This is the second part of the TvP midgame game plan piece. I’ll start off by summarising the key points from part 1 before I continue with the topic of interest. Part 2 discusses about the two main midgame styles.
Key points from part 1
This post is not about opening, and it focuses on the midgame game plan. What is mentioned here is applicable to almost all openings as long as the midgame threshold has been reached.
The midgame starts when Terran have Stim and Medivac on two bases.
Do not commit to an attack if Protoss are sitting defensively on two bases. Delay the third for as long as possible.
The window to attack opens when Protoss are stretched to defend three bases.
Things to consider
I will be discussing the scv pull game plan and “the Maru style” based on the things to consider below.
What to do with the army
The game plans of the two styles are very different, and consequently, the tactical decisions are vastly distinct too. Scv pull is relatively passive until the boys are pulled, while the Maru style is all out aggression.
What units to get
The unit compositions of the two styles are different. You need a certain composition for the specific style to work.
What buildings to build
At the start of the midgame, Terran usually have
– 2x Command Centre
– 3x Barracks
– 1x Factory
– 1x Starport
– 1x Engineering Bay
and will add either another two Barracks or a third Command Centre. Terran should also have Stim, Combat Shield, Concussive Shell and +1 Infantry Attack. What you build next depends on which style you’re playing, because a certain style requires certain unit composition. The sequence of the subsequent buildings matters.
Scv pull
I have previously written about Scv pull. There have been some new minor variations since that post. For example, you can choose whether to add three Ghosts into the mix, build a third, no add-ons for the 4th and 5th Barracks etc. Nevertheless, the general idea remains the same, pull the Scvs for an attack for the kill. As mentioned in the previous Scv pull article, the key is to line up the upgrade timing. The actual timing to pull the boys is not based on the time stamp or supply count, even though there may be a relationship there (e.g., upgrade finishes at a certain time).
What to do with the army
The game plan emphasizes on winning the game when the Scvs are pulled, and it requires Terran to maximise the damage output in that one single attack. In line with this thought, it is important to minimise unit trades before the attack. This is because the rationale behind Scv pull is to let the workers to tank the damage, while the bio deals damage behind the Scv wall. In other words, the Scvs prolonged the survival time of the bio during a straight up engagement, and in turn increases the damage output. Therefore, the average damage output of a bio unit during the Scv pull fight is generally higher than any other trade situation in the midgame, unless Protoss made a glaring mistake. Therefore, there is little value to try out-maneuver Protoss by running around and dropping, as the small benefits from these trades do not outweigh the damage output lost in the Scv pull battle. What the Terran player should do is to be relatively passive and macro up until the Scv pull timing hits, and avoid unit trading as much as possible.
Below is a recent GSL vod of Flash vs. herO. Pay less attention to the opening and result, but focus on what Flash did in the midgame.
As you can see from the above vod, the Terran army do not attempt to make any engagement once the midgame approaches. Flash simply moves a small chunk of his army around the map to pick things off, and it is easily “retreatable” if Protoss tried to engage. So this really sums up what to do with the army.
What units to get
Almost no one pull Scv against Templar builds anymore, at least not knowingly in normal circumstances. Thus, the Terran army composition revolves around dealing with Colossus composition in general. The army composition consists of
– Marine
– Marauder
– Medivac
– Viking
– Ghost*
The first three sets the core of the TvP army, so it is self-explanatory. The viking is to take down the Colossus. The common question asked when Scv pull is brought up is that, “how many Vikings do I need?” The answer is hardly precise, as the attack timing relies on the upgrade timing. In general, if you produce with two Starport, there should be 10 to 14 Vikings. Assuming there is no upgrade discrepancy, it takes 14 Vikings to one shot Colossus. Thus, it is safe to assume that it will take two shots to kill one Colossus during the Scv pull engagement. The number may vary if opponent goes for a Phoenix Colossus composition.
Another thing is, when do you stop producing Medivac and switch to Viking production? A good rule of thumb is stop Medivac production after two pairs, but you can make minor adjustment here and there according to the situation. In the vod above, Flash stops at one pair of Medivac. I will say the range is at one to three pairs of Medivac, before switching to Viking production.
Ghost is optional. It is a good source of aoe damage for that one single attack. You usually get three Ghosts from the three Barrcks with Tech Lab (the other two have Reactor).
What buildings to build
It is rare to do a two base Scv pull unless you cut Scv production, because of how the left over mineral bank up. It is weird to put down two additional Barracks instead of the third Command Centre, as it is hard to support 7 Barracks and 2 Starports on two bases.
The convergent point for Scv pull is usually
– 5x Barracks
– 2x Starport
I left out the Factory on purpose as it does not matter. You do not produce Widow Mine, unless you happen to land the Factory as you commit the attack. Of course, you may have leftover Widow Mines from the opening you chose, but you do not produce more in the midgame stage.
Whether you build the 4th and 5th Barracks or the 2nd Starport first has marginal differences, it should not affect the game plan. Before you start to theorycraft too much, the option only occurs if you put down the third Command Centre before the 4th and 5th Barracks. By putting down the 2nd Starport before additional Barracks obviously gets you an additional pair of Vikings tiny bit faster, but the interesting thing about this sequence is how you manage the 4th and 5th Barracks. You can basically not build add-ons for these two Barracks, and just produce Marines to max out the unit count for that small time frame to hit a strong attack. In my opinion, unless you are a professional player, I doubt it is wise to do this as you probably won’t be able to squeeze the benefits of this minor difference out anyway. The vod below shows Flash doing the no add-ons version.
Most importantly, when you start building the 2nd Starport, the Factory should also build a Reactor. The Reactor is for the 2nd Starport.
Below is a bad example of a Scv pull (timestamp at the midgame move out timing). I am not referring to the execution of the attack itself, but how well the decisions made line up well with the game plan as a whole. Taeja attempted to engage the Protoss army while he was building up for a Scv pull in the midgame, and he lost a decent size of bio army. Moreover, Taeja started +2 infantry attack but pull Scvs at the same time, so it is evident that there is a lack of game plan.
Multi-pronged attack (Maru style)
The multi-pronged attack style is not new. But given the fact that Maru plays it so well to the extent that he redefines how this style is played, I will just call it Maru style.
Long story short, the game plan is to keep making small to medium battle trades where Protoss are weakest, and subsequently overwhelm the opponent. This style emphasizes on being mobile and aggressive. Much has been said about the micro required to execute this style (extremely important to excel with this style), there is little discussion beyond that.
What to do with the army
When people say attack on multiple locations against Protoss, they will usually drop in the main and attack the third. While this is an option, it often does not work as well as intended. Protoss often position the Stalkers in the main to catch the drop, and the rest of the army at the third. Thus, if you drop in the main and attack the third, you are basically playing into Protoss’ hand.
Perhaps the more accurate way to think about the approach is that Terran should attack where Terran can overwhelm the Protoss at the specific location. The below vod best illustrates this approach. Maru did not blindly split his army in two groups and attacked in two locations. He moved around and tried to deduce where the Protoss units were positioned, and picked his attack spots accordingly.
One way to ensure you have more army than the opponent at a location is to have everything you have together and attack at one location. This is based on the fact that both sides have approximately same army size. Another thing is to rally near your own base and move out by loading them into the Medivacs. This reduces the chance of Protoss spotting the army position, and you can deal with Warp Prism harassment better.
There is little I can say about how to execute this as it requires on-the-fly decision making, but the approach can be summed up as follow.
– Attack where opponent is not.
– Attack where you can overwhelm the opponent’s force.
– Keep the Colossus count low by targeting them down.
– Non-stop aggression
Maru often figures out that the Protoss has the Stalkers elsewhere, and he will attack the rest of the Protoss army with everything. In the process, he will try to kill the Colossus, then back off when the Stalkers have arrived. The vods below show just that.
What units to get
Unlike the Scv pull style, this does not produce Vikings early on and keeps producing Medivacs. Viking adds little value to this style as just a few accomplish nothing, and the mobility of Medivac is what makes this style possible. With that being said, you can get a second Starport and make Vikings as you get more ahead and finish the game with a single attack. The unit composition looks like this.
– Marine
– Marauder
– Medivac
– Widow Mine
The Widow Mine complicates things for Protoss. At face value, it trades decently against Gateway unit. Terran usually drop at the mineral line and burrow the Widow Mines, while attacking whatever is there with the bio. Terran have to get out of there when the Protoss army moves there, the Widow Mines often get nice shots due to the pathing of the Zealots. Subsequently, Protoss have to wait for the Observer to move there to clear it (if there isn’t one positioned), so it stretches the mechanics and attention of Protoss. Occasionally, Protoss forgot about the Widow Mines, and put the Probes back to mining. Well, you know what happens.
What buildings to build
The 4th and 5th Barracks are usually built before the third Command Centre for this style for obvious reasons. More importantly, there should be a Reactor on the Factory to produce Widow Mines. Thus, the midgame building convergent point looks like this.
– 5x Barracks
– 1x Factory
– 1x Starport
Subsequently, add three more Barracks to the mix and expand.
Conclusion
It is important to have a game plan, and the decisions made should also be aligned with this plan. The unit movement, unit production and the building sequence should sync together for the plan to work. It should not be just “build Vikings when I see Colossus”.
Hi, I just want to say again that i enjoy your posts very much. Keep up the good work. There is one question that i want to ask you. When I tried to do the 14 min mark +2/+1 scv all-in in TvP. A lot of players would just go back to the natural, build sentries and abandon the third. I wonder if it is a good idea to attack into his small choke in the natural, or should i just destroy the third base and send all the SCVs back to work. Thank you very much for clearing things up.
If you are going for a +2/+1 timing, I think it’s better to get the +1 air/mech upgrade too since it requires armory for +2 upgrade. Personally, I do this when I go an early ebay upgrade build. Ideally, you want to engage when you pull scvs, but if opponent sacrifice the third and use force field, just kill the third and get the scvs back to mine. I tried to move up the ramp at times, I always regret it.
Great posts, I’m glad I came across your blog. I’ve reached Top Diamond as a new player (3-5 weeks) but was held back by my winrates (88% TvT, 72% TvZ, 29% TvP) Due to the MU being totally different, whereas TvT/Z Micro and Positioning is a lot different.
Thanks for boosting up my knowledge and skill of the MU by tenfold!
Quick question, does going for a WM Drop instead of 3rax benefit this style (maru’s) more due to the amount of medivacs coming out earlier or is it the irrelevant by midgame? Thanks
Think of 3 rax as +10 points for yourself, and wm as +7 points for yourself (with the potential to -1 to -10 points for your opponent). The points are like how much advantage you get as you go into mid game. Clearly, I didn’t give too much thought about the ratio of the points, but you get the idea.
Yea, that helps a lot, thanks dude!
Very useful post!
(top master player who hasn’t the time to check all pro games, this blog is a perfect place) :D
=)