The HotS Terran Help Me Thread

TheDwf wrote:
Mines are technically time bombs. Whenever a target enters the radius of a burrowed Mine, the target is acquired, and after a 1.5 seconds delay the Mine shoots if the target is still within its range. You can manually target with Mines: during this 1.5 seconds interval, manually selecting another target in the radius of the Mine will reset the time bomb and thus the aforementioned protocol (1.5 seconds delay → is the target still in range? if yes, shoot; if no, auto-acquire another target if there is one available). If, for some reason, you need to hold fire with a Mine, you can keep switching targets and prevent the Mine from shooting until you're ready.

Mines have 5 range, but can sometimes trigger against 6-ranged targets (such as Stalkers, Marauders or Hydralisks with the range upgrade) attacking them because instead of stopping at exactly 6 range, it seems units sometimes do not decelerate properly and step a little further forward, thus entering the range of the Mine and suffering its wrath.

Mines cannot auto-acquire the same target at the same time. If one Zergling walks in the range of 5 Mines, only one of them will shoot. If one Immortal enters the range of 3 Mines at the same time, they will not fire at once but successively shoot (i. e. the first Mine will lock on it and shoot, after which the second Mine will lock on it and shoot, after which the third Mine will lock on it and shoot). You can circumvent this problem by manually targetting the desired target so Mines fire at once, but in this case Mines can overkill.

Mines need detection to acquire cloaked targets.

Mines deal splash damage to your troops, but not to your other Mines.

Mines do not trigger on morphed Changelings.

Mines do trigger on Hallucinations unless you have detection to reveal them.

Extra info can be found here.

On March 29 2013 04:47 TheDwf wrote:
99% of the time getting a PF on your third is a bad idea (unbalances your mineral/gas ratio + kills your development and ability to scan, which is critical when meching to know your opponent's army location and scout a possible air transition).

On March 29 2013 04:47 TheDwf wrote:
Pick up - Ctrl + click all Marines, shift move on Medivacs, move Medivacs over your soldiers in distress.
Sending drops - select group, shift fill two Medivacs, shift remove them from the group, reassign group, send full Medivacs.

TheDwf wrote:
Standard is rax 12 gas 12 with a very brief (~5-6 seconds) SCV cut before resuming production. A more agressive option is to go rax 11 gas 11 and cut one SCV for an earlier Reaper.

TheDwf wrote:
Hellions and particularly BFH are better for land-based raids in the mineral lines because they can chase workers, unlike Hellbats who need a Medivac boosting over them to keep getting some damage. Hellions and particularly BFH are better to deal with Marines-heavy small drops (TvT), Zerglings raids or Zealots/DTs harass because they come faster and can quickly wipe out the threat (Hellbats too, but they're slower). Hellions can also wander around on the map with more ease due to their high speed.

Hellbats are generally better for head-on engagements since they have +50% hit points and better AoE damage if they connect.

On March 27 2013 08:23 TheDwf wrote:
When opening Reaper you're more or less forced to scout around your base with a SCV before starting your first or at least your second Reaper, else you run into troubles against proxy Marauders.

TheDwf wrote:
If you don't SCV scout and blindly start a second Reaper while your opponent is going proxy Marauders, you've just lost (watch Bogus/Innovation vs Flash, Cloud Kingdom, MLG) barring epic SCV micro to surround and kill the first Marauders while you build a Bunker on your ramp.

If you scout it in time, make Marines instead and delay your Reaper until his attack is stopped. Use this Reaper to scout his transition.

TheDwf wrote:
Considering the amount of gas first openings, playing 1 rax FE into 3 rax is too risky against the threat of Marines/Tanks/Banshees pushes. Building the 3 necessary Turrets to cover your bases will leave you without enough resources to hold the attack.

TheDwf wrote:
You may try it on maps like Nimbus or Deadwing, but otherwise it's risky. The 3 rax transition has the same problems as the 1 rax FE opening (see above).

Ver wrote:
When he comes with a banshee you don't just blindly 1a after it and let him kite you perfectly in a rhythm. Instead you form a perpendicular line to get the maximum firing power and when you think he's going to fire and pull back, you retreat instead of walking forward so he whiffs. By doing this right if he's going to get shots off he's going to take hits as well.

On March 21 2013 09:31 TheDwf wrote:
No real way to stop a determined opponent. You can have Sensor Towers, Turrets rings and some Mines in your main, but if your opponent wants to commit and land 4+ Medivacs, he will. With mech, having your Vikings correctly positioned is crucial, and leaving 1-2 Tanks in your main is virtually mandatory. In Marines/Tanks mirrors, try to be as active as possible; if you're constantly in his face he won't be able to load 40+ supply of bio like that. Rearrange your base layout so your defence can come unhindered; the only thing worse than a doomdrop is a doomdrop having a natural concave against your reinforcements streaming in a line to defend.

TheDwf wrote:
Pure Marines/Tanks does not work well against mech, you have to include Marauders. Pure bio is another option. Switches to Vikings/Banshees once you're on 3 bases are possible too; you build mass Vikings to overwhelm is Viking count so your Banshees have free reign, at least until he gets a Thor.

On January 06 2013 04:49 TheDwf wrote:
Marauders are mandatory against mech if you want to play Marines/Tanks as pure Marines obviously get shredded by mass BFH/Tanks. Don't switch to mech, you can't and anyway you don't need to; you can keep playing biomech with 2-3 Factories (on 3 bases) so you muster enough Tanks to prevent him from advancing. Bank scans to negate the vision advantage Vikings provide and scan repeatedly so he can't slowly edge forward. Get a Thor so a single Banshee can't force your Tanks to unsiege and retreat. Use the map control you usually have by midgame to restrict his movements (you can even try to set up a hard contain, with Turrets and a Sensor Tower, if the opportunity arises and the map layout allows you to do so); ultimately, transition to air if the game reaches the Tank stalemate point. I suggest watching Polt's games, e. g. Polt vs Bomber @ IPL5 on Entombed Valley and Daybreak, or the Polt vs Bbyong series from their Code A match (RO24 2012 Season 5), to see the kind of maneuvers you can use when biomeching against mech. You can also check ByuN vs Ryung, Neo Planet S, Code S RO32.

For HotS games, Bogus vs Flash, Neo Planet S, MLG is a nice example of how to dismantle a mech player on a bio-friendly map.

TheDwf wrote:
Vikings/Ravens/Battlecruisers. See Polt vs aLive, Daybreak, NASL Season 4 and ByuN vs Ryung, Neo Planet S, Code S RO32.

On March 20 2013 02:43 TheDwf wrote:
Generally you make only 4 Reapers (more would be a waste considering your opponent is supposed to get Metabolic boost to push back the Reapers on the map) then transition into triple OC: CC fact [or the reverse depending on Zerg's answer] mine bunk reactor CC. The Bunker can be built lowground, in a corner at your natural, or behind your wall, again depending on Zerg's reaction. Same for the Reactor on the Factory by the way, you need to be sure he's not going for a Baneling bust before building it. Be sure to retreat your Reapers early enough so Zerglings do not catch and kill them on the map. From the last MLG you can check Last vs Life, Akilon Wastes and Newkirk City; Polt vs Life, Akilon Wastes; and Last vs Killer, Akilon Wastes.

If your opponent answers with early Roaches, he should not be able to force your ramp (Bunker/Mine/wall/SCVs for repair) but you will need a Banshee to break the contain, so get a Starport and an earlier second gas.

TheDwf wrote:
No. Zerg gets an early gas, stalls with Queens/Zerglings until Metabolic Boost is done and threatens retaliation with a Baneling bust or a Roach attack/contain; or assuming his defence was correct, he simply remains ahead. See Maru vs JKS (set 5), Icarus, GSTL Pre-Season FXO vs Prime; YoDa vs Ret, Ohana, IEM [starts at 0:33:15]; Flash vs bly, Akilon Wastes, MLG. Another possibility for Zerg, if he Drone scouts, is to proxy his Hatchery in your main and build Roaches from the eggs there while defending at home. See Goswser vs DeMusliM, Cloud Kingdom. It's also virtually a build order loss against 14/14, which can be used reactively if Zerg sends an early Drone scout.

On May 05 2013 00:37 TheDwf wrote:
Let us say that it is the kind of build which can net you lots of free wins, but is stopped dead once the opponent knows the "technical answers" and execute them correctly.

The coinflippy aspect of "drone scout or not" (unless you know the habits of your opponent) also makes the build very fragile, contrary to gasless proxy 2 rax in which there is still a lot of ambiguity even when Zerg has identified it. Here, the only tiny ambiguity is "single or double rax," (though honestly the 8-8-8-8 variant barely exists at high level because it's a complete all-in) and the "gas → stall for metaboost" answer can deal with both.

There are succesful instances of the build at pro level, of course: MarineKing vs bly, Cloud Kingdom, MLG Consolation Bracket; Maru vs Bboong, Star Station, Code S; Maru vs Soulkey, Akilon Wastes, Code S. But in each of them (the outcomes were respectively massive advantage, win and win), you can clearly see they worked because of the Zergs' mistakes and no scout.

What thus happens, when Zerg scouts and/or answers correctly, is that of course you can't hope to kill him straight away, and Zerg remains ahead with the possibility to slow down your development threatening a Baneling bust. That being said, yes, if Zerg's defence is suboptimal, you can have a standard macro game afterwards (Bomber vs RorO, Star Station, Code S).

Maps are also a factor when playing this opening. 8-8-8 is stronger on maps in which there is a longer main ↔ natural distance, because Zerglings and particularly the second Queen from the nat hatch have a longer offcreep distance to travel, so connecting them is harder for Zerg, especially if you build a bunker between the main and the natural (as Maru did in both games, and Bomber too).

On March 29 2013 04:47 TheDwf wrote:
No. Wall natural, and after CC rax gas fact reactor get Marines x2 + Mine before swapping for Hellions. Mine and Hellions behind the Bunker to roast Zerglings, SCVs surrounding the Bunker to mass repair. Something close would be Last vs Stephano, Daybreak, MLG.

Tanks are the wrong approach because they let Zerg get away with a failed attempt at little cost since he can drone uninterrupted and end up still ahead in the worker count after the attack, while Hellions threaten a counter-BBQ if he sacrifices everything for SCVs.

TheDwf wrote:
Short answer: yes, but greedier builds that were in danger in WoL against certain agressive Zerg options are now safe, so they are played instead.

On March 21 2013 09:31 TheDwf wrote:
Still viable, but since you can now be safe even when skipping the Starport thanks to Mines + Reapers being able to scout whatever Zerg is doing, players forgo the Starport (at least when going bio). The trend at the end of WoL was already to skip the early Starport in favor of earlier thirds; at this time it was a risk against 10+ Roaches pressure, but now it's OK, especially with Reapers being able to see the back of Zerg bases.

On March 21 2013 09:31 TheDwf wrote:
The build order Flash used numerous times at the last MLG is CC rax gas fact reactor CC lab gas stim EB EB rax rax gas gas rax rax port, building everything behind 6 Hellions and 2 Mines (assuming no Zerg agression), with the Factory making Reactors once Factory units are done. Bunker as necessary depending if Zerg banks larvae for a 6+ Zerglings poke or not.

A similar development can be played out of a Reaper expand.

On March 23 2013 01:24 TheDwf wrote:
Not really, because Zerg takes a fast gas, stalls until Metabolic Boost is done and then threatens a Baneling bust. The second rax would come too late (unless you're going 8-8-8), heavier Reaper harass is best done out of a single proxy rax.

TheDwf wrote:
Yes.

On March 26 2013 00:15 TheDwf wrote:
No, Mines alone are horrible against Baneling busts because he can just trigger the ones you have with Zerglings, then you're left defenseless (if you have your Mines behind the wall, Banelings still have time to bust the wall). Stick to Hellions. Dream vs Golden, Daybreak, IEM.

Ver wrote:
You can really never know for sure if they are going speedling allin, speedling expo, ling/bane allin, roach timing, roach/ling allin, or roach/bane/ling. All you can really do note what is possible from the gas timing. i,e, if they get gas around when pool finishes, it will be something 7 mins +, while if they go gas pool, you have to account for a 1 or 0 queen 550 baneling bust or 630 40 ling runby. Another possibility, map dependent, is leave an scv in front of their ramp so you can have advance warning if units stream out.

TheDwf wrote:
This is something you can attempt when going heavy Reaper harass (more than 2 Reapers); otherwise, it does not threaten much.

TheDwf wrote:
Your choice as it's wasted minerals most of the time. Besides, 6-pool doesn't even auto-win against CC first (MarineKing vs Suhosin, Ohana, Code A 2012 Season 3). To parry the threat of a 10 pool, building your CC as part of the wall is also possible.

Flash sometimes use an early scout, either to be sure his opponent is not going 10 pool so he can build his CC directly on the natural (Flash vs Life, MLG Fall Championship 2012, Antiga [starts at 17'20]) or to go for an early EB block (Flash vs Revival, SPL, Ohana).

TheDwf wrote:
Zergs can get dual gas after they take care of your scouting SCV, so scouting no gas does not mean he will not be agressive; it just means the fastest pushes are not possible.

TheDwf wrote:
Against 10 pool you cancel your CC on low ground, wall and expand inbase. Against 14 pool using 6 Zerglings to pressure, pulling some SCVs to shield Marines should do.

TheDwf wrote:
With CC first there is none.

TheDwf wrote:
Yes if you build your CC and Barracks as part of the wall (Flash vs Life, MLG Fall Championship 2012, Metropolis [starts at 51'40]. If scouting after Barracks means after you started your Barracks (2'45 - 2'50), your scouting SCV should come across the oncoming Zerglings, so you can quickly close your wall with a second Barracks (or a Depot) if you built your CC on lowground; obviously you will have to cancel this CC. See MarineKing vs RorO, Whirlwind, Code S RO32.

TheDwf wrote:
No, you have no minerals to spare going early scout. Scout one position with the SCV building the first Barracks (after the Barracks is complete, I mean), then another position with the SCV building the second Barracks (same), so by ~3'45 at worst you will know where Zerg is. Rally Marines to the middle of the map until you know his starting position.

Ver wrote:
3 scvs for the normal semi-allin version (the 2 producing the rax, and 1 other), 5 scvs if you want to win or die there.

TheDwf wrote:
The way you play 11/11 (against Zerg) depends on whether you want to all-in or pressure heavily. Naturally, the closer your Barracks are to his base, the more SCVs you pull, the more SCVs you cut (to start Bunkers earlier), the more you have to damage.

Having your first 3 Marines surrounded is the thing you want to avoid as it stops dead your attack. They have to survive until Marines 4 and 5 join the party (after which you kill workers much faster); it means that you have to temporarily retreat if your opponent pulls 10+ workers while your reinforcing SCVs are still on the way.

Whenever I play 11/11 with the intention to get a macro game I now build my first Bunker at the bottom of his natural's ramp. See SuperNova vs Snute, Ohana, Campus Party Europe; the Bunker at the bottom of the ramp supports your attack (preventing Zerglings/Drones from chasing your Marines too far) while containing the Zerg should you fail to complete a second Bunker in front of the Hatchery. Another example is MMA vs Stephano, Iron Squid II, Ohana.

If you want to play macro behind a proxy (proxy ≠ forward; say it's proxy when your Barracks are around his Tower [Ohana, CK], forward when it's on your natural's ramp or slightly more advanced), you have to be more conservative with your Marines because you don't want to be in this awful spot in which you have to fly back both your Barracks without any Marines behind your wall. (If you don't want to all-in, be sure to build your Depots as part of the wall.)

Knowing whether you're ahead or not can be difficult, yes, as it depends on many things: how many Drones you killed, how many SCVs you lost (beware, unintentionnally cutting SCVs because you're busy microing also really hurts), did you retain map control, are you able to build your expand directly on your natural, how many Spines did he make, how many Zerglings did you force, did he delay his first Queen, etc. Counting Drone kills can help; Zergs have 15-17 Drones (to 13 SCVs before OC is done) then usually go full Zerglings until your attack is stopped. Each completed Spine also counts as a Drone kill. If you don't see a Queen spawning at ~4'50-5'00 at the natural it also indicates that the Zerg is likely still at the Stone Age.

You don't bring all SCVs but 2 when all-inning because you need a bit more to afford an additional Depot and constant Marine production. As far as I remember pulling 7 additional SCVs (+2 that built Barracks = 9 total) already makes it difficult to afford a second Supply Depot. See Polt vs Symbol @ the IEM Guangzhou Korean Qualifier, both games; Bogus/Innovation vs Stephano, Daybreak, Code S RO32 (7 SCVs pulled total, late second depot); ByuN vs Revival, Cloud Kingdom, Code A 2013 Season 1 (9 SCVs; compare with the Bel'shir Vestige game).

SCV micro is a mixture of attacking, retreating and holding position to protect Marines. You have to use corners + SCVs to make sure your Marines are not going to be surrounded and killed.

TheDwf wrote:
Agressivity aside (11/11 > 11/12 > 12/12), it's a matter of optimizing SCV movement depending on the Barracks' location. I send my 9th SCV to the ramp to make a Depot, which finishes around the 1'30 mark; so if I want to make my first Barracks just at the natural's ramp with the same SCV, I cannot have a 1'30 11 Barracks since my SCV has to go down the ramp and walk a bit further (and bringing back this SCV to mine while pulling another to build the first Barracks would waste some minerals). With this movement, I have at best a ~1'37 Barracks anyway, so I might as well start a 12th SCV right away since it does not delay the first Barracks. Thus I have 12/12 instead of 11/11.

On Ohana I make my second proxy Barracks with the SCV which builds the first Depot as part of the wall. But by the time he reaches the location just outside of his Tower's range, I sometimes pool 200 minerals, so I might as well start the 12th SCV after starting the first Barracks. Thus I have 11/12 instead of 11/11.

TheDwf wrote:
Your Bunker needs to be located in such a way it's not easily surrounded by Speedlings and yes, you need to surround said Bunker with your SCVs at the natural and you need to hit as many times as possible with your first 2 Hellions. Your goal is to kill as many Speedlings as possible; you will get overwhelmed anyway if he makes 25+ of them but the more Speedlings you kill, the easier it will be for you to take back your natural.

Depending on your build order, taking note of his Metabolic Boost timing with your SCV or Reapers and fully walling your natural in time is another possibility. Once a gas is complete, Zerg will have Speedlings ~165 seconds (2'45) afterwards.

TheDwf wrote:
Gas doesn't necessarily mean agressive play, same as gasless doesn't necessarily mean passive play. Zergs can get Metabolic Boost early to defend (particularly if you open with gas) / have some map presence / prevent you from building your third on its location, you see players like Leenock or Life do that even if they intend to get a third (e. g. Life vs Lucifron, GSL 2012 World Championship Round 2, Antiga). Your two first Hellions should scout whatever he's up to.

On December 24 2012 03:03 Ver wrote:
I don't think that 12/14 eco builds are particularly viable. MKP used to do this for a time pre-queen patch but Zerg will know its not 11/11 by the marine timing and scv count (unless you load scvs in your cc, which might work). I'd say if you want to fake a bunker rush you hide your rax at your 3rd or something and just do rax cc cc.

What can work is what bogus did against soulkey on entombed in gsl ro8, a 12/12 with only 2 scvs pulled. Soulkey underestimated it and didn't pull a ton of drones, so bogus was able to inflict huge casualties and basically auto win from damage dealt.

On December 26 2012 00:44 TheDwf wrote:
12/14 light pressure was severely weakened by the Queen patch. Before, it was guaranteed you could force a Spine and additional Zerglings if you retained your initial Marines and rallied reinforcements in front of his natural; nowadays, Queens can defend this kind of delayed Marine pressure on their own. Realistically on current maps 12/14 can't complete a Bunker in front of the Hatchery unless Zerg does not react, so you will end up with a later expand while not really hindering Zerg's development (Zerg is able to defend your pressure while still mining gas and going 2 Queens). You can always hope he overreacts with Zerglings but even then you won't achieve much.

If you want to use 2 rax pressure I suggest using 11/11 with 3 SCVs; you sacrifice more economy but since the pressure is much stronger, Zerg is forced to respect your Bunker rush attempt and proceed cautiously, forgoing gas (or sacrificing his natural) and often delaying the second Queen (if he goes 2 Queens you should be able to complete a Bunker in front of the Hatchery) on top of forcing Spine(s) and Zerglings.

TheDwf wrote:
WIP

On March 21 2013 09:31 TheDwf wrote:
To deal with Swarm Hosts when going mech, use your old friend the Tank; a few are enough to prevent Locusts from nibbling your position, and once you have 15-20 of them you can slowly push back even large amounts (20+) of SHs. If your opponent has neither Mutalisks nor Queens nearby you can force them to unburrow with Banshees.

Swarm hosts don't seem to be a valid first tech choice against bio (because Zerg is left with nothing to deal with constant multi-pronged drops/attacks, see Flash vs RorO), but Zergs can get some of them after they stabilize. Swarm hosts are siege weapons, so don't forget that you always have the option to circumvent them via drops. Just like against Broodlords, when fighting them head-on you have to be sure to have what it takes to overpower them quickly if you attempt to scan and focus them ignoring Locusts.

On April 01 2013 17:59 KawaiiRice wrote:
Bio/mine/hellbat players posture/drop aggressively and intelligently to keep map control. Without map control a z is being really suicidal moving swarmhosts without army support; if they do that they are really weak to drops so you can slowly pick them apart. But if you don't have map control and just sit back it's quite easy for them to get in front of your base like that.

TheDwf wrote:
Being minerals-heavy, 4M tends to bank gas indeed. If you get a second Factory with a Lab attached to search Drilling Claws, you can build Thors against lings/banes/mutas. Building a third Factory for extra Mines is another possibility. Other questions you should ask are: do you have too much gas or not enough minerals? Do you produce enough Medivacs? By lategame, you can spend this gas on Ravens (and/or Battlecruisers with air upgrades) or possibly Ghosts.

On March 21 2013 09:31 TheDwf wrote:
Mitigating Mine splash damage comes down to the position of your Mines and your army control. You can play 4M like you would with Tanks, i. e. backing away the main part of your troops from his charging army while leaving the Mines triggering + some bio to kill stray Zerglings in front. If you face lings/banes/mutas you should be pre-spread anyway.

On March 21 2013 09:31 TheDwf wrote:
Tanks: 45 seconds of production, one at a time (150/125), 2.25 movespeed, 3 seconds transformation time.
Mines: 40 seconds of production, two at a time (150/50), 2.81 movespeed, 1 second transformation time with Drilling Claws, possibility to hit Mutalisks.

Tanks have advantages (much longer range, easier to focus), but Mines have a strong synergy with bio because their producibility, higher mobility and lesser gas cost (i. e. possibility to get more Medivacs) allows you to constantly be in Zerg's face, thus having strong chances to build a momentum (once Zerg falls behind you keep smashing him and eventually he crumbles), while Tanks are produced more slowly and don't work well in small-medium supply situations in the open, thus leading to an accumulation dynamics (against dedicated lings/banes/mutas you need like 8+ Tanks to start trading efficiently on creep against large amounts of Banelings). When you have a strong multitask, 4M gives you even better possibilities to capitalize on this.

On March 23 2013 01:24 TheDwf wrote:
Of course, on paper a sieged Tank fires 13 times faster than a Widow Mine, but realistically, when Zerg armies charge, most Tanks die after one or two shots when you have small or medium counts. Therefore, even if they have the theoretical potential to deal more damage, they don't always do, hence why Mines having a 40 seconds cooldown is not always as problematic as it seems… unless each of them fires on a single Zergling.

To have an idea about the firepower of the thing, a Mine hitting a clump of Banelings or Zerglings at his very heart can decimate up to 20-25 of them (vs 5 kills + some splash damage for a Tank in the same situation).

On March 29 2013 04:47 TheDwf wrote:
1) Tanks are still valid, but 4M is currently superior against lings/banes/mutas as you have more ways to outplay your opponent.
2) Tanks are still the core unit in mech.
3) Tanks are still needed against certain Roaches timings and are better than Mines against Roaches/Hydralisks.

On March 23 2013 01:24 TheDwf wrote:
Didn't work in WoL (Baby/TY just massively outplayed Symbol during their match) and I doubt it will in HotS, especially since Infestors as first Lair tech is currently nearly extinct (which means increased odds of Mutalisks, against which you don't want to try expensive and risky Tanks drops). There are simply too many gas constraints when meching vs Zerg, and Tanks are horribly inefficient in small numbers against Roaches.

It could possibly work against someone going pure Swarm hosts (as the first Lair tech), but pure Swarm hosts typically seems like the kind of mistake in the "how to use them" learning process; besides, you would need to make sure you have enough Tanks at home to hold the continuous Locust waves before being able to send some of them to drop. Furthermore, Tank drops are not super threatening, even if you mix Hellbats; Drones can easily flee, and you're left with the simple possibility to snipe tech buildings, which is not enough given the investment. The risk/reward ratio is just too high/too low.

The style looks cute and all, but when you remember that 5 Medivacs = 4 Tanks gas-wise while it already takes a long time to get 15+ Tanks to trade efficiently against Roaches-centric armies in the open… Having 1 or 2 Medivac(s) to harass with BFH or Hellbats drops is a possibility, but many of them doesn't seem doable.

I have seen ForGG do this kind of "Mechivac" style in mech mirrors, though. Works better as people don't get Turrets rings in mech mirrors, and you can turn his base against him as Tanks have such a hard time driving among the intricate mess that is your typical Terran base. And recently, it seems people (at least on Europe) tend to skip Vikings in mech mirrors.

On March 20 2013 01:16 TheDwf wrote:
Keep at least 6 Hellions alive, wall your third and build a Bunker there.

On March 23 2013 01:24 TheDwf wrote:
Mines, Bunkers, wall at natural, spread bio behind with Marauders in front, keep Hellions alive so you can search around your base (otherwise use a Medivac) to see if he's there with mass Zerglings. If your opponent is too lazy to spread his Banelings, get money hits on them with your Mines. Last vs Life, Akilon Wastes, MLG and YoDa vs Lucky (set 2), Akilon Wastes, GSTL FXO vs LG-IM.

TheDwf wrote
There are two viable approaches:
1. Bio/Tanks. You may need Bunkers and a strong defensive position to hold the 2/2 Roaches/Hydras timings on maps with a very exposed third like Merry. See Polt vs DRG, Merry, Red Bull Global; Polt vs HyuN, Overgrowth, Red Bull: Online Qualifier #3.
2. Pure bio. See YoDa vs ByuL, Catallena, Red Bull Qualifiers; Bogus vs Petraeus, King Sejong, Destiny I; Flash vs Symbol, Catallena, Code A.Your goal is to survive in a decent state until 10+ Medivacs and max supply, after which you comfortably massacre Zerg's army in a head-on engagement assuming Zerg doesn't have a much wider concave. Some Mines can help defensively but you should cut them later on.

On March 24 2013 05:25 TheDwf wrote:
Marines/Marauders/Tanks/Medivacs. WoL game, but watch Flash vs EffOrt, Ohana, SPL. For a HotS game, see MMA vs Vibe, Star Station, ATC (reactive 3 fact Tanks).

4M is not impossible to play against Roaches/Hydralisks, and you will have no troubles if you have the advantage and/or the initiative via drop play, but you may be vulnerable in some windows because it takes time to get enough Marauders and Medivacs to fight on equal terms with Zerg. A compromise can be to produce Tanks on your second Lab Factory as displayed by Last at the MLG.

TheDwf wrote:
Due to the production asymmetry between Terran and Zerg, assuming an even game it takes too much time to build enough Marauders/Medivacs to start fighting on equal terms against Roaches/Hydralisks, which leads to the existence of windows of vulnerability in which Zerg has simply too many Roaches/Hydralisks for your army to handle, even if you're actively working towards a higher Marauder/Medivac count. Two games to examine this vulnerability are Ryung vs YuGiOh, Newkirk Precinct, ATC [starts at 13'05 and goes on the next VOD] and Flash vs Soulkey, Star Station, Code S. In both games Zerg overextended and ended up losing their advantage (and subsequently the game), but they did have one after they forced their Terran opponent to sacrifice all their SCVs at the third to hold the Roaches/Hydralisks timing. Naturally, short distances (close spawns, certain maps) only make things worse.

Additionally, Mines are not as helpful against roaches/hydras as they are against lings/banes/mutas, since they have to move towards the opponent (unless you're defending at home of course) and burrow in front of a ranged wall of beefy units (since Roaches are supposed to be at the front) as opposed to the fragile, clumpy melee swarm of lings/banes/mutas rushing towards them when engaging the Terran army.

Hence the Tank; while not being tremendously efficient at first when in low numbers, they help secure your position when carefully positioned (particularly on high grounds overhanging the third if the map layout allows it), and allow you to build an army which will stomp Roaches/Hydralisks when fighting at the appropriate timings (see below).

Switching from Mine production to Tanks is not harder than swapping your fact reactor with a lab rax, and getting Vehicle Weapons instead of mech armor. I recommend getting 3-4 lab rax to build a healthy Marauder count, while producing Tanks out of 2 or even 3 fact lab later.

Since you build Tanks, the logical next step for Zerg is to get Vipers, which is why you want to hit a timing before them (which, negatively, also means that as long as your opponent has no Hive, you have no need to rush into a premature attack and you can passively keep building your Tank count until your army is unbeatable by his first army). This timing is critical as you want to hit neither "too early" (when lacking the critical Tank count + ranged mass to prevail) nor "too late," when Zerg has several Vipers with enough energy ready to nullify your Tanks and disrupt the position of your bio troops. An example of the immediate disaster which can occur if you hit "too early" can be found in ThorZaIN vs Stephano, Neo Planet S, Dreamhack.

Basically, waiting ~170-180 supply and 6+ Tanks is better to engage. If you can wait longer (against late Hives) and afford to max before with 8-10 Tanks, don't hesitate.

As bio/Tanks vs roaches/hydras/(infests) is a ranged war, make sure you get a wide concave before engaging.

TheDwf wrote:
1. You can transform your initial Hellions into Hellbats when going for your Medivac push.
2. You can decide to play bio/Hellbats/Thors instead of 4M. With a slightly heavier emphasis on Marauders, the composition is OK but retains some vulnerability to really high Banelings numbers.

TheDwf wrote:
Basically you get your third + double EB + gases at natural when you move out (if it's a Stim timing, CS timings hit earlier), around 9'30 – 9'45. Mvp vs sLivko, IEM Cologne VII, Antiga (also shows what's wrong with this kind of build).

Ver wrote:
There are two main versions to transitioning out of it, and it will also depend on which build you did to get there. In general, the version I feel is best is:

After banshees/3rd cc/ double ebay/stim

1 gas at nat
fact lifts off, makes another reactor. 4 rax, including one on the open reactor, then 4th gas.
Starport and Fact swap. 5th and 6th gases, armory when needed.
3 more raxes and 1 more fact when money comes in. If muta 2 more facts.

Alternatively some people start on tanks right away instead of a second reactor, but this makes you more turtlish and lets them spread more creep.

Ver wrote:
The most popular mech build is the one Supernova, Kas, and MVP use, which is:

Rax CC -> Reactor Hellion -> Cloak Banshee -> CC -> Fact 2x -> Armory 2x

You have a lot of potential damage from hellion/banshee, you are safe vs allins, and the banshees are very useful in the midgame to stop the Zerg from bullying you around with roaches. The only downside is your upgrades are very late and if they scout and defend correctly you end up behind. If you see very early mass roach, you can get siege mode first, if not get blue flame.

Personally I have always preferred getting an ultra fast armory after the factory finishes to rush +3 attack, thus delaying cloak (or not getting it), with the build otherwise similar.

If you are okay stretching your multitasking then I would recommend making a Medivac and viking after the 2nd or 3rd banshee and attempting hellion drops as well.

TheDwf wrote:
Personally I always runby my first 4 or 6 Hellions if his Queens are out of position (not only for Drones kills, but also for information) and/or he has no wall; obviously you retreat if you see lots of Zerglings. During the game, it depends; you can mostly kill Drones when his army is out of position, otherwise they just die for nothing. It also depends on your opponent's creep spread and the map layout since you'll obviously get more kills if you surprise him.

Constantly suiciding Hellions for Drones looks like a good trade but things are actually more subtle. Whenever you lose Hellions during midgame, you weaken and/or delay your pre-Hive tech push; meanwhile, your opponent just replenishes his Drone count right away and voilà, you didn't necessarily gain something valuable... Hence those infuriating games in which you can kill 100+ Drones and still lose. By lategame, yes, you can dispose of Hellions with far more munificence but now I am a lot more conservative with my Hellions during mid-game.

Ver wrote:
I think they are two key ingredients . Aggressive movement attention/commands goes something like this:

1) Load up send out a drop, waypointed to arrive at around hte time your attack will hit the target area (usually the 4th).

2) Attack move the army to the area, macro a round of units

3) Stim a handful of marine/marauder ahead of the rest of your army, generally around 6 or so units.

4) As you are walking forward into a potential danger area, start pre splitting parts of your army to the left and right then attack moving back to the target, while keeping your advance guard ahead.

5) Once you get to creep scan ahead and have your screen force pick it off. Remember the Zerg can't see where your army actually is, so for all he knows, your screen is right in front of the rest of it.

6) Once you see the location of the Zerg's army, start pre splitting your army a lot more and begin spreading and sieging your tanks, depending on when you think he will attack, and make sure to produce another round of units if the timer is almost up.

7) Once the Zerg attacks, stim, do whatever immediate declumping you need to in order to avoid massive fungals. Then target fire banelings and infestors with tanks while pulling some of your bio back and trying to form a concave over their area of attack. Retarget fire with tanks as needed. When all your bio is spread out enough, then attack move towards his army and focus on spreading individual parts.

A key ingredient of this is watching pro games very closelessly for micro, preferably streams/fpview or the first person in replays, then practicing it in an isolated setting (micro trainers). For an example of the seeing, my TvZ push micro actually improved a lot after two weeks of playing as Zerg only, because I felt what was annoying to play against and what was easy, and so when I went back to Terran, [TheDwf: incomplete sentence; Ver probably meant that he knew better what to do from the Terran side].

TheDwf wrote:
When going 4M:
Keep producing 4M with a heavier emphasis on Marauders; you should also have many Medivacs by this time. Be sure to spread your bio, hit & run and focus Ultralisks with Marauders. Mines help to alleviate the burden since one hit removes 25% of the hit points of an Ultralisk, or can kill numerous lings/banes. Producing Thors on your Lab Factory is another possibility. Ultralisks are also vulnerable to drop play if Zerg no longer has mass Mutalisks.

YoDa vs Scarlett, Akilon Wastes, GSTL.

When going mech:
Mines/Hellbats/Tanks/Thors for mech. (Tanks unsieged if there are mainly/only Ultralisks.) In open space, use Autoturrets if you have Ravens to prevent Ultralisks from closing the gap too quickly.

TheDwf wrote:
Yes, whether you went bio or mech. Once you're on 4+ bases (which thus assumes a favorable map) with a stable position, you can slowly incorporate Ravens and/or Battlecruisers with Viking support if you're still on a high economy and the game calls for it (the transition takes some time, so if your position in the game is too fragile you might lose while attempting to get those expensive units). See Dream vs Stephano, Akilon Wastes, IEM [starts at 0:46:54] and ThorZaIN vs Killer, Newkirk City, MLG.

On March 26 2013 00:15 TheDwf wrote:
One Turret in your mineral line, 3 Bunkers and one Turret in front (you can go up to 4-5 against the 2-bases variant), mass Marines (no Marauders!), pull SCVs in front of your Bunkers and quickly react to the focus (Bunkers evaporate in 2-3 seconds). Shift focus Voids with your Bunkers. Retreating to your main is mandatory against 1-base variants. Void busts generally come after a proxy Oracle, so taking no damage from the Oracle is critical (4'45 Turret in your mineral line). 1-1-1 openings defend with Marines/Tanks/Vikings.

The best opening to defend this is Reaper expand into 4' Ebay.

Old examples: Aphrodite vs Eve, Atlas, GSTL Pre-Season; Polt vs Creator, Akilon Wastes, IPL Fight Club 48 and Polt vs HerO, Entombed Valley, MLG Showdowns.

Recent examples of 1-base variant: Mvp vs Has, Overgrowth, King of the Iron; Mvp vs Has, Frost, King of the Iron.

Recent examples of 2-bases variants: Bogus vs HuK, Deadwing, Destiny I; Flash vs Pigbaby, Merry, IEM Toronto Asian Finals.

On March 26 2013 00:15 TheDwf wrote:
Depends on the opening and the tools you have at your disposal.

You opened with 1-1-1 after or before expand: Marines/Tanks with the help of some Bunkers and possibly your Mines (which forces some caution from Protoss) and one Medivac can hold it fine. Lift your natural and defend in your main. Generally you're on a dangerous spot though; your Tank may be out just as the attack starts, and retaining this first Tank is critical, so make sure its position is hard to break for Protoss (who may typically try to blink all his Stalkers forward to snipe this Tank). Be sure that your base layout does not work against you; you need some open space for your SCVs/Marines to chase Stalkers.

From the last MLG check Mvp vs Feast, Star Station and Bogus vs MC, Cloud Kingdom. Add Bogus vs Rain, Akilon Wastes for a case of 1-1-1 expand. Another textbook hold can be found in Ryung vs hellokitty, Akilon Wastes, WCS America.

You opened with Reaper expand into 3 rax Medivacs: scout with Reaper, then Bunkers + SCVs/Marines/Marauders are your best bet. Searching Concussive helps getting extra kills. Be prudent with your Stims as long as you don't have Medivacs.

You opened with gasless expand into 3 rax Medivacs: same as above. Mvp vs Feast, Akilon Wastes, MLG.

On March 26 2013 00:15 TheDwf wrote:
3 Bunkers in front (you can go up to 4-5 against the 2-bases variant), mass Marines (no Marauders!), pull SCVs in front of your Bunkers and quickly react to the focus (Bunkers evaporate in 2-3 seconds), focus Sentries with the one casting Guardian Shield as a priority (not Immortals as first; they're too beefy). Retreating to your main is mandatory against 1-base variants (except perhaps if he doesn't proxy the Robotics, in which case you may try to defend your natural if it doesn't have backdoor rocks). 1-1-1 openings defend with Marines/Tanks/Medivacs (focus Sentries with Tanks). Be careful with your Bunker placement, particularly on maps with backdoor rocks like King Sejong or Deadwing.

Old example of a Marines/Tanks defence: WoL game but watch SuperNova vs ReaL, Ohana, IEM Cologne VII.

Old examples: Watch Polt vs Creator, Cloud Kingdom, IPL FC48 and YoDa vs PartinG, Ohana, IEM.

Recent or less recent examples of 2-bases variants: Bomber vs JYP, Overgrowth, The Big One; KeeN vs First, King Sejong, Redbull: Online Qualifiers; Flash vs Jangbi, Bel'shir Vestige, Proleague; TaeJa vs Try, Bel'shir Vestige, ATC.

TheDwf wrote:
Now that the arsenal of Protoss all-ins has been weakened/reduced, playing CC first or 1 rax FE is less risky. Nonetheless, without a Reaper scout you may have to play in the dark in a few situations, so Reaper expands are still the standard TvP opening. It is recommended to be active on the map with a SCV and/or a Marine to avoid unpleasant surprises.

The superior economy can be put to good use to fuel a stronger 5 rax pressure or deadly SCV pulls.

On March 24 2013 05:25 TheDwf wrote:
No, it was already bad at the end of WoL and it's now completely horrible because of MSC and/or Oracles.

TheDwf wrote:
Now that Zealots/Stalkers/MSC pokes are extinct and Blink attacks less common, walling is safe anew most of the time but it's still rarely done because walls can quickly become a liability against certain 1-base all-ins (in particular, a Reaper expand cannot afford to lose the Reactor + losing depots is painful overall).

TheDwf wrote:
Nope, completely dead because of MSC (Photon Overcharge) and Oracles.

TheDwf wrote:
If Protoss is vigilant and doesn't play in autopilot assuming you expanded, it won't work because of the MSC (Time Warp, Photon Overcharge).

TheDwf wrote:
Against 2 Gateways in my base, I use 11/11 and I build my Barracks between my CC and geysers; against a single Gateway, 12/14 should be enough. Regardless of your building layout you will have to micro anyway, so pull back SCVs after they take 2 hits and remember that a Marine cannot take any hit from a Zealot if you move it manually with caution.

Proxies outside your base are simply thwarted by a wall.

TheDwf wrote:
When scouting a Protoss you want to look at the following things:

- Energy on Nexus.
- Number of gases (and look how much gas he mined to know if he took them at the standard timings) + how many Probes are mining.
- Number of Pylons (should be 3 shortly before your SCV has to leave).
- Pylon and building placement (Pylons in a corner on in the back should make you suspicious).
- His first units.

TheDwf wrote:

Usually signs are:

- Dual gas
- 3 Pylons in his base, usually with one of them being able to power buildings in a hidden or semi-hidden corner of the base
- No early expand
- You saw only Stalkers (sometimes they build a Zealot, but usually they don't; sure thing is they don't get a Sentry)
- One Stalker stays in front of his natural to deny further scouting
- No early proxy around your base (though sometimes the Robotics is proxied...).

The map can be a big indication too, some maps are more favourable because of blinkable ridges.

If you're unsure, have a scan ready around 7'15 - 7'30, this is when DTs show up. This way, if it happens to be Blink Stalkers, you can try to snipe the Observer if the Protoss player is careless.

Ver wrote:
Not many koreans do the double reactor style, it's really only MKP that will use it consistently. Basically, double tech is more flexible and versatile, while double reactor is better versus very specific openings. Essentially:

Vs Colossus -> double tech lab >>> double reactor. double reactor has a terrible game vs this.
Vs fast templar double reactor > double tech
Vs double forge -> double reactor >> double tech

Marauders are better for dropping and raiding than marines except for killing probes. Marines are more cost efficient and exponentially better in straight up fights if the Protoss doesn't have aoe. Marauders are still useful if you have a lot of room to kite though. So essentially it boils down to what you predict the Protoss will do, so it's kind of a coinflip either way. I double reactor on maps where colossus are trash (i,e taldarim) and double tech on most others unless I know they dont like cool.

On March 26 2013 00:15 TheDwf wrote:
Nothing changed. Like before, you just poke; if for some reason he has not enough you can try to commit, otherwise you simply back away and soft contain him, threatening drops and so on. Offensive possibilities are increased when you have your third and fourth Medivac.

On March 21 2013 09:31 TheDwf wrote:
The fundamental TvP plan did not change, you push with your first Medivacs, try to delay Protoss' third as long as possible dropping or threatening drops while adapting to his tree path, etc.

Ver wrote:
Your 2 medivac poke shouldn't actually fail, it should just be repulsed without much loss on either side. You cannot afford to commit too heavily with it. When it does get repulsed freely, you shouldn't worry or think you are behind, it is simply natural. You can continue threatening drops, getting upgrades, and taking your own third. Ultimately, the Protoss has to mess up for you to really gain an advantage, but you can also mess with htem by using empty medivacs to pull their units out of position. The decision between 3 and 5 rax really is more dependent on what kind of player you are and how confident you are if the game drags on. 5 rax obviously has much more offensive potential and can punish greed a lot easier, but feels weaker if the Protoss knows what they are doing and the map affords easy thirds.

On December 13 2012 02:39 TheDwf wrote:
When you move out with your first 2 Medivacs, it's more of a poke than a real timing. As you probably experienced, painfuls things do happen if you try to force your way through a ramp (see Yoda vs Socke, Ohana, IEM Katowice to see what happens when you commit while you shouldn't), so unless Protoss went some kind of horrible build not allowing him to defend this simple pressure, just soft contain him while waiting reinforcements / more information about what he's up to. When Medivacs #3 and 4 arrive you have more possibilities; you can try a quadruple drop when Protoss moves out his army near his warping third while moving the rest of your army to cancel the warping Nexus, etc. But there are games in which nothing special happens and you just get away with a faster third (preferably built directly in its location).

On January 25 2013 17:37 Type|NarutO wrote:
Your timing if you play it well should hit before the 2nd colossus and before storm (low units in that case) but its not a push that you can actually win the game with against any Protoss that is not completely bad. He can stall time if he needs with forcefields (…).

If you catch him out on the map (aka not watch tower, not his base) I'd be a bit more defensive but other than that, you can apply pressure. Pressure doesn't always mean attacking. Pressuring can be showing presence on the map, denying his vision. As soon as your medivacs are there, poking is fine but actually attacking is dumb. Protoss has the best defensive ability when it comes down to defending ramps, terrain in total. He can make up for supply deficit with simply cutting your army, he can deny fire power or forcefield-circle you to bug you.

TheDwf wrote:
A common misconception is that Terran systematically has the advantage at the time the Medivac poke arrives at Protoss' natural. That's actually untrue: while with some tech-heavy build orders Protoss might need to stall with Forcefields if you threaten to commit (i. e. approaching the ramp as if you were going to Stim then retreating), some build orders put more emphasis on units and therefore allow Protoss to comfortably massacre your army should you try to commit. Use a stimmed Marine, your flying Factory or a Scan to evaluate the strength of Protoss' army (keeping in mind that Stalkers are usually in his main) and act accordingly.

On February 10 2013 05:25 TheDwf wrote:
There are multiple goals [with the Medivac push]:

• seizing map control: you deny vision (scanning Observers is key for that) and Pylons, trying to keep him in the dark as much as possible; map control means you can usually build your third directly on its location for example.
  
• scouting his tech path/unit composition;
  
• keeping him modest, i. e. he has to deal with your troops and the threat of drops if he wants to take a third.

Soft containing means being in his area, threatening drops so he can't freely move out on the map, or threatening an actual attack so he can't comfortably skip units in favor of greedier tech.

TheDwf wrote:
5 Barracks (usually 2 Reactors + 3 Tech Labs) on 2 bases, 8 Barracks (3R + 5L) on 3 bases. Beyond that it depends on your economy, if you have several OCs you can get 12+ (you should rarely need more) with Tech Labs on all those additional Barracks.

TheDwf wrote:
Generally, don't attack in a head-on fight a Protoss having Storm unless (a) you have Ghosts, preferably with several EMPs ready or (b) you are confident in your ability to dodge Storms while outnumbering him (situational).

Ver wrote:
It also depends on whether you think the Protoss is being greedy and if you can punish it or not. If he plays defensively and you 3 rax cc, you end up ahead, while if you 5 rax and he is defensive, it remains even unless your upgrades are delayed.

TheDwf wrote:
The threat of drops currently allows Terrans to safely 3 rax third on the spot without much trouble. (Be sure to search the surroundings for possible proxy Pylons; a single Zealot interrupting the construction is annoying). Therefore, there are specific reasons to go 5 rax before third:

1. The earlier and extra production from 5 rax can be used to hammer down Protoss when they're behind from the opening. See MarineKing vs Seed, Akilon Wastes, MLG Showdowns (consolation bracket).
   
2. 5 rax can also be used to reactively punish a failed all-in or pressure from the Protoss. See MarineKing vs Seed, Whirlwind, MLG Showdowns.
   
3. You're planning to head for a 2-bases all-in (in this case, you don't get a third, or only late when you see you won't be able to force his defence; see the game below).
   
4. 5 rax has a better game against Colossi as first AoE tech (because you can get Vikings on 2 bases, unlike Ghosts), while Zealots/Archons/Immortals and/or Storm with a superior armor upgrade makes it very hard to press on even when you have the supply advantage. See MarineKing vs Seed, Star Station, MLG Showdowns.

Ver wrote:
Second starport is a stylistic thing but is safer. I prefer to get it right away, but some players like Taeja will just stay on 1 starport and rely on having really good sense to know exactly when to make vikings. However, if the Protoss does the trick where they just make 1 colo with no range and tech to templar, then it's mostly wasted money. However, having more medivacs early on is really helpful anyway and rather undervalued.

Ver wrote:
Templar in general are more defensive, stop drops [more] easily, and make it hard to press into their base even with a supply advantage. The downside is templar have no offensive potential for a long time and the templar player gets slaughtered if he tries to push out across the middle because then he can't land a storm and gets kited all day. See Marineking vs Parting from the KSL Finals on Daybreak for an example of this, or Taeja vs MC on Entombed from Asus. Versus Templar players you definitely want to add a 4th cc very quickly and focus on all your tech upgrades because they can't threaten much. In general you want to add a ghost academy as soon as you get a whiff of templar, because if they have templar and you have no ghosts you can't really threaten/attack much.

TheDwf wrote:
The Dream vs PartinG series at IEM Katowice is a nice illustration of the above point about Templar play, see the Daybreak and the Entombed Valley games.

TheDwf wrote:
3-4 Vikings per Colossus. No need to make Vikings against a single Colossus without range since it means Protoss is actually teching Templars.

TheDwf wrote:
You certainly don't want to be Marauder-heavy against Templar tech as they always come with Charge Zealots, and Marauders are terrible against Zealots. You want lots of Marines with some Ghosts and a handful of Marauders; having a high Medivac count also matters a lot. See Bomber vs Grubby, Ohana, IEM Cologne VII; and Bomber vs Squirtle, Metropolis, Red Bull Battlegrounds.

Ver wrote:
It is somewhat a stylistic choice and map dependent but in general you want ghosts sooner than later. You first want your 3rd cc, 5th rax, and double ebay/armory before even considering ghosts. You can't really attack the Protoss beyond a certain point without ghosts, as storm and feedback are just too strong defensively. On the other hand, getting ghosts will limit your mobility and harass potential a good amount, which is why Marineking, an initiative based player, will delay them as long as he can. Usually once you have the above pre reqs and you see templar its worth it to grab the academy and research energy upgrade, but you don't want to start massing ghosts until later on unless you feel the game will be very passive. 4-6 is a good amount for midgame pressure options, as too many will leave you with too small of an army.

Ver wrote:
Can't really put it in words...it really has a lot to do with feel and knowing the practical range of your ghosts/placement in the army/timing. Basically snipe has longer range but feedback is instant cast, so if they both walk in range then try to kill, templar wins, but if Terran scans ahead and does it from range, he wins. Alternatively, you can lead with emp and use the outer radius of it to hit them, though that isn't as good as it used to be.

TheDwf wrote:
Mass Marauders and 20+ Vikings (don't be afraid to make that many Vikings, a Protoss can't recover from losing such an expensive Stalker/Colossus army). You want a Colossus to die at each volley at the beginning of the fight so that his firepower quickly decreases.

TheDwf wrote:
HotS changed the way Phoenixes/Colossi is played. The composition went from fragile/unviable in WoL to potent. First, be aware that there are different levels of commitment to Phoenixes. Some builds get a second Stargate and upgrades (notably the Fleet Beacon range upgrade), eventually going into Tempests or Carriers, some stay on one Stargate and transition towards Archives tech (Zealots/Archons or Storm). As of now there are a few ways to deal with Phoenixes/Colossi play:

1. Cripple/kill before critical mass with 5 rax pressure. Maru vs Classic, Merry and Habitation Station, Code S RO4.
2. Turret contain into Viking harassment. Maru vs eMotion, Frost, Proleague.
3. Thors at your own risk. They may help but they're far from mandatory, and in the end they're probably a liability against Templar tech switches. Bbyong vs Stats, Nimbus, Code S; Bbyong vs Caliber, Deadwing, Redbull: Online Qualifiers.
4. Pulling SCVs. Bbyong vs Stats, Frost, Proleague Playoffs; Bogus vs Adonminus, King Sejong, Destiny I. Avoid doing that if your opponent walls his third with Gateways (Ryung vs Macsed, King Sejong, WCS America Challenger).
5. The classic mass Vikings answer: Flash vs Classic, Nimbus, IEM Toronto Asian Finals. Flash went up to 3 Starports with Reactors, outnumbered the Phoenixes with his Vikings and ended up slaughtering both the Phoenixes and the Colossi. Ghosts/Vikings timings also work: Bogus vs MyuNgSiK, Nimbus, Code S RO32. Examples of lategame play can be found in MMA vs Starnan, Merry, WCS Europe Challenger and Happy vs Adonminus, Overgrowth, WCS Europe Challenger. Note that against high Phoenixes numbers, Ghosts are critical to deplete their energy.

Below are the ancient answers; they still partially apply to non-committed Phoenix play, so I let them for history purposes.

On August 25 2012 15:50 Ver wrote:
Bio/ghost/double starport viking. Don't bother dropping, just play defensively til 200/200. In battle you need to emp the phoenixes and have the vikings focus fire the colossus. You will win the big battles very easy as he'll have so much wasted gas in phoenixes that don't do anything. Hence why nobody does this anymore.

On September 04 2012 06:50 Ver wrote:
The way to beat colo/phoenix 100% reliably is turtling until maxed on 3 base with double port viking and getting ghosts. You can get 2/2 and double port viking, those are not mutually exclusive. Don't bother dropping: it's far too risky and little reward, unless he tries some early 2 base allin and you know where his phoenix are.

In that 200 battle you just emp their phoenix and focus fire colo with vikings and they just wasted tons of money for nothing, meaning you win very easily.

TheDwf wrote:
Yes, actually you overmake Vikings so you can ignore Phoenixes (don't waste shots on them, they're not the damage dealers in his army) and kill Colossi reasonably fast without losing all your Vikings to Phoenixes before they get the job done. Be sure to get air attack upgrades as well.

If you have enough Ghosts, try to hit Colossi too with EMP so you can divide your focus fire and kill 2 Colossi instead of one at the beginning of the fight when you still have enough Vikings.

Don't even try an all-ground approach, it's generally impossible to deal with 3+ Colossi without Vikings unless you badly outnumber him. Uncontested Colossi just deal way too much damage, you will lose like 10+ supply each time they attack and his Zealot wall won't even be down by the time most of your bio is reduced to ashes.

Needless to say, you need a second Starport to deal with Phoenixes/Colossi.

Ver wrote:
Practicing engagements in TvP is really hard because it requires you to simply be active with your army but have a good eye for knowing when the Protoss army is in bad formation (i,e colossus not able to all fire at once, scrunched up in choke, you have wide arc, stalkers in the front blocking zealot/archon, templar clumped up etc) and instantly seizing it. The best way to learn it would just to watch a large number of top Kor Terrans from their pov's and rewatching the battles and sequences leading up to them over and over.

Ver wrote:
You need 2 [TheDwf: I assume he meant 3] hotkeys at a mimum. 1 for marine/marauder/medivac, 1 for ghost, 1 for viking. You can potentially have another for medivacs, and yet another for dividing the bio force into two.

Ver wrote:
The "standard" composition is marine/some marauders/mass ghost/medivac + vikings depending on their colossus count. You want to slowly mass as many ghosts as possible over time. Usually players will try to keep a couple marauders in front to absorb charge, then tank with ghosts while pulling marines back to kill zealots. Then the ghosts will retreat once they are under colossus shots and the entire line will engage. If you do not have many ghosts then typically you will simply let the vikings shoot, preferably from a cliff, while kiting far away with your bio until the zealots are nearly dead. The ideal way to engage is simply attacking with a wide arc when their units are not in proper formation. For example, Taeja vs Younghwa on Daybreak, where Taeja attacks Younghwa at the 4th and wins easily because 3 colossus cannot shoot due to being cramped in between a pylon and nexus.

TheDwf wrote:
Ghosts: progressively head towards 20-30 by lategame. Ghost count in your initial max depends on his army composition. Note that Ghosts are extremely expensive and quite slow to produce so depending on how midgame went you might not always have the necessary time and/or economy to get as many Ghosts as you would need.
- If mostly Zealots/Archons/HTs aim at 10-15; no more since he should inevitably get a second and perhaps a third Robotics going for a Colossi switch, and seeing your whole Ghost squad evaporate to mindless lasers because you overmade Ghosts instead of getting Vikings really hurts your Terran soul.
- If Storm-less max for some kind of 3-0-3 timing try to get 6-8 Ghosts in your max (ideally you would want more but from experience sometimes you barely have time to have the first round of Ghosts ready, not to mention resources issues).

Vikings: 3-4 per Colossus, don't forget air attack upgrades. If you manage to get 20-30 Ghosts and 20+ Vikings you have gold in your hands. See ByuN vs HerO, Daybreak, Ups & Downs.

To scout his army composition, you mostly scan his army or use your flying Factory or poke with a stimmed Marine.

If Protoss opens with Colossi and produces more than 3 of them he will likely max without Storm, possibly going for a 3-0-3 Zealot/Stalkers/Archons/Colossi timing.

By (advanced) lategame you shouldn't really have to wonder if he has Colossi or Storm since he should have both: Colossi-free armies can't really handle mass Ghosts outside of defensive positions and Storm-less armies lose much of their appeal once upgrades are equal (assuming your Viking count is appropriate and you have several Ghosts to carpet EMP his army).

TheDwf wrote:
Heavy Turret rings + Sensor Towers by lategame. One Viking patrolling in a critical air path can also come in handy.

TheDwf wrote:
If you have too much gas it can be for severals reasons: gas 4, 5 and/or 6 taken too early, constant trades causing you to remake Marines and Marauders en masse, not having enough SCVs mining minerals, ... When playing TvP, I find myself gas starved because Ghosts, Vikings, Medivacs and upgrades (including air attack upgrades) absorb everything on 3 bases. When I'm taking my fourth while being maxxed and still on 65+ SCVs, I take gas 7 and 8 right away because I head for mass Ghosts, which takes a lot of gas (Vikings too). So, basically, assuming you have 65+ SCVs on 3 bases, you should be able to dump all your gas into Ghosts/Vikings/Medivacs/upgrades. If you're in a rather low-econ action-packed game in which, for some reason, you constantly trade Marines and Marauders, you should pull some SCVs away from your Refineries because you will have no way to spend your banked gas if it goes too high. Adapt your economy to the type of game you're playing.

TheDwf wrote:
No, definitely not! If you're stuck with a low-tech bio army with some kind of banking (or a high income) you must increase the efficiency of your army; if you fight with your low-tech or inadequate army hoping you will be able to remax on a better composition, you will likely end up not being able to remax at all since if you lose the fight too badly (and you will if Protoss has many Archons/Colossi/HTs while your Ghost/Viking count is too low) Protoss will trample upon you with Zealots warp-ins. Get rid of your excess of Marines/Marauders (depending on his composition) with suicidal raids or simply kill them yourself to replace them with Ghosts and/or Vikings. Do this progressively of course.

On December 26 2012 00:44 TheDwf wrote:
Retreat. Unfortunately most of the time when you narrowly win a fight in TvP lategame all you have gained is the right to have another fight later, hopefully winning this one convincingly enough so you can at least destroy an expand before being forced to run away from Zealots' wrath. For sure, overextending against Zealots warp-ins and leftover defensive Templars is the worst thing to do; the exhaustion of your army by dint of repetitive Stims (which means Medivacs will eventually run out of energy) means you will lose troops for nothing, throwing away the edge you might have acquired. Depending on his expands' location you might be able to destroy a remote Nexus but do not sacrifice whatever is left from your army in order to do that. According to the situation you can also use this true-false victory to secure an expand after you retreated. TvP lategame is a war of attrition usually won after several major engagements when Protoss is forced to file for bankruptcy.

On November 06 2012 13:44 Ver wrote:
For example, in TvP if the Terran loses a battle from an even scenario on 4+ bases, he is either massively behind or its game over as Protoss remaxes almost instantly, Terran has no/few ghosts and will likely lose medivacs/energy on them and his units take too long to make. On the other hand if Protoss loses a battle, the Terran almost always has to retreat and remake their army instead of pressing on (at most killing one nearby expo), as if you do you run into templars and full warpins that decimate your weakened army (Taeja/MC in NASL for example). This is why in lategame TvP's the Terran often has to win 3-6 big battles in a row in order to win the game, while the Protoss needs at most 1 or 2 victories.

TheDwf wrote:
No, 3 Reactors is fine.