There are many pvt guides (see end for list), but none of them cover builds and detailed timings well. My guides are meant to fill that void. I assume a more or less standard style map for this guide.
See my PvP guide
And the PvZ bible
This PvT guide is meant to be a third piece.
It is popular to divide SC up into strategy and mechanics, the mental and physical parts. SC is a very demanding game on both accounts. All of the things I am going to talk about in my guide are on the strategy side. I will be covering know-that knowledge, not know-how knowledge. Knowing-that is knowing that you should play a certain way. Having know-how means that you can pull it off in practice. More specifically, I will be covering builds. Nevertheless, the know-how of execution and mechanics are a large part of SC. Your ability to carry out the strategy and actually play well requires a lot of hard earned skill through practice. The best way to improve is though deliberate practice, which is when you concentrate on actively trying to go beyond your current abilities (including knowledge).
How does this guide work? Day9 has said about his podcasts that he tries to explain the theory behind things so that you have the right mindset to work out solutions on your own. I think that I great, but my style has been more to show you the end, and then explain how I got there. This way, you have someting you can use now, and hopefully you begin to understand how the process works.
Part 1 of this guide deals with pre-game strategy, which means opening builds and ideas. I talk about efficiency and lay out some very good builds for you to use. Part 2 deals with in-game & post-game strategy. This is adaptation, where you take your pre-game strategy, put it to the test, and modify it as necessary. This guide is meant to get the reader initiated in both parts of the game. Part 3 includes resources such as game replays, other pvt guides, and answers to questions.
How should you make use of this guide? This guide is very dense, so I suggest taking your time to work though it. Read it thoroughly and asking any questions you may have. Practice the builds a bit in single player and make sure your timings match so you know you have good efficiency down. It is very important that you do this so that you demonstrate that you can play with good efficiency, so that you know you lost for other reasons. Though reading and practicing my builds I hope that you begin to have a greater understanding of good efficiency, because making how you play more efficient is a huge part of playing well. Then head out to meet opponents, learn, and adapt.
There are many ways to play PvT, but it was pretty easy to decide on a standard BO. I decided that one gate range to robo, then 2nd gate and take a quick expo is about as plain vanilla as it gets. That means just goon and ob, no dts or reavers. It's the middle of the road build. I am not saying that this is always a good way to play, which I will get into more later. The build has its weaknesses, and is not a strong choice sometimes. The purpose of choosing this as the standard is to have a good basic starting point for PvT, and to show good PvT macro though this style of play. I have also included many other good opening BO variations on the standard style for you to try right here if you want to do something different or examine timings:
+ Show Spoiler +
Other PvT example BOs:
1) Slow Reaver with range/ob BO
8p, 10gate, 12assim, 14core, 15p, 17goon, 20range, 21g, 24p, 25robo, 26goon, 30pylon, 31goon, 34shuttle+observetory+pylon,
37goon, 40robo bay+ob, p+zealot, 45reaver, zealot, Nexus (6:00)
5:30-5goon, ob out soon
*You can skip the ob for 20sec quicker reaver.
Example rep: G5 vs Bisu @ longinus
2)Quick reaver without range/ob BO
8p, 10gate, 12assim, 14core, 15p, 17goon, 21robo, 21goon, 24pylon, 25goon, 28pylon, shuttle, bay+zealot, 34 reaver, p, 39zealot, observetory+range, goon, ...
6:20 5goon/ob/2z/1reavershuttle range almost done
*What is the difference between this and Slow reaver BO? Basically robo timing depends on T scv scout. You don't typically want terran scouting that you are going tech, but if they don't have a scout on, go for the quicker robo.
Example rep: Bisu vs leilei @ python.
3) Ground dt BO
8p, 10gate, 12assim, 14core, 15p, 17goon, 20range, 21g, 24p (can proxy this pylon), 25citadel, 26goon, p, 30gate+archives, goon,
2dts, p, nexus(5:30)
*This is just a very basic constant probe to dt build. Like reaver, If you want faster dt it basically comes down to T scv scout. If they don't have a scout on you, you can get away with a quick citadel before range. You could also try to proxy 2nd pylon to build citadel and hope they don't notice.
Example rep: Nal_Ra vs. Nada @ Luna
*Also see #8 for a ground dt buid.
4) Dt drop BO
8p, 10gate, 12assim, 14core, 15p, 17goon, 20 +1air weapon if T scout (cancel after T scout dies), 21pylon+goon, 25citadel+robo, 26goon,30p, 31gate+archives, goon, shuttle, 2dts, 40p, ...
5:45-4g/2dtshuttle (might want to build zealot instead of goon to drop with dts)
*if you make citadel asap after 1goon you can have 2g/1z/2dt shuttle at 5:20
Example rep: Legend vs Strelok @ RoV
5) 10/15 gate BO
8p, 10gate, 11assim, 13core-scout, 15gate, 15range, 15goon, 17pylon, 17two goons, 21pylon, ...
3goons with range at T nat ~4:25
5:00 - up to 7goons out
*after 3goons you could decide to tech and do citadel or proxy/robo. If not, after 5goons you typically want to make a deicsion to expand or if your attack is looking good keep making goons.
6)1gate range expand BO
8p, 10gate, 12assim, 14core, 15p, 17goon, 20range, 21g, 24p, 26goon, 29nexus(~4:10), 30goon+p, ...
Example rep: Oysten vs hobang @ colosseum
7)1gate range-robo(ob) expand
8p, 10gate, 12assim, 14core, 15p, 17goon, 20range, 21g, 24p, 25robo, 26goon, 29p, 30goon, 33observetory+goon, 35 Nexus (5:00), goon+ob, ...
Example rep: Best vs. fanta
Example rep with speed reaver followup: Nony vs. Yosh
8)Other two gate builds:
There are a range of 2gate goon builds that are not as agressive as 10/15, but not as weak as 1gate. The two worth mentioning:
20range->20gateway->20goon->22p has up to 6goons out at 5min.
20range->21goon->24p->25gateway has 5goons out at 5min.
*I will show you two good BOs from this opening:
Timing for attack is 6:20 with 6g/4zShuttle or if you build observatory+ob ~6:50 8g/4zShuttle/ob.
2) 2gate goon to quick dt
Around 5:30 you have 5g/1z/2dt.
9) Fast expansion with range BO
8p, 13Nexus, 14Gate, 15gas, 17core, 17Gate, 17z, 21p, 21gg, 25range (done ~5:35), 27p, 27gg, 33p, 35gg
*At around 5:20 you have 6goons/1zealot
Example rep: similar to Pojuk vs Bratok @ Andro
- Clearly I have not accounted for every kind of build/strategy. I am painting broad strokes here, and there is variation left open based on game circumstances. Also, I am missing builds such as 1proxy gateway that is mostly done on destination at times. While these sorts of builds deserve mention, they are not done terribly often, and I do not know enough about them.
Example of standard PvT BO
8p, 10gate -> scout, 12assim, 14core, 15p, 17goon, 20range, 21goon, 24p, 25robo, 26goon, 30p+gate, 31goon, 34observetory+p, 36twogoons+ob, 42nexus(5:30), 43gg, 47pylon, 49gg, ob, 55p, ...
+ Show Spoiler +
An important thing to note in SC is that every choice you make has an opportunity cost. The opportunity cost of a decision is all of the other possible decisions you give up in order to make the decision you made. What this means is that you could have done something else rather than what you did do, and this carries a cost. This is complicated by the fact that you usually don't know what choice is best. For example, for this BO I pause goon production after producing 6 to make a nexus. It is safer early on if you don't pause goons and make 8 before expanding. But, sometimes you don't know which way to go. If you make more goons then you don't have your nexus done as quickly, and thus your macro is not as strong later. Trade offs are everywhere in starcraft, so start noticing them and what they mean. While I aim to provide good timing for standard PvT ultimately the end is in realizing how everything relates and how doing one thing rather than another changes your possibilities and timings.
Standard PvT BO unit timings
5:30- 6goon/ob/nexus started
10:00 ~140 psi, 14g/18z if going mass probe/unit.
Or tech/cannon route:
10:00 ~125psi, 14g/14z, 3rd secured with cannons, +1started, tribunal started.
Example build replay
I will briefly I will cover the main aspects of good efficiency that go along with above rep:
+ Show Spoiler +
- Build probes constantly one right after another unless otherwise decided.
- Rally probes close to mins and catch them on the way and send to mins. As the game goes on and multitasking increases some people prefer to cancel rally so that probes are right bellow nexus and don't move towards mins. Try both methods and see what works for you. Ultimately though, for best efficiency you should rally close to mins, and give the new probe a command to gather on its way.
- You can use shift to create a waypoint back to minerals after you build something. For example, b+p+click where pylon goes then shift+right click on mineral patch. This goes along with don't leave probes laying around idle.
- Have a good split. send all 4 to different min patchs. Usually send to closest mins, but sometimes closest mins are not the fastest mining ones.
- There are some minor things such as not sending 3probes to gas all at the same time. If you are making a building and want probe to go back to mins afterward. Don't make the probe travel further than its needs to in order to make the building, so that it has less travel distance to get back to mining. Sometimes it helps to make probe travel corrects by placing a building just so. Be aware of where your probe is when you are making buildings...say you want to build 2gateways, if your probe is in the middle ground of where a gateway is supposed to go the warp will push the probe around and you will lose control of it, rather place the probe right in between the two gateway plots and then make them. And so on.
- Fighting units should be built like probes, one at a time and one right after another, unless otherwise decided. Don't make more gateways than necessary.
- Unless decided otherwise for good reasons, I would generally suggest building early pylons near minerals so that probes do not take much time off mining in order to build pylons.
- Builds things with the appropriate timing. For instance if I wanted citadel for leg upgrade, but after it was completed I waited 30seconds before making leg upgrade, that is inefficient. I may as well have just build citadel 30seconds later, which means I could have built other things quicker.
In PvT there is usually not any opening adaptation involved. Usually you do your build and Terran does his unless either one of you does something like proxy. Usually your scout will not gather any significant information for you to act on. The scouting probe's purpose is mainly to to locate your opponent and perhaps harass a bit. Terrans usual opening builds fall into these categories:
1) FD (I consider this standard tvp, it's a strong and versitile build)
-weak fd: This is a fd meant to expand at ~4:30. 4rine/1tank/1vult with mine.
This is like siege-expand, you don't have to worry about them attacking you.
-strong fd: This is a serious attack unlike weak fd, and this expos ~5:00. Two varieties here. One that attacks you ~5:00 with 1tank, and one that attacks you ~5:30 with 2tank and a bit more marines.
2)2fact (many styles of this and it can be very dangerous)
Most common 2fact is perhaps 3tank to vulture.
Take Stylish's low eco 2fact numbers:
3tank/6vult attack at 6:00 with vultups and scan.
3)Siege expand varies from placing CC before 4:00 before tank is even started to floating it and placing it down past 6:30 if they are facing goon pressure. They usually have 3-4tank with siege at 6min. No way they can hurt you early here, but you may be able to hurt them by breaking their wall, reaver drop, etc. The main issue here is that they will expand earlier than you, perhaps giving them a stronger economy later.
4)Drop - most usual is perhaps 4vults, but there is 2tank siege, 1tank/2vult, and other unusual combinations.
Standard PvT BO vs. Terran openings
1)Vs weak fd you have no problem, but they probably have a slight advantage since they expoed earlier, and if they play it right you won't be able to attack them successfully early. Still, they may mess up with their units and lose them, or they may not get siege in time, so you have the potential to win early.
2)Vs strong fd it will be very tough without a ramp. This shows how this is not always a strong build on a flat map. Vs strong fd it seems you need to do a build that has many units early to be able to consistently stop it.
3)Vs. 2fact. Vs low numbers of tanks, such as pure vult attack you are in good shape as you have goons and ob. But vs. some sort of hard attack with tanks it will probably be rough going since you went robo and expo. So just do what you can to fend it off. When you scout it, you might even want to cancel nexus, get another gateway or two and go reaver.
4) Like vs. weak FD you have no problem vs. siege-expand but you probably don't have an advantage either. There are significant differences in siege expand builds depending on what the map allows and how your opponent plays. If they can have miners on a 2nd CC before you even start nexus at 5:30, and you are not able to attack them successfully, it shows something has gone seriously bad for you, and it is highly suggested that you do a different build, either a more aggressive one or grab an earlier expansion. This seems to be the case on a map like colosseum.
5) Generally you are in good shape vs. drops since you have goons+ob. Just have goons defending the right places and you are in great shape. Easier said then done though because often you don't know drop is coming. Also, some maps you have to worry about your cliff, which might require a shuttle.
To sum up:
Vs a strong attack this build at least gives you a chance, which builds like 1gate range expo typically don't. Still, this build probably does poorly vs a well done strong attack since you went robo and expo.
Vs. a quick expansion you usually do so-so, and what happens next matters more, so the more skilled player should usually win.
Vs. drops and vults you should do well since you have goon+ob+relatively quick expo.
So, not exactly a great spread, but this is a middle of the road build, so this is expected. This may not look good, but understand that you are doing this build very efficiently, and many other players are not doing their builds efficiently. Still, if strong attacks are popular at the moment and/or on the map you are playing, this is not a good choice. On a ramped map, this is always at least an okay choice since you can defend vs. strong attacks better.
Perhaps the main reason bad players are bad is because they use the fundamental Protoss unit poorly. I have already talked about probes role in good efficiency. In normal circumstances you are making probes as consistently as possible, so a key question is when to stop making them? See this Chinese thread on the topic:
The formula for most popular maps is 3x per patch -1.
So in the case of python:
29for main (9 patch+gas)
23at nat (7patch+gas)
But for actual play you need to figure out how many to build, so you need to figure out transfers, and take into account scouts built (typically only 1), as well as the 4probes you start with. In the case of this PvT build where nexus starts at 5:30, you'll want to transfer around 12 (slightly less than half). So you want to build 38 total out of main nexus (29-4tostart+1scout+12transfer) and 8total out of 2nd nexus +3 more for gas later (23-12transfer). Note that in the example replay I do not transfer probes to 3rd nexus. Usually you will want to transfer them however. But, if you are facing the threat of a push, not transferring will give you more minerals short term since you don't lose time transferring, and those probes will not be vulnerable to harass or attack. In any case, the important thing is that this shows the numbers you are aiming for in a typical macro style game. The formula shows the max number of probes you want on a mineral line. Often you will not reach these ideal numbers right away because you will want to pause for more units/tech/etc quicker depending on what you scout.
Also, because it is difficult to keep track of how many probes you have built exactly, it helps to know your plan beforehand. One marker you can use to know to stop building probes for this build is 100psi, by that point you should have enough probes at first two minerals lines. You should stop there out of you first two nexuses, just built out of 3rd (as I did in rep), or transfer and perhaps build more out of all three.
*Some people suggest counting the number of scv terran has on gas. This is only suggestive and not absolute. If they take all scv but 1 off gas they will almost always fast expo. With 2-3 scv on gas the chance of them going 2fact or drop rises.
*When you send your first several goons out to scout/attack terran you can gather more info, usually how many marines terran has made. Again, this is only suggestive and not definitive. Usually a high number of marines 6+ means some kind of hard FD attack.
*you can place pylons around the map at key locations a drop might go over for advance notice.
*Generally make 3+ ob by 10min. Send first to clear any mines if necessary and to scout their build, and see if you can place it over their factories to see timing/how many factories they build. 2nd ob is for your goons & clear mins for expo if necessary. 3rd at their choke to see when they move out, then more if you want around key map locations. Adapt ob scouting as applicable.
*The main point of scouting is to adapt based on whatever you see/don't see that is significant. Coming up with the correct response to what you scout takes a lot of experience and a long time to learn. But, if you just use some common sense, you can figure any situation out. For instance, if you see a fast 3rd cc with your ob, you could take a fast 3rd and 4th for an economic advantage. If you see fast 3fact attack coming from 1base T, cut probes, get more gates, and a zealot shuttle will sure help. And so on.
Strategy and Adaptation:
After you have expanded with the standard BO I have shown you can follow up with all sorts of things. You can go two reavers and speed shuttle. You can go dt drop. But, since this is meant to be standard play, I assume just getting speedlots to go with your goons, and teching up to arb eventually. Many questions remain, some good ones are: When to get speed? When to take a 3rd nexus? When to stop goon production and start zealot production? And so on.
The trouble with these questions is that the right answers largely depend on how your opponent plays. There are no easy answers. No absolutely correct answers for these sort of things, only situational ones. What Day said in his triggers talk is important because it explains how coming up with a strategy in SC is a practice (not so much theorycrafting). Like Day said, you start with a slightly arbitrary point. You just make up a build that seems to make sense. Now, the problem a lot of inexperienced players will have is that their builds will be inefficient, and thus they will not serve as good models. One thing to do is use pro replay(s) to help you come up with one. At first you will basically just rip their BO, and struggle to understand how they played the way they did. Eventually though, you want to go more towards constructing a build yourself that makes sense instead of just ripping a BO, because perhaps the pro even made some bad build decisions for the circumstances. So, after this process, now you have a basic build that is good vs. some things and not so good vs. others. Now you have a standard to work from and the timings associated with it available to you. As you have seen, I have provided many good opening build orders for PvT which you can expand upon.
So, now it comes down to figuring out how to tweak a particular opening and make it work in a variety of situations. This is a complex matter of scouting your opponent, and realizing their possibilities, and how they relate to your builds possibilities. Then, based on this, you come up with a plan for action for the future. Still, face the fact that a particular build does not necessarily work well in all situations and there are such things as build order wins, where you are in bad shape just because of how you happened to decide to play. In these cases, you go back to the standard opening you set for yourself and make a judgment if it would be better to play a different opening standard, that is, to do something else.
As you can see from the example replay, my standard follow up for this pvt build is to take a 3rd nexus ~7:30 off of two gate after making goons consistently. Then pause probes slightly to get citadel+2gates+legs, and start zealot production. What are the timings associated with this? For one, legs completes slightly after 9:30. If you find that is later than you want then you might build a citadel just before nexus to have legs complete ~8:30. Lets take another example: If you find that you want more units than this standard vs. a strong timing push at 9-10min. There are several things you could do to have more units quicker. You can deviate from the norm by 1) cut forge/+1/cannons, 2) cut tech, 3) add more gates and expand later, 4)cut probes, all of which will give you more units quicker. If you do all of them you should clearly have no problem with ~15goon/20zealot @ 10min, but you can mix and match options. Remember what I said about opportunity costs before. Recognize the full range of choices you have to adapt every little bit of your play.
So far I have covered BOs, done some opening analysis, and given some general advice. Past these things, the game becomes so situational that I would prefer to evaluate each thing individually in context. I can speak about builds easily enough without a particular context, but it is more difficult to talk about something like how to attack properly without considering context. This is why an example of game analysis comes next.
1) Zpux vs Stylish @ Python.
Thank you to the players for 'providing' the rep. I picked this rep because they both play well, and protoss plays similar to what I have termed standard PvT. The purpose is to compare protoss's play with an ideal to strive toward.
+ Show Spoiler +
To sum up, Protoss's loss this game came down to slight decisions in both macro and attacking. Protoss had slight advantage most of game, but lost it though a series of attacks.
- Not the best split, it was not to fastest/closest mins, but it was decent.
- The 5th probe pauses ever so slightly at rally point before being sent to mins. No reason for this 30seconds into game.
- 8th probe is not sent to a separate min patch, so it is forced to wait to mine for a bit, and then later move anyway. This and the above are basic efficiency, yet you see players doing these things all the time.
- Stops probe production for 13core. Generally you don't want to stop pause probe production, so 14core.
- Leaves probe outside terran base after scouting. To what end? Goon will be out soon anyway so just send it home to mine.
- Some issue with killing scv. Even though he went 13core, his goon is late by at least 5seconds which is part of the problem. So protoss decides to build 3rd goon then robo.
- Slightly delayed robo...fine, but his building timing is a bit wonky such that when the makes a pair of goon out of two gates he has 7 goon, rather than 6. He had 5goons at 5:27, then another two 30sec later, while the what I call standard PvT has 6 at 5:30. Building another goon delayed his nexus which certainly does not help vs a quick terran expo.
- Trading 1 goon for 1tank is a good trade. This opens up other early game strategies like goon attack +4zealot shuttle drop even more. Sometimes you can put pressure on with your first 6goons+ob, but terran had made 3tank/4rine, siege, and some mines, so no real opportunity here. Could have killed the vult too, but goons backed off too far.
- A large issue this game is gateways. Protoss makes 9gateways before 10min while I made only 4 by that time in example rep, yet I have made more units. Obviously something has gone wrong. Gateway production is like probe production in that you usually want your gateways constantly building, and only when you actually need more production do you build more gateways. Some players may complain that with so few gateways they have to be very careful and not miss unit cycles. Well obviously that is the case, but it is an important part of efficiency to get away with the least gateways necessary.
- At 12:35 mining rates:
My example rep 17746/3584 (49 probes by 8:30)
Protoss had 17010/3256 (42 probes by 8:30)
Terran had 14466/4264
This just shows how building probes constantly and sending to mins instantly matters. It also seems to shows that you will often have significant room to cut probes since Terran will often not produce as many scv when going for a timing attack. If you stop building probes when 3rd nexus is done (~40probes) you will still have equal mining to terran.
- Protoss's active army at 10min is 16goon (including 1lost and 1 in base) and 7zealots with speed almost done. P has no tech or forge/+1/cannons. As my guide shows, you can have ~14goon/14zealot at 10min with tech/+1/cannons. You can see that he has 2more goons than my standard, but 7less zealots. Enough zealots are necessary to successfully fight a tank army. And if you see your opponent lay down 5fact from 2base you can make more gateways for more units instead of building cannons and tech -- which this protoss did do.
- Another major thing was the way protoss attacked. Well, it was clear he did not have enough zealots, so the choice should have been to stall T with goons giving time for his 9gateways to make zealots. He did attack though, and his zealots died off rather quickly. Typically after zealots are gone and it looks like your attack will not obviously work, you want to back your goons off so they do not turn into blue goo. However, this protoss intentionally moved his goons near the tanks to be able to attack tanks. He was not quite able to finish off the terrans army while mines and reinforcements finished his off.
- So by 11min terran has obtained map control, while the protoss still has a better economy. 11:30 protoss makes an attempt to break out, but loses a lot of zealots to mines and is just able to beat off terran, but now terran has a clear advantage. As a general rule, don't attack if zealots are not able to get to tanks, as you will end up wasting your army. At 13:30 there is a last desperate attack, but P loses more zealots to mines, and that is game.
To sum up, Protoss's loss this game came down to slight decisions in both macro and attacking. Protoss had slight advantage most of game, but lost it though a series of attacks.
Replays of standard PvT build
+ Show Spoiler +
1)Pusan vs Nsp_Fancy @ python
Pretty much my exact opening displayed with reaver followup, which is a decent option on python since you can use shuttle to expo to semi-island. @10min 14goon/reaver/9z speed half done.
I would be willing to hold up the way T plays as a standard for seige expand as well. T unit production is impressive with ~10tank/20vult/3rine @10min.
2)Bisu vs nada @ arcadia
With an easy to take 3rd bisu gets 3rd nexus off of only 6goon. He
also goes low goon numbers, only 9, in order to have lots of zealots more quickly. After he spanks Nada's army he goes for expansions, and ends up losing a long game.
3)Bisu vs test[3.33] @ rush hour
8goons before nexus here. More solid macro with shuttle/dt midgame strategy.
4)Bisu vs Nada @ reverse LT
After expo he goes two reaver speed shuttle to carrier.
5) Nony vs. Artosis @ python
Commentary of game on Youtube
Nony sees fast CC so he expos after only 4goon.
Nony lets T get almost all the way to his choke uncontested though which is always bad. He was fortunate in getting citadel for legs and 3rd gate before even scouting with ob, so I suppose that was either because he knows artosis style and/or it was close positions. Still, I think a very early zealot and legs build is not quite as good as consistent goon production, which helps fight off a terran army traveling towards your nat better.
6)Best vs bbyong @ othello
Best starts with zcorez build (2z/3g at 5min) but after that it is the same style since he adds robo+2ndgate. Then he gets expo and 1reaver.
Liquidpedia strategy database
Other PvT guides:
+ Show Spoiler +
IMO the best of the bunch. A decent general pvt guide.
Lastly, this guide is not static. It is a work in progress. I am here for you guys, and am working towards a better understanding of the game. So if better understanding and knowledge comes along that is relevant I will include it in the guide. I will add sections as necessary. There is always more to add, so if you have something you think should be added to this guide, feel free to say so. Please ask any strategy build related questions. If you play a game following Standard PvT BO and want me to look at it go ahead and post it.