CatZ on: Swarm hosts
Published on 04/10/2016 20:21 PDT by Kavik
Most of what's written in this 'article' is my opinion, arrived to through my own logic, thoughts, and beliefs. My opinion is always subjected to change through my own experience and watching other players explore their own thoughts in a way that makes sense to me (and patch notes).
Whispers of the Old Gods: Swarm host
Once upon a time, swarm hosts and colossi ruled the earth with little effort required of their commanders. These beings had such immense power that they drew the attention of the developer gods who smote these powerful creatures into the dust.
Tales of the Missing Key
There are tales told about many human commanders of the highest of ranks who unbound the key to call for swarm host reinforcements onto the field of battle, which begs the question, is the unit really THAT bad? I personally think not.. But I understand the decision from a logical perspective; It could be as simple as the small risk of making an undesired unit that takes up 300 resources & 3 supply versus the reward of the perhaps smaller chance of deciding that swarm host would be the best unit for them to make at that point in time.
I asked on Twitter to see what other pros thought on the matter of unbinding the swarm host key to not make them by mistake:
A pro making a conscious decision to remove the SH (swarm host) hotkey means that they've evaluated the risk/reward based on experience, and decided based on what they believe is best for them. However that doesn't necessarily mean that the unit is completely useless.
Swarm hosts - Why? When? How?
I think swarm hosts are usable in certain match-ups / situations. Even though I have used them in a few situations where I determined that they'd be the best overall choice for my army, I don't use them often.
Before going into the "Why", I think it's important to mention "Why Not" to better understand the weaknesses of the unit as potential alternatives for improvement or change.
Defense: A Big No-No
Getting them defensively is almost out of the question, since the attacking army can more often than not just retreat and wait for the very slow locusts to be gone before taking an engagement, because SH have long cool-downs; They leave you exposed and you run the large risk of having spent a lot of resources and supply on units that just didn't pay for themselves defensively.
Attack: A Big No-No
Getting them offensively for timings has similar issues. I thought maybe lings into SH could be good vs roaches in ZvZ because of the crazy DPS of locusts. Lings would be the only melee unit zerg has early-mid game that would naturally stand in front and allow locusts to output their max damage. Having said that, I haven't figured out anything I like with that concept yet but I think it might be possible in certain situations.. but there comes another problem.
Swarm hosts Overlap
SH + lings is a concept that I was exploring with good success towards the end of HotS after they made the SH change. If ROOT's hydra was to play a protoss at Blizzcon, a 2 base SH build was an option we discussed seriously. One of the issues with swarm hosts in LOTV is that they overlap. Let's say you want SH + lings as a unit composition, you get a full ling surround, allow the locusts to be dropped and kill a given Protoss / Zerg Army (since T would more often than not pick up and get out) great job, you managed to use SH effectively; But you could've done the same thing with ravagers. For example - ling surround into Corrosive Bile accomplishes something very similar, but it ALSO gives you a more stable early game from the ability to make roaches, a better defensive stance if you chose to go that route instead of going for a timing attack, and arguably, a better and easier to manage resource dump more often than not. I'll touch on the resource aspect later. In short, the ravager route seems to me much more flexible and reliable. Like with ravagers in the last scenario, swarm host overlap with muta and run-bys in general, often not as well rounded job of harassment, and increasing for the player a risk of the unit dying/trading inefficiently because of their limited mobility and 'expensive' cost. Compared to these, SH are also low-impact units in the sense that they never have that explosive flexibility of potentially ending a game by sieging production or constantly picking off reinforcements, for example. That last bit has more to do with design than anything else, but that same design is what allows the SH to have its own advantages! With a few simple tweaks I think it could become a little bit more popular of a unit.
So when can you use swarm hosts? The best place I've found for the unit in its current form is by harassment, but more specifically (and sadly almost exclusively) to meet an ultimate goal of cost efficiency. These are the rare really long games where you determine that you can have a win condition if the map gets split or if you get X number of bases as long as you start or continue to trade more cost efficiently. For me personally, these are often times "comeback games" where I decide to get higher risk / reward units or utilize units like the lurker for example to give me 'a chance' to survive against a greater production while forfeiting map control (and any economic advantages) as a result.
In ZvZ I believe one of the strongest/most cost efficient late game unit compositions is:
Lurker, viper, static defense, SH! How much I end up getting of each, or whether I avoid making any 'piece' altogether, varies from game to game. I am not a man of formulas and I don't think that's the right way to play StarCraft, much less Zerg. From the perspective of that particular unit composition and speaking generally:
Swarm hosts are good vs ultra (locusts have so much dps they don't care much for ultra's armor). Because you're basically splitting the map, SH provide value picking off bases, lurkers, spores, spines of your opponent. You can also use lurkers to create smaller sieges and launch locusts from that siege onto your opponent's main to kill tech, or any base. Lurker sieges are very difficult to break cost efficiently, so in many situations, that's how you would safely use SH aggressively once you have all your defense setup, usually not before. SH will allow you to at low risk to slowly pick away at your opponent's buildings and units. You don't have to kill much ; you just need to kill something, anything. This concept only works at a certain part of the game where you're sort of at a stalemate, because the attacking player will be at a disadvantage almost no matter what. The advantage that the attacking player will almost always have is a bigger bank as a result of you forfeiting map control earlier, but as long as you can be cost efficient and trade effectively, which is what your army is designed for, that bank advantage shouldn't be as important; if it comes into play in a big way - then it's likely you didn't achieve your goal to be cost efficient when your opponent attacked.
In ZvT you can use swarm hosts (though I don't recommend making many because of how supply heavy they are) to go up against mech armies that aren't yet heavy on vikings (though I personally think Mech isn't good atm, a lot of T's seem to like it). Flying locusts can do well against tanks and low numbers of thors and they aren't too gas intensive so they can serve as a mild mineral dump when you're looking for one. Of course they'll overlap with other units like lings, roaches and hydras, but there are a few situations in which I would strongly consider the SH over the alternativves. Similar to ZvZ, you likely want a setup to fall back to when going SH. Swarm hosts in both matchups are the units that put a timer on your opponent, a REALLY slow countdown forcing them to engage you on your terms. There's only been a couple of times that I decided it'd be best to get some SH in the match-up.
I rarely use Swarm hosts in ZvP. Funnily enough, it is the matchup in which I used SH the most in HotS (after the change). I think that this is where the most harassment-basey-snipey type SH would come into play the best, but that has issues of its own. Like in ZvT, swarm hosts would hardly be useful unless there is a lack of an air counter army. Alternatively you could also land the locusts and attack them manually, but then that takes a lot more time and attention for possibly still no reward. Stargate(s) / Phoenixes early on are very prevalent in the meta and hard counter SH, here comes the problem that I have with the current cost of SH and at the end of the article, i'll present a simple solution which would in my opinion make them far more flexible without changing the design of the unit.
It's important to keep in mind that there are always other things to consider, ideas that may have gone unexplored for a lack of game understanding or simply because of the constrains and weaknesses that the SH has as a unit.
I'm sure there are plenty of other uses that haven't crossed my mind, when looking at a unit I recommend you think of its strengths, weaknesses and what units it synergizes with. From there you can figure out what situations would be best and when you're in that position you can give it a shot / test. All of these situations I've considered tend to be very rare and somewhat impractical, I think that swarm hosts have a place in the game; but I don't think we'll see them used often at all in their current state.
The Fix: Change (Revert) the Swarm host cost
Swarm hosts with the old cost-supply after their design got changed were still rare towards the end of HotS, so there was definitely no reason for a nerf. I think that Blizzard might have thought they were doing SH a service by making them cost less gas and more minerals, but I strongly disagree.
The old cost was a far better for the SH to become viable because of the similar mineral-gas ratio to the infestor. Even though a lot of people think of gas as the "More Valuable" resource compared to minerals, which is not always the case. For bio terran players that is rarely the case, for example. Put simply, resources gain / lose value based on which you can spare after making the units you wanted to to make without a consideration for your bank, whichever resource is limiting you.
If the cost (and supply) of SH was reversed, you could get an infestation pit, take all the gases you think you need/want and then have the flexibility to decide whether you want SH or infestors. Your bank would be far more ready to go for either. Whereas if you want to go for swarm hosts as a "timing", you want more minerals and less gas; but still need a lair and an infestation pit. This means that often you'd start planning to go SH and shifting your resource intake in that direction without yet knowing what your opponent will chose for their tech path. Say you do this and your opponent opens phoenix or muta.. you're dead! If SH cost more gas rather than minerals instead, you can be flexible with your resources and go for a SH timing attack or infestors to defend with fungal (or to an infested Terran Timing-attack with burrow if your opponent opened stargate, which often means no robo/observer, allowing your infestors to get to their door undetected). Also, the change increases the economical synergy between SH and lings which we talked about earlier in the article, making for stronger attacks. The disadvantage would become the extreme late game situations we mentioned earlier, where gas tends to be the more precious resource generally speaking. But then again, I don't think that would make SH useless in those situations either.
Hard to say whether this change alone would make the SH strong enough (or even too strong) but I think it's a healthy place to start tweaking. I say if you want to go WILD, make the cost 100m-175g and try keeping it to 4 supply, making it even closer to the infestor.
The Invisible Cost of Supply
I was thinking about this and I realized 'supply only costs minerals' - it might seem obvious but there is an important implication to it, the cost per supply of a unit can be simplified, without taking into account things like: being maxed out (where replacing units starts being cheaper because you don't have to make the overlords again with them) or from an abstract but very real perspective of attention and time as resources in SC2 and taking out of the equation however much of that you spend making overlord(s) (or any other supply structure).
An overlord frees up 8 Supply and costs 100 bucks, that means one supply costs you 12.5 minerals.
With this in mind let's take a look at the current Swarm host cost in relation to the infestor, and how the proposed change would create much better synergy in terms of cost, thus making the SH more of a viable and flexible alternative.
So you can see with some basic math and utilizing my art degree, I've managed to illustrate how currently: both units that come off the infestation pit are so far apart from each other in terms of cost that it does not allow for much flexibility at all. Where a swarm host that costs : (3) Supply (100) minerals and (175) gas has great synergy with infestors in relation to their cost. I believe that would translate into players making more swarm hosts reactively without changing the current design, which I personally like quite a bit.
One Last Idea
One more thing that could be done to potentially improve Swarm host design would be to have 2 options with SH :
a) Spawn Flying Locusts
b) Spawn Ground Locusts
This would make the SH more versatile in many situations. Flying locusts are great at harassing, but not so great in heads on engagements, against an army with good anti-air locusts can be rather underwhelming.
You might be thinking, why not just land them? The problem there surges when you view attention and time as resources (and you should).
To land the locusts you need to:
1. Select SH
2. Press Spawn Locusts Key
3. Click attack location
4. Select all of your locusts
5. Press landing Key
6. Click landing location
7. Click attack location
Attention and time are limited by your own mechanical skill and there is a limit for every player ever. The better players will do more things, but they'll still need to decide between those things and where they allocate most of their attention; no one can do it all.
I think having the option of a hotkey that spawns Locusts on the ground automatically would help greatly in making the SH a more versatile/usable unit.