TvZ: A Summary of What's (Still) Broken

I’m a mid-high masters Terran (I’ve finished top 8 more than once). Some of you may have seen me sporadically cast games with Halby over at his channel on youtube. I’ve appeared in a few small tournaments in the New England area, but I am by and large a casual player. I live in Boston, am old (31), and have a relatively low APM given my level of play.

Blizzard recently stated that the game is essentially well-balanced at all levels of play. It’s this statement, among a handful of others between developer forum-posts and interviews, which inspired me to write this article. If they truly believe that the matchups are equal, they’re ignoring some pretty significant trends.

For starters.their own tournament sported a meager 12.5% Terran representation, none of which made it past the RO16.

I won’t belabor this point since it has been repeatedly made, although I will offer the following additional pieces of research which support my claim:

IPL5 winrates

Race winrates at major tournaments over the past 4 months

Recent Lings of Liberty thread on Team Liquid

If you take the aforementioned list of examples as indicative of the unequal TvZ pairing (as I do), then the next logical question would be: what exactly is going wrong? Well, I'm glad you asked. The reasons- or at the very least, the symptoms- behind this unequal skew can be summed up by exploring the following five themes in detail: (1) Timing Attacks, (2) “Unforgiveness”, (3) Infestors, (4) The Late Game, and (5) Attention Imbalances.

About five months ago, Blizzard extended the queen attack range by 66%, thereby giving Zergs a vicious new increase in defensibility. A new meta-game quickly emerged: Zergs found that by building 4-6 queens early on, they could spend nearly all of their larvae on drones while simultaneously being safe from the vast majority of Terran timing attacks. After several months of failed attempts to break this build, Terrans finally reasoned that the only solution was an equally economically-aggressive opening; the 3-OC Hellion-Banshee was born, and with it, a (temporary) equalization of win-rates between the two races.

However, while supposedly providing some equilibrium to the metagame, there remained some striking differences between these two openings which ultimately illustrated that they were not equal in strength. First and foremost, Terran remained quite vulnerable doing the 3 OC opening, whereas Zerg, with their 4-6 transfusing queens, did not. As such, Zergs quickly caught on to Terran’s greed, and punished it with nydus, ling-runbys, and assorted forms of all-ins. Within weeks, the fleeting equilibrium in win-rates shifted once again, back in favor of Zerg.

All of that is to say that timing attacks for Terran were (and are) no longer reliably effective versus Zerg. TvZ is presently the only matchup where timing attacks do not work. When a Zerg plays beyond economic greed (such as using all of the queens for creep tumors rather than saving the energy to transfuse), they become vulnerable again. However, this is a rarity in practice.

TL;DR: Zerg can punish Terran greed, but Terran cannot reliably punish Zerg greed in kind.

The term “unforgiving” has been circulating a lot recently among forums for Terran players. The general consensus is that there are simply too many common, benign mistakes— a Terran player looking away from an army for a fraction of a second, for example—that end the game in the Zerg’s favor. In general (excepting tier 3 armies, which will be covered later) a Zerg’s army can always retreat from a Terran’s army, but not visa-versa. Fungals and a speed/mobility advantage means Zerg players have more options regarding if/when to engage a Terran force. Good creep spread—a staple of any high-level Zerg player—gives the Zerg plenty of cheap forewarning of a pending Terran attack as well.

Additional examples of common game-ending Terran snafus in TvZ include:

• Encountering a Zerg army with tanks unsieged.
  
• Lowered supply depot (see zergling harass below).
  
• Not scanning ahead to scout burrowed banelings.
  
• Clumped-up army getting surprise-fungaled.
  

By contrast, the most costly mistakes a Zerg can make in the current meta-game revolve solely around infestor miscontrol. Although these can and do occur, they are less likely and common than the previously cited Terran examples.

The aforementioned article on the the rise of Zerg offers additional support to my claims above, and includes some interesting statistics on Terran versus Zerg play between Korean and Foreign players.

TL;DR: The potential for unrecoverable mishaps made by Terrans are far greater than those made by Zergs within a typical TvZ game.

No TvZ imbalance thread would be complete without a section on infestors. Unsurprisingly, they have been a staple Zerg unit in ZvT over the past several months. Let’s look at why:

• Infestors are strong versus all Terran compositions from the mid-game and beyond. Unlike ravens, they are helpful immediately upon finishing, especially with the pathogen glands (starting energy) upgrade, and stay viable—even increasing in value as they survive and reserve energy—throughout the game.
  
  
• Infestors are very difficult for Terran to kill. Fungal, infested terran, and neural parasite mean that it has no natural hard counters in the Terran force. Furthermore, while cloaked banshees are good low-attention unit to provoke a high-attention response from infestors (see the Attention section below), their high cost makes them an expensive option in practice. Ghosts are a theoretical counter, but serve no other practical purpose in a Terran’s army, are very difficult to micro properly (while also splitting, sieging, and performing other Terran micro-necessities), and very susceptible to fungal. Ravens with seeker missile are another oft-suggested counter, but they require a significant up-front investment before they become viable (starports, tech labs, and worst of all, all of that precious Time spent waiting for raven energy to accrue). Finally, since fungals outrange seeker missile (even with the new patch), good infestor control can keep ravens at bay indefinitely.
  

TL;DR: Infestors remain a strong unit versus all Terran compositions without a reasonable counter in response.

In order to compete against Zerg production in the late game, Terran is required to spend an inordinate amount of money on infrastructure. This has been a dynamic—not necessarily a problem—of the matchup since day 1. What this means, however, is that Terran players must commit to bio or mech early on, so as not to waste a large amount of money on infrastructure that will be useless to them in the end game. I will briefly review each of these tech paths and how they become problematic in the late game for TvZ. For Zerg, I will focus on the infestor/brood lord composition, since this is generally the most common (and the most deadly) late game ZvT approach.

Bio
+ Show Spoiler +

Mech
+ Show Spoiler +

TL;DR: Terran has no comfortable late-game option TvZ. Zerg’s ability to continually build towards a “value-heavy” 200 unit-space army gives them a definitive advantage in the late game.

We can think about Starcraft as a competition of resources; each player is attempting to leverage their own set of resources to overcome their opponent’s. There are the obvious resources in SCII such as economy, army, and production. There are also the not-so-obvious (but just as critical) resources of attention and actions per minute (APM).

Attention is a close cousin to APM, but yet is distinctly different. Loosely defined, attention is the resource required to actively look at a situation in order to adequately respond. It’s a combination of situation assessment and action execution. One of the most effective aspects of harassment in any matchup is that it is usually a low-attention attack tactic (drop, ling runby, warp prism warp-in) that necessitates a high-attention response by the opponent (looking at attack, judging how much of a threat it is, examining army, deciding what to pull back if anything, pulling it back, then micro-ing the army once it has retreated to the point of attack some moments later).

As an example, mutalisk harassment is generally roughly a 1:1 ratio as far as attention goes, meaning that the Zerg needs to spend just as much attention to avoid turrets and the defending army as the Terran does to find and defend against the mutas.
However, while the muta example above is equally balanced, it is unfortunately the exception rather than the rule as Zerg has a far greater arsenal of low-risk viable options for harassment which provoke a disproportional response in required attention from Terrans. Specifically:

• Base prevention. Most Terran players would describe a burrowed ling (or an overlord pooping creep, or general creep spread) at a desired-expansion site annoying. I would describe it additionally as attention-consuming. In the case of a ling, a Terran may move a base to an expansion site and attempt to land it, but not know for some time that the landing has been denied. This requires the Terran to
  
  1. examine the base,
     
  2. look for the source of the block (ovie? Unburrowed ling? Burrowed ling?),
     
  3. move a clearing unit into position
     
  4. come back to the base later when the unit is in position,
     
  5. execute a valuable scan to kill the ling, and
     
  6. return the unit to the pack.
     
  
  None of these actions are difficult, but they do amount to a large amount of attention, especially considering the multi-step process of the task and the minimal attention required on the part of the Zerg who placed the burrowed ling there in the first place. Waiting for creep to recede from a base creates yet another step of required attention before the delay has been fully dealt with.
  
  
• Ling runbys. Lings are fast! They’re often a blink away from charging into your main or mineral line. It doesn’t matter how 1337 you are, if you are a Terran you know what I am talking about; Terran needs to have lowered Supply Depots to successfully rally their army without forcing attention-consuming raising/lowering of depots. This, plus the fact that Terran is the only race without cheap, zero unit space, static defense means that the Terran’s main base is almost always defenseless against a Zergling pillage-fest. Because lings can burrow and sneakily hide in various corners, it can often require multiple iterations of army retreating scenarios in order to deal with this threat fully. While a Terran is waiting for their defending army to get back to their base, they will often need to lift buildings and re-rally SCVs, all of which require attentive focus (both on the front end and on the back when the buildings are put back into place). In general, a Terran’s base requires more attention than any of the three races to rebuild and zerglings are one of the best units out there for creating just the right circumstances to necessitate this attention-consuming process (and generally making a royal mess of things).
  
  
• Burrowed infestor harass. Burrowed infestors and roaches are the only units in the game that can move invisibly without being detected by the human eye. This makes infestor harass particularly brutal as it cannot be anticipated without ample cloak detection. More importantly, however, is that burrowed infestors in a Terran’s main have the most lopsided attention ratio of any form of harassment in Starcraft II. A few sets of infestors can move unseen from corner to corner of the Terran’s base, requiring multiple scans, turrets, and often multiple trips back home to finally rid the main of them for good. By contrast, this form of harassment requires very little attention from Zerg: they need simply to look at their infestors in order to spam infested terrans every once and awhile, then move the infestors to a new hiding spot. Other than this, there is very little that Zergs need to focus on with their harassing infestors relative to the attention it requires from the Terran to deal with them.
  

TL;DR: Harassment tactics from Zerg require a disproportionately larger amount of attention for Terran to deal with than Terran harassment does for a Zerg player.

I feel it is important to state explicitly that I have nothing but the highest respect for the developers at Blizzard. Starcraft II has sustained my interest longer than any other game in my extensive game-playing career, and I have the folks at Blizzard to thank for this. They are a creative bunch who put extensive thought and planning into their game, and it’s clear they’re passionate about the result.

I honestly don’t have any significant opinions on how to modify the game to account for some of the imbalances that I have described in this article. I think the development team at Blizzard really excels at coming up with fresh solutions to balance problems. My concern lately, however, is that Blizzard isn’t paying close enough attention to what is broken at the high level of play in TvZ matchups specifically, and as such, their proposed solutions (so far) have been largely inadequate. The development of Heart of the Swarm has further enhanced my concern; they are introducing several more late-game options for Zerg (ultra burrow, blinding cloud, abduct), without giving Terran anything new to deal with them in the late game. (Only the widow mine: another unit that has low value given its unit space requirement).

This is obviously a Terran post written by a Terran-loving Terran player. It is not, however, my intention to whine or complain, and I’ve strived to make my conclusions as objective as possible. Most of my assertions have already been made by players far better and smarter than I am.

But please, feel free to disagree with me! If my language or examples seem too Terran-leaning, I welcome counter-examples. Rather than insist that points 1-5 are absolutely true for all high masters+ Terran (versus Zerg) players, I'm more trying to paint a broad picture of a general imbalance. If you feel as though I have under-represented Terran advantages in the matchup, I am open to hearing your contrasting opinions (especially if your experience is first-hand at the high level of play).

Nebbish over and out.