make uber turrets. So why not 3p/12t? Because the front of his base will be facing you, which will be well defended and not vulnerable to a drop. You'll have to go all the way down around the back to do anything, which quadruples your travel time over the 12p/3t spots. DT Rush, conversely, is only effective in the 3p/12t spots because it relies on you being able to move ground forces around without shuttles. Obviously you can't really do that when he can have tanks overlooking your ramp. The Double Elevator really only works from the 6/9 spots, rather than 12p/3t, because it comes so much slower than the double reaver. By the time you're ready to double elevator him, the spot that you would normally be double elevatoring into is already turreted and tanked up, and you have 0 cushion to fall back on if that fails. At least in 6/9, if your double elevator is not feasible you can still expand and hold. In 12/3 if your first attack doesn't hurt him, you're fucked.
That is all of the aggressive builds that I will ever use against a terran. Now let me go into a little bit of detail on my scouting probe. If I get my probe inside his base before the wall, here is a list of things I do to piss him off.
1) Teh Mannar Pyron. This misspelled offensive was created by Smuft, another protoss user from my hometown of Vancouver. If the terran is at 6 or 12 on most temples, you can execute this. Basically you build a pylon in his mineral line across 2 mineral patches in a way that traps 2 or 3 mining scvs and forces other scvs intent upon mining those minerals to go around to the back of the minerals, greatly slowing down his mining time. In reality, if he uses the 2 or 3 trapped scvs and 1 or 2 marines to attack the pylon, he will only have lost a little more then 100 minerals in mining time, putting him barely behind at all considering the 100 mineral cost of the pylon. But psychologically this attack is always worth your while.
2) The Factory Block: This is another way to annoy the enemy by building a pylon beside his factory in a way that prevents him from making a machine shop. He will be forced to lift off in order to make his machine shop. At that point you can put your probe underneath the factory to keep it from landing. You can delay tank production by several seconds using this technique, allowing your first harassing goon a lot of time to fire on his wallin. You can then make him waste money and mining time repairing the depot, and perhaps kill an scv or two. Almost every terran now makes at least one marine to stop this since it's so effective.
3) The Wallin Block: This technique involves building a pylon where his wallin would normally go, which would allow a zealot to get inside his base. Often times terran will react to this possibility by building a bunker beside the pylon, anticipating a zealot rush. You can then cancel the pylon at the last possible second. Why was that worth it? Because terran had to waste mining time on the scv that built the bunker. If you did make a zealot, perhaps because he randomed, and terran does fail to make a bunker, then your zealot rush could actually be cost effective as well.
All of that is in addition to the standard attack scvs that are building things. Also note that if terran is producing marines out his barracks, none of those tactics will work too well. Most likely you'll just be concentrating on keeping your probe alive. Those tactics are mainly for games where terran doesn't want to make any marines at all, or at most a single marine just to chase your probe around.
Unit Choice:vs pure vultures: ranged goons/obs
vs pure tanks: speedy zealots/dts/possibly templar (dts depending on how much gas you have, how many turrets the enemy has, whether he has a science vessel, but against pure tanks usually dts are safe bet and can often pay off big time if he has no detection and scans late. Templar depending on the flow of the game. Sometimes in pvt the attacking is constant and all over the map, and a templar would never have the chance to even get enough mana to cast a storm, nor would you have the time to micro one effectively. Other times it's the opposite, both sides mass for minutes at a time before leaving their base with huge armies. In that case templar are invaluable)
vs vulture/tank: speedy zealots/ranged goons/possibly dts/possibly templar (if you already have a second gas expansion switching to carriers is certainly viable). You want more dragoons if he has more vultures, more zealots/dts if he has more tanks. Also remember that in general dragoons are more defensive and zealots/dts are more aggressive. If you're looking to finish off a weakened terran or take out an expansion attempt to a second main, massing pure zealots and dts in the face of anything but pure vultures is feasible.
vs vulture/tank/goliath: speedy zealots/ranged goons/probably templar/possibly dts (in all but the most chaotic of games you want templar when the terran starts to show large numbers of goliaths. Goliaths are as effective against zealots as vultures are (not counting mines) and more effective against goons (even counting mines) but they are more vulnerable then tanks or vultures to storm so you almost always want some templar. Also templar make great company for carriers. They are invaluable in adding firepower against a goliath or bc army.)
vs m&ms: ranged goons/speed zealots/templar/possibly dts (in small numbers, dt's rape even with detection, but once terran starts to get 10+ marines, dts just drop like flies)
vs m&m&f: ranged goons/templar
vs infantry/tank: ranged goons/templar/possibly speedy zealot/possibly dt (you'll want dt's if he proves the least bit lax in his scanning, as dts can maul infantry in a hurry with their one hit kills. If the enemy has no bats zealots are worth their while.
vs valkyraes: assuming island maps, carriers pull their weight against valks.
vs wraiths: sairs are a must if you're going carriers on a land map and he counters with cloaked wraiths. Otherwise your carrier force is a waste.
vs bcs: templar are great, so are speedy scouts if you can get them in decent numbers in time (2 scouts to a bc at least). Sairs and Carriers are largely inneffective in a straight fight, although well microed carriers in conjunction with arbiters and templars are pretty good. Spend the excess money on goons.
vs valk/wraith/bc/sv: against a terran air superiority force I think toss is a little outgunned except for 2 things: templar and arbiters. You want the standard mix of sairs/carriers but include arbiters and templar wherever possible. Stasis and psi storm give toss the edge against the superior terran air force, even in the face of science vessels/lockdown