Snute came into Toronto looking strong, but hardly anyone expected the performance that was to come. With stunning capacity, Snute's run through IEM Toronto included: Chaos 2-0, CJ_herO 2-1 TWICE, Masa 2-0, and sOs 2-0 before narrowly losing 2-3 to Flash in the Ro8, going out in an outstanding and unbelievable performance. This weekend proved that Snute's long and hard road to becoming a top tier pro was finally starting to pay off, and he was looking dominant in many of his games.
Back in March and April of 2014, Snute was looking shaky in ZvT and ZvP, as his long term game plans were still shaping up, but his ZvZ was easily at the highest level of competition. With multiple wins over HyuN in three different tournaments, a win against TRUE (when TRUE was in the Ro4 of GSL), with only very narrow losses to Jaedong at IEM Shenzhen and the World ESPORTS Championship, Snute has looked unstoppable in his ZvZ. While he looks a little bit more vulnerable in ZvZ today, Snute's ZvP and ZvT have finally appeared to come together through the use of swarm hosts. His patient, boa-constrictor style of ZvP looks inescapable, as several Protoss players have desperately tried to stop the slow swarm host constriction and have been met with futility.
More recently, he has started using a unique swarm host transition in ZvT to help him get to hive. This fascinating and innovative style has become Snute's trademark and popularized a new style. While many are still clinging to the old traditions of muta/ling/bling and chanting the mantra, “get to hive, get to hive, get to hive,” Snute’s unique swarm host play provides many solutions and convincing arguments to the storied problems plaguing Zerg since the mine and hellbat buffs.
In this particular game, Snute not only demonstrates the power of such a transition and how to make it work, but also puts on a brilliant display of cascading aggression to keep Flash pinned back all game long and unable to break the SHs before a critical mass is out. Although Flash eventually comes out on top in the series, Snute shows a vulnerability to Flash's recent hot streak in this game. Contrasted to his slow, patient ZvP style, Snute uses many of the same concepts of SHs mixed with more aggressive tactics to fulfill the same purpose: safely transition into SHs and prevent your opponent from expanding until they starve to death.
Snute's Opening: Less Queens, More Aggression
- 9 overlord
- 13 pool
- 15 hatch
- @100% spawning pool, build three sets of lings
- 19 queen (extractor trick)
- 19 overlord
- @100% queen, build second queen
- 24 overlord
- 23 extractor (4:00)
- @100% hatchery, build third queen
- 32 overlord
- @100 gas, begin zergling speed (5:40)
- 40 overlord
- @6:00 start second gas
- @6:30 start roach warren
- **@6:30 Scout with ling/overlord**
- @6:45 start third base
- **Overlord swell (~4)**
- @7:30, roach warren finishes and zergling speed finishes; start building nonstop roaches
- @7:50 baneling nest
Benchmark: At 9:15, you should have 9 roaches, 14 banelings, and about 35 lings (~16 more on the way). At home, you should have approximately 40 drones, three queens, and triple hatch.
Flash is notorious for playing greedy and not scouting in the early game. Often times Flash will neglect SCV scouting entirely and patrol his reaper at home to prevent early pools from sneaking lings in to cancel the CC. Naturally, this leaves him in the dark about a lot of things and creates a very exploitable environment. DRG showed this recently in GSL and Snute, who had obviously studied Flash's gameplay quite a bit as well, both exploited Flash's lack of scouting with early attacks that completely shut down Flash's ability to early expand as well as somewhat "contain" Flash to allow for any number of possibilities at home.
Snute's opening is a 13 pool, which is much more aggressive than an economic 14 or 15 pool, but significantly more economical than other early pools. With this opening, Snute still manages to get a hatchery down close to the same timing as 15p16h while also squeezing out six lings to pressure. Aside from the metagaming of Flash, this opening still shows strong potential for unsettling CC first openings and, because of the 3-player dynamic of the map, makes it much harder to scout early on, even if Flash had opted for an SCV scout. All in all, this opening provides Snute with a powerful, disruptive build while also giving him the flexibility to catch back up in economy, even if very little damage is actually done.
In this particular game, Flash happens to scout the wrong direction first with the reaper, which is a direct result of not scouting with an SCV. As a result, his reaper is on the other side of the map when the lings show up in his natural at 4:00. The command center is cancelled, and Flash is forced to build it again on the high ground while microing his reaper furiously against Snute's six lings, setting him quite far back but also limiting his intelligence by forcing him to spend time cleaning up the lings.
These lings are sacrificial, but for a good cause: time and space
The second part of this opening is the map control it gives Snute. By opening with an aggressive build, Snute buys himself a lot of space by pinning the reaper back at home and keeping the Terran from getting an accurate scout. The map control Snute creates keeps Flash in the dark and unable to know exactly how much gas has been mined. With this in mind, Flash MUST play safely and prepare for the possibility of a follow-up speedling attack, a one-base baneling bust, etc., etc. These unknown factors force Terran players to pool the first four to six hellions before moving out, in the event that 40 speedlings just run in and surround the hellions. This is an advantage for Snute, as it delays all of the normal information that the hellion pokes give: presence of a 3rd base, number of queens, roaches walking across the map, etc. In addition, he also gives himself space for uninterrupted creep spread. By the time the reaper gets a chance to scout at 6:00, the creep spread has already reached the natural ramp; by the time the first hellions get there at 7:30, the creep spread is already down the ramp and headed to the 3rd. Naturally, this serves to set up the mid game, but it also prevents an easy reaper or hellion scout due to the threat of a ling surround.
The next step to Snute's plan is a roach/bane bust. By using the map control and uncertainty of the Terran gained through his early pool opening, Snute is able to just about completely deny any real scouting information up until 7:30. When Flash finally has a chance to get scouting information by checking the 3rd and looking for gas saturation, Snute already has his second gas up and the roach warren is about to complete. Flash checks the third base at 7:00 and sees that it's already building. Unsure still of a potential bust, he then runs his reaper into Snute's main base, barely dodging past the queens and lings, and manages to scout the two gases just before the roaches pop out. Without actually seeing any roaches or the roach warren, Flash still cannot be sure if Snute is doing a roach bust or hiding two evos somewhere, or even about to tech up to lair off of two and a half bases, so he is forced to scan Snute's natural at 8:00, confirming his suspicions. However, having just started +1/+1 from his ebays and only one reactor factory and one tech lab barracks to produce from, Flash looks to be in dire straits.
You run over there without a second to lose, and what comes next?
Snute's roach/bane bust hits at 9:15 with nine roaches, fourteen banelings, and endless strings of lings. This timing is about 45 seconds later than the traditional two-base baneling bust, but due to the early game openings, both players are in fairly similar situations tech and unit wise, with the difference being that Snute already has a third base up for extra larva and creep spread nearly connecting all three of his bases. It must be said that Snute could have outright won here. With a bit of miscontrol and sloppy micro, he didn't break the bunker or kill as many SCVs as he probably could have. However, after the dust settles at around 10:00, when he realizes the pressure isn't going to outright win, he starts double evos for +1/+1 and begins to flood drones. After killing 22 SCVs, Snute is sitting at around 55 drones to about 30 SCVs with a healthy three-base economy, lair, and +1/+1 on the way.
The biggest sacrifice of this opening is queens. The typical ZvT opener generally relies on queens and their power to deflect reapers, hellions, banshees, and any kind of marine attack while also getting heavy creep spread going to connect the natural and the third base as well as setting up a powerful mid game. Snute, however, throws all these preconceptions out the window and uses forward pressure to buy time and create space away from the hellions for his third base and creep spread to get up. Although he only gets a total of three initial queens with this opener, he actually has relatively stronger creep spread than a standard build because it has gone completely uninterrupted.
Snute's Transition: Lots of Drones and Lings
After dealing significant economic damage and killing off 22 workers, Snute looks to transition into a standard muta/ling/bling mid game. At this point in the game, each player has certain advantages to work with:
- Up 20 workers
- 3 bases
- Great creep spread
- Lair and +1/+1 on the way
- +1/+1 and stim/CS are finished
- Medivacs are out
Knowing that he is behind in tech, Snute then focuses entirely on ling/bling production while getting his spire tech up and sneaking in the occasional drone or two. At this point, he knows Flash needs to put on a little pressure and has the arsenal to do so. With upwards of 60 lings and 20 banelings in good positions, he is very much poised to defend any counter aggression as well as secure a fourth base. The uninterrupted creep spread from the beginning of the game plays a huge role during this stage of the game, as Flash is unable to really start clearing creep until it's already halfway extended out on the map.
With his heavy emphasis on ling production, Snute is able to launch ling counterattacks and threaten Flash into staying at home with his army, buying more time for his mutalisks and drones for his fourth base. When all is said and done, his first mutalisks don't pop out until 14:30, a time when we typically see 20-25 in most games.
When Snute secures eight gases on four bases, he is able to really ramp up the mutalisk and baneling count and defend versus any mid game pressure the Terran is throwing at him while even banking some gas. During this time, he is constantly using ling runbys and mutalisk harassment to keep Flash pinned back while constantly spreading creep, building the muta numbers, and banking gas. The most important thing about this harassment is that it prevents Flash from maxing out and doing a doom push before Snute is ready to defend with his swarm hosts. Again, this is in contrast to the normal ZvT thought process, which is usually to secure four bases, defend with ling/bane until you can max out and smash the Terran army before teching up to hive. Instead, Snute uses the same ideas from earlier in the game of using pressure and counterattacks to buy space and time and still achieve the same goals.
Buying time with constant counterattacks
Snute's Late Game: The Fabled SH Switch
Once Snute has about 20 mutalisks and a decent number of banelings, he begins his unique transition to swarm hosts. When +2/+2 finishes, he starts +1 ranged attacks and an infestation pit and prepares the swarm host switch which he is becoming well known for. His goal at this point in the game is to buy as much time with mutalisks and runbys as possible in order to squeeze out a dozen SHs. From here, Snute's choice of units seems to be a mix of muta/SH/baneling. At first glance, this appears to be a strange mixture of Zerg units to deal with bio, but closer analysis reveals that each unit plays a key role in going toe to toe with the traditionally "superior" Terran biomech ball.
Sword, Shield, and Dagger
Snute uses the mutalisks to continually poke and harass. The job of the mutalisks is to dance around the edges of the maps to check for drops, pick off stray Terran units, and put pressure on the Terran's economy with counterattacks. When he gets the SHs up, Snute tends to use the mutalisks almost exclusively for harassment and relying on his SH/baneling combination to blunt any direct assaults. In fact, in the final engagement, Snute abandons defense entirely with the mutas and just uses them to counterattack into Flash's natural, confident in his ability to defend with only the SH/baneling combo. In this way, the mutalisks are considered the "sword" of his strategy, constantly attacking, always seeking a hole in the armor of the opponent.
That moment when you realize your natural is gone
The second key component to Snute's army composition is the swarm host. With swarm hosts, he is able to constantly create a barrier against any direct engagements, and buy time in between locust waves for banelings to morph, counterattacks to do damage, etc. Therefore, it's easy to think of swarm hosts as the "shield" of the strategy because of their ability to deflect or lesson the blows of the Terran player.
Swarm hosts can deflect most pressure by themselves, but the most dedicated attacks will be able to quickly clear locust waves and step on top of the vulnerable swarm hosts, which is where banelings enter into the picture. The locusts constantly push back the Terran, and the only way to deal with the swarm hosts is to rush forward. However, with the threat of banelings looming just behind the locusts, it becomes particularly hard for the opponent to push in and get on top of the swarm hosts. Banelings are the "dagger" of the strategy, keeping players from getting too close to the swarm hosts and punishing any risky attempts to seize the moment in between locust waves.
Hive Transition and Game End
Snute seizes map control with his high muta count and abundant creep spread, along with multiple ling/bane counterattacks. Snute transitions safely into swarm hosts and techs up to hive without Flash ever reaching his side of the map. After reaching the comfortable number of twelve swarm hosts, Snute uses his mutalisks entirely for harassment while calmly camping his fourth base with swarm hosts and a small group of banelings.
When Flash reaches 3-3 at 20:00, he immediately pushes into Snute's fourth base in somewhat of a desperation push. All game long, Snute's constant attacking has worn down Flash's economy and prevented him from taking a fourth. This is the point at which Snute shows the power of his swarm host style: he is so confident that he can hold against Flash's maxed out push with only the swarm hosts and the banelings that he counterattacks with the mutas and kills Flash's natural, putting the final nail in the Terran coffin and taking the game.
Why This Works and Proper Transitioning
Whenever swarm hosts enter the picture, the immediate response is to just start dropping all over the map and doing multi-pronged attacks. In all of the macro games we've seen so far, games where Zergs use swarm hosts -- mostly in ZvP, but we've also seen a few games of swarm host ZvT over the past three years -- almost always end in the Zerg player either shutting down all harassment attempts with static defense and good positioning or being harassed far too much and crumbling under the pressure and losing. So when looking at the context of this strategy, one has to ask questions like, "Will this die to mass drops? Why or why not? Wouldn't it be better to just go into hive? Why get swarm hosts at all when more banelings would do just as well?" With these questions in mind, let's look at the concepts behind Snute's play and what he does to diminish his weaknesses and bolster his strengths:
- Swarm hosts and mutalisks are a one-time expense
- Drop play can be blunted by mutas
- Defending only four to five bases is possible with swarm hosts
- Closer expansions are stronger than spread out expansions
The most striking fact about swarm hosts is that they are essentially a one-time investment: you pay for them once, and then they continue to pay for themselves consistently while you use your money on other things. Once the Zerg player reaches the amount of mutalisks or swarm host numbers that he needs, those units have enough sustainability to continually be useful without actually trading resources. Even if Zerg stays on muta/ling/bling or transitions straight into ultralisks, they will still be trading resources with the Terran player, and the outcome of the game will rely on how many bases each player can have, a much more difficult task for a Zerg player who cannot easily spread creep to connect all of six or seven bases. Trading with muta/ling/bling also becomes significantly harder once the Terran player starts adding a lot of marauders and thors to their unit mix, as banelings are no longer very effective. When Snute has the swarm host numbers he needs and the necessary muta count, he is safe from starving out, and this is represented by how slowly he tends to take bases afterward, despite having plenty of money to do so.
The most difficult part of Snute's style is during this transition period, as there is a certain vulnerability to direct attacks as the Zerg player attempts to free up gas and supply to start making the transition to swarm hosts. Almost all Terran builds center around a 3-3 timing during which the Zerg player is still stuck on 2-2. From this point, Terrans tend to continually push until the Zerg is starved out of resources taking inefficient trades against the Terran army. Snute’s swarm host transition is designed to deal with the ongoing crisis of starving out in the later stages of the game by taking more cost efficient trades. However, it is still very much vulnerable to the initial pressure. This timing is the crux of his entire ZvT and everything that he's been building to from the 13 pool at the beginning of the game. How does Snute safely get to the swarm host numbers he needs without dying first?
A String of Counterattack Pressure (14:30 - 21:00)
Snute's solution is counterattacks. Snute's string of counterattacks begins at 14:30 when he runs close to 30 lings across the map into the natural. Another ling/bane counterattack sweeps into the third base of Flash at 15:45 while the mutalisks and some banelings threaten the rally point outside of Flash's natural. After regrouping and beginning his transition to swarm hosts with the addition of an infestation pit and +1 ranged attack, Snute starts to shape up for more pressure by 17:00. At this point, Flash has not been able to comfortably step out onto the map, and the creep spread has pushed all the way through to the center of the map unpunished.
Are you serious with this creep spread?
At 17:10 Flash sends a portion of his army to clean up some creep in the map center. Seeing this, Snute swings around with the mutalisks to cut off the line of reinforcements and harass the natural of Flash. After pulling the army all the way to the left, Snute runs some lings and banelings into the third. After Flash cleans up the pressure at his third and begins to feel safe moving out, again the mutalisks swing back into the natural and harass again. Meanwhile, Snute has already set up a counterattack to hit the third base again. This back and forth motion between Flash's third and natural keeps Flash from ever managing to move out on the map and threaten the fourth base of Snute from 14:30 to 20:00, almost six whole minutes! This allows him to comfortably get to twelve swarm hosts -- which appears to be the magic number -- and it even buys him enough time to finish hive.
The combination of the swarm hosts and mutalisks allows Snute the freedom to tech up to hive without dying to a frontal assault, as long as he can get there. This is readily evident in this game, where Flash attempts a maxed out 3-3 push at the earliest possible time and is still unable to break the swarm hosts and fourth base of Snute. The use of counterattacks plays a big role in slowing down the Terran and preventing frontal pushes until Snute is ready to receive it, and even allows him to keep Flash off of a fourth base so that both players stay equal in economy.
Adapting to Swarm Host Play
When the swarm host and muta counts are up, Zerg can start trading cost efficiently versus any frontal assault. However, the second key question going into the late game for Zerg is how to defend multiple bases. When talking about swarm hosts, the natural response is to start worrying about drops and how to defend outlying bases; with such an immobile unit, how can you possibly keep from being torn apart?
In ZvP, the solution to this question is normally to get a lot of static defense and just allow the swarm hosts to slowly chip away at the opponent, maybe using very small numbers of units to augment the static defense. However, in ZvT drops can hit much harder and much more quickly. This is where Snute’s previous emphasis on ling/bane and getting a healthy muta count comes in. With a sizeable flock of mutas, Snute can easily swat down drops, even fairly large ones while still comfortably defending army pressure with swarm hosts and banelings. If more firepower is necessary against a doom drop, some banelings can be pulled to deal with it. In this way, Snute is fairly drop-proof against the Terran player while still being able to control and defend the center of the map.
The other key piece to Snute’s defense is his willingness to expand more slowly. Since drops in ZvT are so much more mobile and potent, swarm hosts can only realistically defend about four bases, even with ample static defense. Because of this, Snute adds on more static defense in his outlying bases instead of extra hatcheries to spend his minerals and only expands to a fifth base when his main is starting to mine out. Expanding this slowly in a standard muta/ling/bling ZvT would be disastrous, as the Zerg would constantly be trading resources to defend and would easily be out-expanded by the Terran player (See: Polt vs Scarlett on Nimbus). However, since swarm host/muta play relies on survivability and cost efficiency, this is not a problem. The high numbers of mutas can also be dedicated solely for harassment, thereby preventing the Terran player from establishing bases all over the map by simply lifting and landing.
With swarm host play, expanding toward the Terran player is actually stronger than expanding away like the typical muta/ling/bling player. Using closer, more forward bases creates a cluster effect of mining bases and helps the Zerg player defend more efficiently. The swarm hosts/banelings don’t require a lot of room to set up and deal well with direct pushes, and the mutalisks can both defend and harass along the same axis. For instance, consider Overgrowth: taking the gold base is far stronger than taking the low ground base outside of the main for the fifth base. Swarm hosts can move between the natural and the gold base and essentially defend all pressure while the mutalisks harass the bases on the other side of the map and spot for drops going into the main.
Snute's map vision and more likely fifth base
Using all of these concepts, Snute is able to create a safe environment to tech to hive and secure his 3-3 upgrades and ultralisks without the risk of dying to 3-3 timing or slowly starving out on three to four bases. Swarm hosts, when viewed this way, become a safe “stepping stone” to hive. The leap between lair tech muta/ling/bling and hive tech is a rather large gap, making it difficult to actually make it to hive tech without dying, but swarm hosts provide the much needed in-between step to make the transition smoother and safer. In many ways, this makes swarm hosts a more proper, but not completely necessary, transition into hive.
This game represents not only Snute’s trademark style, but explains a lot of the concepts behind Zerg economics and swarm host play. From the very beginning, Snute uses early pressure to secure himself a map control lead; with no interference from the Terran player, the creep spread and droning stage is uncontested, buying him a lot of time in the mid game. The following emphasis on ling/bane before teching allows him to extend that early map control and establish a solid economy. Snute then sets up a series of constant counterattacks with mutalisks and ling/bane runbys, and always the creep continues to spread uncontested. By the time the Flash ever really steps foot onto creep, he cannot clear it fast enough to pressure the fourth base or stop Snute’s swarm host and hive transitions. In short, Snute’s constant attention to attacking and setting Flash’s timing back further and further allows him to cover half the map in creep, which in turn defends him while he techs up.
The swarm host style used by Snute is based on the principle that Zerg doesn’t have to trade inefficiently going into the late game. The endurance of swarm hosts and mutalisks allows Zerg to do damage without trading and allows a bank to accrue, which can be used on upgrades and static defense. Even the nature of how and when the Zerg expands is changed by the stylistic choice of using swarm hosts as a stepping stone into hive tech. Static defense is made more of a priority, expansions are taken much more slowly, and the Zerg expands more aggressively toward their opponent. The rules of Zerg tacitly agreed upon are turned up on their head with swarm host play, and one has to learn to think in an entirely different fashion.
Perhaps this is the reason why most players have still avoided swarm hosts in ZvT. The conventions of just playing the standard “better” have never been disputed much, and there have rarely been major upsets in the metagame that dynamically change the way a matchup is played. But every once in a while, a player makes a substantial contribution. In 2011, Stephano was playing gasless triple hatch in ZvT and ZvP, and people considered his style gimmicky and unlikely to work against top Koreans; years later, the core concepts behind ZvT and ZvP openings revolve around the ground work laid out by Stephano’s seemingly misguided judgment.
Snute’s swarm host play in ZvT could be similar. All of the logic is there, and the understated genius is unmistakable. Even in practice, it’s been successful, as Snute demonstrated against Flash. As the best foreigner around today, Snute has proven that he doesn’t just win on gimmicks or timings; his mechanics are solid, his logic is inspired, and his strategic thinking is flawless. It’s not possible to tell if this is the verge of a metagame shift right now, but only time will tell if Snute’s unorthodox strategic thinking will eventually become the new norm.