Swarm Host Design Flaw: Assessment and Solution

1. Introduction
2. The issue
3. An example
4. Why is this a problem?
5. Opportunity cost of delaying Hive tech
6. Examples of Tier 2 units that provide utility and allow for transitions
7. Recent changes to the unit
8. The solution
9. Conclusion
10. Summary of key points
11. Vods/Replays of Swarm Host use
12. Q & A with CultisChan

I would like to begin a discussion on the design of the Swarm Host. I believe that there is a major issue with the design of this unit and how it is actually utilized in competitive play, and would like to see what everyone’s opinion is on the subject.

This is not a balance whine/balance discussion/QQ thread, and I do not feel the Swarm Host is overpowered or underpowered. Rather, this is a discussion about how the unit is actually used and the type of gameplay and situations that arise from investing into the unit.

I believe that I have been able to discuss why the unit does not always function the way it was intended to and does not always promote dynamic, interesting gameplay.

In addition, I believe that I have provided a concrete solution that would fundamentally fix some of the issues that players are encountering in using the Swarm Host, and would help the unit to better fulfill its intended role.

I have added a Summary at the end explaining my main points as briefly as possible, but I encourage you to read the entire post if you have the time. There are a lot of finer points that cannont be articulated effectively in such few words.

After testing the unit at the GM level in the beta, and watching/hearing other players use the Swarm Host, it appears that the Swarm Host is very weak in low numbers, but begins to gain considerable strength in very large numbers, roughly 15+. Really, to use this unit effectively at a higher level of play, you need to invest heavily into the unit, as making a small investment in Swarm Hosts will not benefit your army enough to justify the investment.

The problem that this relationship (weak in low numbers, considerably more effective in large numbers) creates is that by investing into the unit, you essentially become all-in, and the outcome of the game is decided by whether or not you can set up a deadly contain and break your opponent’s defensive position.

Let’s use an example here. The matchup is ZvP, and the Protoss player does a Forge Fast Expand while the Zerg player takes a fast third base before taking gas. Protoss does a build that allows him to secure a third base around the 9 minute mark, and in response, the Zerg player decides he wants to pressure his opponent with Swarm Hosts. The Zerg player knows that 5-6 swarm hosts won’t create a critical mass of locusts needed to break the Protoss players defensive position with Storm or Colossus on the way, and in response he/she proceeds to make 15 swarm hosts. This represents an investment of 300/300 for the infestation pit and the locust upgrade, and a 3000/1500 investment for the Swarm Hosts themselves.

This investment is so huge, that if Protoss can break the contain that Zerg has established and start picking off the Swarm Hosts, the Zerg player won’t have any possible way of holding some sort of counter-push, as he invested all of his resources into making the Swarm Host push a success. Zerg is all-in and has to win the game right here or the game will end.

On the other hand, this army can be devastatingly strong, and gives the Zerg player the opportunity to win the game outright if the Protoss player cannot deal with the Swarm Hosts in time.

This all-or-nothing scenario is a problem, because if your investment into a critical mass of swarm hosts is not successful, the Zerg player cannot transition.

So what kind of composition are we seeing that makes the Zerg all-in in games? In a typical ZvP, the composition will be almost entirely Swarm Host/Corruptor/Overseer, with the Corruptors serving the purpose of sniping Obs/Colossi/air units.

How can the Zerg transition from this? He really can't just start sacrificing Corruptors, as they are providing the support to keep the Swarm Hosts alive. He really can't just start sacrificing Swarm Hosts, because then his army would be weaker, and if the Protoss is spending all his resources into breaking your position (as he must, with the exception of Zealot/Warp Prism base harass etc.), then he can now break through and win the game.

Even if you wanted to transition into a Broodlord/Corruptor/Infestor composition, you don't really have the infestor/infestor energy necessary to make the transition without maintaining the vast majority of your Swarm Hosts. At this point, having a few extra Broodlords vs Swarm Hosts wouldn't change the dynamic of the battles in the ZvP example, and wouldn't change Zerg being all-in with his one army in one location on the map.

Of course, if you make an investment that puts you all-in and it is not effective, you should be very far behind. But the Swarm Host was never designed as a unit that puts you completely all-in when you invest in it. It was designed as a way to apply some real mid-game pressure to a turtling opponent, in a very “Zergy” way.

MVP TheSkunk has provided his input on this topic, and I feel that he does such a good job of summarizing why this specific problem that I will be quoting him in this section. Thank you for your contributions

MVP TheSkunk's thoughts on facing this critical mass of Swarm Hosts, from the Protoss perspective of a PvZ match:

"It seems like you have to reach the tipping point where there are enough Locusts spawning that they can survive long enough (and clog the pathing of enemy units) for their sick DPS to take effect. With too few, the locusts are too easily killed or ignored, but once the right number is reached it is impossible to engage them without taking massive losses.

You really have to be absolutely confident that you'll be able to kill actual Swarm Hosts when you engage an army, because if you just kill locusts (and no Hosts) then the Zerg has lost nothing and you've lost real units -- and they do damage so quickly, you can't really afford to be indecisive. You have to choose. This also contributes to the "all or nothing" feeling."

MVPTheSkunk provides his thoughts on the difficulties that Zerg players face utilizing this unit:

"This thread reminds me of a game of Idra's I saw on his stream. It was a ZvZ, and he was ahead. He decided to get a lot of Swarm Hosts to contain his enemy.

Unfortunately, he misjudged and had barely to few Swarm Hosts to contain the enemy, and as a result lost all of his Swarm Hosts and was quite far behind.

This is just an example of how the way it works sucks for the Zerg player and not just their opponents -- just like it is hard to engage a critical mass of Swarm Hosts, if you fail to reach the critical mass then you lose all of your Swarm Hosts and are massively behind."

I would like to explore a little bit what the opportunity cost of investing into the Swarm Host entails, and why some players would rather tech straight to Hive rather than invest heavily into a Tier 2 Zerg army.

In Wings of Liberty, both in ZvT and ZvP, the unit composition that has dominated the Zerg lategame in the last few months has been infestor/broodlord. This has proven to be an extremely strong unit composition, and successfully teching to this army often meant that the Zerg players chances of winning were extremely good. This composition has proven itself in Wings to be one of, if not the most reliable way for Zerg to win a game, and by investing into things that are not infestors/broodlords, one is delaying/avoiding this winning composition.

Heart of the Swarm is a new game, and Blizzard has taken steps to making the Infestor/Broodlord composition considerably weaker by nerfing two spells of the Infestor: fungal growth and infested terrans.

However, Zerg has received other buffs to its late-game Hive-tech, including the addition of the Viper and the improved Ultralisk. These two units are proving to be extremely powerful in the lategame, and acts to fill the void that the infestor/broodlord nerfs (infestor nerfs really, but the composition relies on the strength of the infestor) have left. The following game illustrates the power that Blinding Cloud and Ultralisks can have against a Lategame Terran army:



To understand why this is significant, let's look at some of the stats of Swarm Hosts and Vipers, specifically the amount of time it takes to gain access to them. Currently, Swarm Hosts are available at the Infestation Pit, and require a 120 second 200/200 upgrade to achieve their maximum usefulness. In contrast, it takes 140 seconds for a player to morph hive, and spawn any number of Vipers.

The issue here is that Vipers are proving to be extremely powerful, even game-changing units for the Zerg Swarm, and fulfill the siege-breaking role of Swarm Hosts much more effectively for less cost. In addition, it only takes 20 extra seconds to have Vipers compared to fully functional Swarm Hosts.

Why would someone invest into Swarm Hosts, when they can have a unit that performs the same role, only much more effectively and for a lesser cost, and the only downside is waiting an extra 20 seconds for the unit to be available in its fully functional form? This creates a problem, where Vipers/Hive-tech is simply more effective than the Swarm Host, and barely takes any extra time to gain access to it. There should not be a situation where two units of the same race perform the same function but one is just better. No one would build the weaker unit, but that is the case here.

I would like to provide an example of a Zerg unit that allows for a significant investment to be made, reaping the benefits of the unit, but still allows for effective transitions into the later stages of the game. The unit I will use for this example is the Mutalisk in Wings of Liberty ZvZ. We are going to assume that, in this mirror race game, one Zerg player go 2-base lair into Mutalisks, and another will go 2-base Lair into Infestors. There are several advantages to going Mutalisks in ZvZ. The Mutalisk player gains map control, can clear the map of overlords, has an opportunity to try and deny his opponents third with his Mutalisks and a burst production of zerglings, and prevents his/her opponent from moving out without anti-air. In the words of Dustin Browder on this unit, “Party, right?” [1]

But what I find so interesting about this scenario, and so different from our Swarm Host ZvP example, is that there are diminishing returns to investing into the Mutalisks. Typically players build around 8 Mutas, and assuming the opponent is not also going Mutas, will begin to transition into other tech. Having more than the initial 8 Mutas does not offer too much of an advantage to the player, and he/she can accomplish everything he needs to with those units. He won’t necessarily die to a counter-push after this investment, and can begin to invest in other tech, including double upgrades, infestors, Roaches, Hydras, and even Swarm Hosts.

These transitions create interesting, dynamic gameplay, and allow for a more exciting playing/viewing experience. Furthermore, they encourage players to invest into the unit in the first place, because there are different ways to transition out of the unit after the investment is made.

In HOTS Beta Patch 8, Blizzard buffed the health of the Swarm Host from 120 to 160. David Kim explains “The reason we went with a health buff is to allow easier repositioning of Swarm Hosts between Locust spawns. We think this type of positional play is one of the most fun ways to use Swarm Hosts, and wanted to promote more of it.” [2]

I do not feel that this type of buff will significantly alter how Swarm Hosts can be used in a competitive game, and I would like to explain why I feel this way.

This buff will indeed allow players to move around and reposition their swarm hosts effectively in between locust volleys. However, the issue still remains that you must be able to justify the huge investment into the unit, and because you really can’t transition out of the unit, you still need to end the game with them. This buff does allow for more micro-potential of the unit, and we may see some very interesting ways that players siege their opponents with this change. Unfortunately, this does nothing to address the do-or-die nature of the Swarm Host.

I would like to discuss possible solutions to the design flaws of the Swarm Host. Several members of the community have offered their feedback on what could be done to improve the unit, and I will include their input here as well.

i. My thoughts

I think the main thing that need to be addressed right now is the low health/high DPS nature of the Swarm Hosts’ locusts. This relationship adds to the snowballing effect that occurs with the unit. The low health of the Locusts ensures that a smaller number of Locusts will be unlikely to reach their destination to inflict damage. On the other hand, when the locusts do reach the critical mass required to inflict damage, the results are often devastating.

I propose changing the Locusts to being units with higher survivability and lower DPS. This could include increasing the hit points of the locusts, while also decreasing the amount of damage inflicted per shot. By making this type of change, the locusts will be more likely to reach their destination to chip away at the opponent, but the effects will not be as devastating as they are now. The locusts could truly act as support for your army, instead of the other way around.

ii. Engineer's thoughts

has had a lot to say about the unit, and I would like to try and summarize some of his suggestions.
-Give Locusts an attack more similar to Broodlings, where the initial attack does more damage immediately, and the subsequent ones do less. This would be a way of frontloading some of the damage, while still allowing Locusts to eat away at the enemy if left untouched
-Increased survivability of locusts through increased speed, health, and collision radius of locusts. Decrease in DPS of the locusts to compensate.
-Roaches morph into Swarm Hosts, because it would be awesome.
-More burrow utility. This can be done by allowing locusts to burrow move, and allowing burrowed roaches to morph into Swarm Hosts

iii. Existor's thoughts

Existor has brought up some interesting points about how the Swarm Host might function better if Locusts were air units, instead of ground units. There were some older sketches shown at the last Blizzcon of an older “air version” of the Swarm Host, so this is something that Blizzard has considered in the past. Existor has done an excellent job of summarizing his arguments, and I will quote him rather than butchering his explanation:

-Widow Mines will still counter them, as they still hit air
-Stalkers will be able to shoot them
-Phoenixes will have good damage against them as light units
-Thors will counter them as they have anti-light damage
-Zerg Swarm-Festors will not work anymore, because you will need anti-air against Locusts. Hydralisks/Corruptors/Fungals to stop those small flying kamikazes
-Siege Tanks and Colossies will not work against Locusts anymore -> more usage for swarm hosts
-New locusts will be able to fly meaning that cliff becomes less a problem for Zergs
-Because Locusts can hit only ground, every air unit will counter it, if it has air-to-air attack
-Allows ground zerg units not stuck between locusts
-Forces enemy to make more air. Many current [units] already counters new kamikaze-Locusts. Phoenixes, Thors anti-light will be very usefull, because locusts are small and they always will have some splash, corruptors will be more useful as air-to-air counter against locusts, fungals will stop locusts from reaching your units

A more detailed explanation of this possible solution can be found at Existor's original thread:

http://eu.battle.net/sc2/en/forum/topic/6160896658#1

iv. The Lurker

Finally, I have had many people respond to my post, both at Battle.net, and at TeamLiquid.net, suggesting that the Lurker would be a better choice for a Zerg Tier 2 space control/aoe damage dealer/siege unit, and that we should just replace the Swarm Host with the Lurker.

I feel that there are some extremely valid reasons for feeling this way, and many people have brought up some very good supporting arguments. I will do my best to explain my reasoning for why I have not taken this stance:

It is my goal in making this post to help Blizzard understand what the current flaws of the unit are, so that Blizzard can make the changes to the unit that it feels are best. I do not feel that Dustin Browder is interested in bringing back the Lurker in this specific case, and intends to see the Swarm Host as one of Zerg’s new units with the expansion. This can be seen in many interviews, where he expresses his opinion that the Swarm Host is incredibly “Zerg-y,” and that the concept of spawning locusts works. If I were to have as my recommendation, “Ditch the Swarm Host, give us back our Lurkers!” I don’t think Dustin would take my thread very seriously, and he may just dismiss it entirely. By trying to work with Blizzard’s unit, rather than against it, I feel I am in a better position to influence Blizzard in making some positive changes to the unit. If anyone else believes that the Lurker would be a better addition to the Zerg in HOTS, I urge you to say so, and explain why. This is not a position that I can afford to take, given my goals for this thread.

That being said, if someone could offer a detailed argument for why (re)introducing the Lurker would be more appropriate than keeping the Swarm Host, I would be happy to it here.

I would like this to be a discussion about how everyone has been able to use the Swarm Host, its effectiveness in small vs. large numbers, the all-in nature of the unit, and possible ways to improve the way the unit performs in competitive matches. I encourage everyone to discuss how we can make the Swarm Host a powerful tier 2 option, without necessarily making the unit stronger or overpowered in large quantities.

I believe that there is still time to address these problems of the Swarm Host in the HOTS Beta, but because the unit suffers from a design problem, simple numbers buffs/nerfs will not fix the unit.

I. The Swarm Host is largely ineffective in small numbers (4-8), but becomes exponentially more effective as you reach a critical mass of the unit (12-15+). This is a case of increasing returns.

II. Investing into a critical mass of Swarm Hosts represents a huge investment. It is extremely difficult to transition out of this composition, as the investment essentially puts you "all in." Battles can become very lopsided very quickly, much like in lategame ZvP with Archon toilet vs Broodlord Infestor. One side often wins very convincingly here, and comebacks are not possible.

III. There are very attractive Hive Tech options for Zerg in HOTS, particularly Vipers and buffed Ultralisks, and to convince people to invest in Tier 2 tech options, there need to be very powerful reasons to do so. Vipers are available 20 seconds after Swarm Hosts with their locust upgrade, and unless the Swarm Hosts offer the Zerg player some key advantages, he/she will tech to Hive every game instead if playing to maximize winning chances.

IV. This combination of weak in small numbers and effective in large, the inability to transition out of the unit, and the extreme desirability of Hive-tech options relative to Swarm Hosts create for boring gameplay, where the player can go all-in with Swarm Hosts, or tech to Hive as fast as possible. Many people complain that strategy of teching fast to hive every game has made ZvT/P dull and boring in Wings of Liberty, and i we are to make Zerg games more interesting, Tier 2 tech should have reasons that justify investments into them.

V. I strongly believe that the correct solution to these problems is to increase the survivability of the Swarm Host's locusts, and to decrease their dps. By making this design change, the locusts would be much more likely to reach their intended destination to inflict damage, especially in less than huge numbers, but would not be able to break entire positions by themselves in critical mass, like they currently can. Instead of the Swarm Host being the main the main damage dealer, they would exist to dramatically increase the effectiveness of the rest of your army by tanking large amounts of damage. This would allow the Swarm Host to more effectively fulfill the siege role that Dustin Browder intended for the unit, and still in a "Zerg-y" way.

*This section is a work in progress, and will be added to / edited in the coming days*

I would like to kick off this replay analysis section a game played by Blade55555, a GM Zerg in the Beta and a high-masters Zerg in Wings. The game can be found at the following link, andit is shown from the Zerg player's point of view:

http://www.twitch.tv/blade55555/b/347106538

The game is a ZvP on Daybreak, and it begins at roughly 1:50:00. I would like to focus on a few key engagements that occur in this game, as I feel they do an excellent job of illustrating the points I have made.

Zerg spawns in the bottom left hand side, and Protoss on the top right. The game opens with a standard FFE vs 3 Hatch before gas, and the Protoss elects for the famous immortal sentry push, ultimately hitting the third base by around 10:00 into the game. Zerg has to sacrifice his third, but is able to somewhat stabalize on two bases and the game continues.

Blade chooses to go for a Swarm Host-based midgame, with the first Swarm Hosts being built around 10:40. By around 11:30, Zerg has driven back the attack for now, and reclaims his third base. Blade continues to invest in his Swarm Host army to hold off the growing Protoss force, and by 12:30 has 10 Swarm Hosts in his army.

The first engagement I would like to cover occurs at 12:40. Blade tries to use his army of 10 Swarmhosts and lings to defend his third base, but some well-placed forcefields trap the locusts, and allow the Protoss army to walk up to the third, snipe it, and walk away unscathed. There are things Zerg could have done to prevent this, but the point I want to make here is this: Zerg had 10 siege units sieged, worth 200/100 each, and Protoss could walk up next to them, deny their ability to attack, snipe a base, and walk away unscathed. For a siege unit, this comes accross as a little unintuitive.

By 14 minutes, Zerg has reclaimed his third again, has 16 Swarm Hosts to try and defend with, and Protoss still has not taken a third base. It is difficult to tell who is ahead here, but it certainly looks like Protoss is in a good spot.

At 17 minutes, another major engagement occurs. Protoss has not taken a third, and is going for a brutal attack with 4 colossi/sentries/immortals/stalkers, while Zerg is defending with 18 Swarm Hosts and 49 Zerglings. In the engagement, the locusts are not able to dent the Protoss army, and Zerg is forced to retreat to his third. He loses some spine crawlers, but does not suffer any major army loss. Zerg is backed into a corner, and Protoss looks primed to engage the third of Zerg for a killing blow.

If Zerg has an engagement like the last one at his third, he won't have anywhere to retreat, and his army will get obliterated while the Protoss player won't suffer any losses. However, Zerg has Corruptors about to spawn, and this may be enough to defend with.

The final battle occurs at 18:30, and Zerg has 13 corruptors to defend. These corruptors, combined with some fatal Colossi miscontrols, lets Zerg snipe the colossi, and proceed to win the battle without suffering any losses.

So, what is the point of all this, and what have we demonstrated?

Battles involving Swarm Hosts can be incredibly, incredibly one-sided in ZvP. In some situations, the locusts are denied the ability to inflict damage, and there is nothing left to defend the Swarm Hosts from their imenent doom. If Zerg loses a large army of Swarm Hosts with no compensation, the game is essentially over.

At the same time, battles with Swarm Hosts can be incredibly one-sided in favor of Zerg in ZvP. In this game, Zerg took out an entire army without suffering losses, winning the game.

This is a huge problem, because it creates an environment similar to that of lategame ZvP in Wings, where the fate of an entire game can be reduced to whether or not Protoss lands one key Vortex.

The Starcraft community has raised concerns over this coin-flippy type of gameplay, and has collectively agreed that it needs to be removed from the game as much as possible in HOTS. Unfortunately, late-game ZvP has been given a new life, and that life is in Swarm Host ZvX Mid-games.

I believe it is important that my arguments presented here are able to stand up to tough criticisms by members of this community, CultisChan has asked some difficult questions, and has forced me to continue to articulate my arguments better, and to provide more relevent examples of games and situations that support my main points.

I would like to present to you a conversation I had with CurtisChan from page 4 of this thread:



Your fix makes the SwarmHost to be some sort of abysmal support caster unit, which at its core violates Dustin Browder's imago of SwarmHost's ideals.

That beautiful bald man smiled and nodded as the world hung on to his every word, here:


And the audience went silent as they saw the locusts do absolutely nothing. I understand the tanking goal you have in mind, and I'd like to see a version that actually worked. Because as is, there's no way this change will allow Swarmhost play in timing attacks, as investing in something that offers no offensive value, or ability to properly tank/absorb AoE is going to be like using Ultralisks in WoL. Transition to loss.

Also note that Zerg has always suffered most heavily from AoE damage. Adding 20 or 40 hp to locusts won't change the cataclysmic effect massed siege tanks or collosi will have on clumped up range 3 units. Both of which will be out in time to counter your newly crippled SH.

Swarmhosts need to remain a threatening presence that punishes turtlers. Your changes would further harm their usage, as the only thing they HAVE going for them is their critical-mass-spirit-bomb technique. Which can effectively be one shot with a sick nuke, or less awe inspiring fungal chain.

You make many good points about their cost being a complete waste if you don't have enough. But doesn't the same risk/reward apply to the Terran if they waste their siege tanks willy nilly?

What exactly were his original intentions for the unit? If patch eight has taught us anything, it's that Dustin Browder is as capricious as a Greek God and this is all really a game to him.

But I digress, the original intent of the swarm host was having locusts that could attack ground and air. A wonderful notion indeed, scooby!


Hi Cultischan, thanks for stopping by! I appreciate the feedback that you have provided here, and hope this isnt the only post you make here.

I'm going to do my best to address each of the points you have made individually, as they are certainly valid concerns. This will be a long post, as I want to address all of the points you have brought up thoroughly :D

It looks like you bring up the question of whether the Swarm Host should be a support unit or the damage dealer itself, the weakness of locusts to AOE damage, the need to have a threatening presense when going Swarm Hosts, the risk/reward inherent in investing into the units (compared to, say, a siege tank), and the original intentions for the unit. Let me know if Im missed anything there!

1. The nature of the unit being a support unit, or the frontline unit itself. You mention that, by making the locusts high health/lower dps, the unit would be relegated to a support function in the army, and I would absolutely agree! As it stands right now, the Swarm Host is your main army, and everything else functions to keep them alive. I think that this creates a serious issue, and prevents the Swarm Host from synergizing with the rest of your army in more moderate numbers, and forces you to go allin with the unit when you make it.. Let me use a couple of examples to illustrate my point.

The Viper is a Tier 3 support spellcaster, with two main engagement spells: abduct, and blinding cloud. What makes these spells interesting, and fundamentally different from, say, the spells of the Infestor, is that the spells do absolutely zero damage by themselves. The Viper exists to allow the rest of your army to engage armies and fortified positions that would otherwise be impossible, and acts to strengthen the rest of your army.

Because the Viper does not do any damage on its own, there are very strong diminishing returns to investing heavily into it, and many players are finding that between 4 and six Vipers is all they need to accomplish what they need to. Vipers spawn 20 seconds after you would have access to Swarm Hosts with the locust upgrade, and despite not having any actual attack, open up tons of timing attack potential, and gives your army legitimate offensive capability. So its not that Support units like these dont allow for offensive/timing attack potential in general, because there are support units that pop at similar timings that offer this.

Here is a recent example to illustrate the offensive capabilities that the Viper, a pure support unit, offers to the Swarm:



For our next example, let's look at Wings of Liberty PvT, and our focus will beon a stalker/colossi based composition, to combat the standard terran biological army. For the sake of using an example, we will examine the first game of the following link:

http://sc2casts.com/cast10594-LiquidHero-vs-Beastyqt-Best-of-3-NASL-4-Season-Play

LiquidHero decides to go for a stalker/Colossi based composition in this game. The question here is, which unit is the support unit, and which is the actual damage dealer? In my opinion, the stalkers have very low dps, and their role in this composition is to keep the Colossi long enough that they can deal their damage. This relationship is very evident throughout the game, and I feel it is relevent to our Swarm Host example.

With the current iteration of the Swarm Host, the unit IS your main damage dealer (not unlike the colossi), and the rest of your army exists with the purpose of keeping your Swarm Hosts alive. Why is this significant? Well, unlike the colossus, Swarm Hosts really can't be used in small numbers, because their locusts are so vulnerable that it takes reaching a critical mass of locusts to reach the enemy to do damage. So you invest in this huge Swarm Host army, and if you can't justify the investment by doing an incredible (almost game-winning) amount of damage, you will be behind. You can't really transition out of Swarm Hosts, because that is your army in this composition, and you haven't invested enough into the strength of the rest of your army. Contrast this to the game with the Viper, where the Viper exists to make the rest of your army more powerful.

I strongly feel that the Swarm Host can fulfill a useful support unit function, rather than functioning as the core of your army, as it does now. This design change will allow for a more moderate investment into swarm hosts to be useful, both in the midgame and in the lategame, and allow the player a better opportunity to transition to later tech if so desired.

2. The weakness of locusts to AOE damage

I agree with you that the nature of AOE damage really limits the effectiveness of what the Swarm Hosts can do right now. However, where we disagree is the benefit that can be achieved through increasing locust health.

If we nerf the locust damage somewhat (as Blizzard has already done twice in the Beta), the unit becomes less all-or-nothing, and would crate less situations where the locust reach this critical mass and suddenly nothing can stop them. By being less all-or-nothing, I think we would have the opportunity to really push the locust health up there, possibly much more than 20-40 health.

Because I love examples to illustrate points, lets look at some of the design changes that Blizzard has implemented so far for the Raven. On the one hand, Blizzard kinda said "well, the Raven needs a cool new spell, but we're out of ideas. Wait, the patch is coming out tomorrow? AAAHHH! Quick, someone take the stats of that spell we never see from the BattleCruiser and throw it on the Raven! They won't notice!"

On the other hand, this type of change is interesting for a number of reasons. In Wings of Liberty, the raven suffered from two major issues from a design standpoint. The first, is that the Raven was really, really hard to tech to in a normal game and it was very difficult to justify an investment in them, because that investment probably wouldn't be realized in a normal game. On the other hand, if the player actaully managed to tech to them, especially against Zerg, teh results were often absolutely insanely one-sided:



This unit created an all-or-nothing type of gameplay, not unsimilar to many situations involving Broodlord/Infestor vs Archon Toilet, especially before Zerg really honed in on how to control this late--game composition. I think that these type of one-sided battles, which make comebacks impossible, aren't necessarily optimal from a gameplay perspective, and that steps can be taken to make them less extreme. Fortunately, this is something that Blizzard has done with the Raven.

The Raven right now does not require an upgrade for high seeker misssle, and the spell itself is a little bit less extreme, doing 300 single-shot damage vs 100 stackable splash damage. Whether or not you think Ravens are viable right now in HOTS TvZ, or if the Seeker Missle spell is in a good spot, Blizzard has taken an all-or-nothing unit/spell and has made it more accessible, while also making the spell itself less volatile.

I believe , that when people use the Swarm Host right now, it behaves very similar to the Raven in Wings of Liberty. Very often, the Zerg player can't quite establish the sieged position he needs with a critical number of Swarm Hosts that he/she needs, and the opponent can break their position, winning the game, before the investment has paid for itself. Equally often, the results are similar to the linked IMMvP vs Nestea game, where the investment pays for itself a thousand times over and the game ends right then and there.

I believe that a step away from this type of volatility associated with teh Swarm Host would allow the unit to be more easily balanced, and would encourage players to invest in it more in high-level games. I also believe buffing locust health and nerfing locust DPS would be a huge step in this direction.

3. The need to have a threatening presense when going Swarm Hosts

You brought up some really great points here. You say that, when you build Swarm Hosts, there needs to be a threatening impact on the flow of the game, and that by making the unit, you should be able to accomplish something with them. You also seem to be of the opinion, that by making the unit into more of a support role by nerfing locust DPS and increasing Locust health, that you won't be able to have this forceful impact on the game.

First, lets look at an example of a unit that can have an immediate, powerful impact on the flow of a game that can be built at Tier 2 for Zerg. I would like to discuss the Mutalisk. We are going to use an example from the GSL Semi-finals here:

http://www.gomtv.net/2012gsls5/vod/71073

The game I am referencing is the first one, and should be free to watch. I think this game erally goes a long way in showing how much a Tier 2 investment can change the flow of the game. When the Mutalisks pop around 12:30 in game, the entire flow of the game switches. Suddenly, instead of defending, Sniper is in Terran's base, forcing the Terran on the defensive. Sniper uses this 7 mutalisk investment to take complete map control, force terran to have defense everywhere, and makes it more difficult for him to move out. These seven Mutalisks won't win the game for Sniper, but they are clearly paying for themselves and fit the role of a powerful support unit really effectively. They also have the sort of immediate threatening impact that you allude to in your post.

So, how is this relevant to Swarm Hosts? I think that the current iteration of the Swarm Host does not allow for this sort of immediate impact due to a fatal design flaw, and that my proposed changes would address this perfectly.

Swarm Hosts are fundamentally weak in small groups, due to the vulnerability of locusts to splash damage, and only gain power when they reach very large, critical masses. The video example you referenced from Blizzcon last year is a perfect example of this! The Zerg player has 8 Swarm Hosts, a very sizeable investment, and can't seem to dent a somewhat fortified terran position. Now, what if the Zerg player had 16 Swarm hosts? The results would be completely different, as the high dps locusts would begin to reach their target, and become exponentially more effective.

Because the Swarm Hosts are so effective when reaching critical mass, they need to be balanced around this fact. Blizzard realized this when they released the Beta, and subsequently nerfed Swarm Host locusts 3 times in a row. The Swarm Hosts were fulfilling their role well in moderate numbers, but a case of exponentially increasing returns made it so that, instead of stopping at 5 or 10 Swarm Hosts, people were building 20+ and nothing could stop them. I will try and provide video examples of this later, but I do not have any offhand right now.

But what if we could make the Swarm Hosts effective in smaller numbers, but weaker in very large numbers? This way, the unit would have that threatening presense you talk about when it first hits the map, and, like the Mutalisk example of Sniper, can have an immediate, tangible impact on the game when they spawn. I believe that my changes address this concern, but will require some actual testing before I can confirm this.

4. Risk/reward enherent in controlling units (using the Siege Tank as an example).

You make a very good point, that if someone miscontrols any unit, they should be punished for it. If someone has their siege tanks or colossi out of position, then they can and will be picked off by a competent user. You raise the question, how is this any different than when a player gets his Swarm Hosts picked off?

Well, I feel the major diference here, is that when a Zerg player loses his Swarm Hosts to a frontal attack on them while sieging an enemy position, the issue isn't that he miscontrolled the Swarm Hosts, but that he didn't reach a critical mass of Swarm Hosts = locusts to properly utilize the unit in the first place.

Siege tanks don't need to be massed to be effective. If you have 2-3 siege tanks added in a ZvT or a TvT, they are often enough to make a huge difference in the flow of the game. The Colossi behaves in a similar fashion, and in PvT/Z, each incremental Colossi that is added onto the Protoss army in the midgame gives the collective strength of the Protoss army a huge boost in value.

Now, production mechanics for Zerg are very different than Terran and Protoss, and the unit must be balanced around them. It would not be realistic to give the Zerg a colossus, because of their ability to produce many of them at once, when the Colossus is balanced around a much slower potential rate of production.

That being said, I do not feel that the Swarm Host is in a good place right now with respect to the number of units needed to have an impact on the game. Going back to the Blizzcon Swarm Host example again, the Zerg had 8 Swarm Hosts sieging a position. THat is an inevstment of 1900/1100, assuming current unit/upgrade/building costs. This represents a huge, huge investment into the unit, and yet it is only able to tickle the opponent.

What if, instead of sitting in his base, the Terran army just stimmed out, scanned, and started picking off the units? This is possible because there is not a critical mass of the unit out. WHat if there were 15 Swarm Hosts, instead of the 8 or so shown in the vidow? Well, the situation would be completely different, and the Terran assault would likely be unsuccessful.

I am trying to find a way to address the design issue where you need this critical mass, or all-or-nothing quantity of the unit to be effective on the battlefield, and I feel that my changes go a long ways in correcting this.

5. The original (Dustin Browder) Intentions for the Swarm Host

Ah! Last but not least, we get to talk about what the role of the unit was supposed to be in the first place! Rather than making myself look foolish, I will be writing down here what the Rock himself said at last year's Blizzcon where they unveiled the idea for the unit:

"I wanna talk about a new zerg artillery unit called the Swarm Host. Now this guy is a great way to gain map control, and what he does, is he burrows and spitts out for free, a bunch of little timed-life units called locusts, which swarm accross the map. Itt's a very Zerg way to get an artilery unit, and a very Zerg way to get a bit of static map control.... We wanted something they could use to also fundamentally apply some pressure. We wanted some way to force them to deal with the Zerg who had some kind of fundamental advantage."



Sorry if I didn't quote him perfectly there, but I think it was pretty close. So, Dustin seems to think the unit is to:

-gain map control
-apply pressure to a turtling opponent, in avery zergy way

I strongly believe that the current itteration of the Swarm Host does not accomplish either of these goals until you reach a very, very large number of Swarm Hosts (12-15+). Anything less than that, and the locust waves aren't large enough to break down a fortified defensive position, and they aren't really strong enough to defend themselves or an area out on the map (map control).

Interestingly, Dustin mentions that the locusts were 90 health in the video, so they have indeed brought down the health of locusts since its first iteration.

I still believe that my proposed solution is likely to be most effective in addressing these issues. However, my main goal for the unit is to be successful in the expansion, and to create interesting, dynamic gameoplay while giving the Zerg player a Tier 2 option that is worth investing, and provides unique advantages that they could not otherwise have access to.

I would like to thank you for bringing these criticisms to light. It is very important that my arguments can stand up to harsh criticism if they are to be taken seriously by Blizzard and some of the Blue Posters, which was the main motivating factor behind making this thread in the first place. If you would like to continue our discussion of the unit I would be more than happy to do so!

*Now, for a question for you about the Swarm Host

You seem to share my dissatisfaction of the unit in its current state, but disagree with my proposed solutions. So, how would YOU fix the Swarm Host if you were Dustin Browder, the Rock himself?

My problem with the Swarm Host is that the doesn't unit functions well with other Zerg units (at least in PvZ) and doesn't offer anything different in PvZ really.

The Swarmhost lacks synergy with the fast moving Roach and Speedling, so it becomes a different form of play in the same way that Colossus play differs from Templar play in PvT. That is fine, but they cost a lot, you need a lot to make them work and the Swarmhost doesn't work that well with Ultralisk or Broodlords, limiting transitions. Compare this to the Infestor, which you can invest a lot into during the mid-game and use them defensively or offensively and then they synergize well with Broodlords or Ultralisks in the late game. Swarm Hosts do not synergize well with Ultralisks or Broodlords in the late game, so if you want to transition into the late game from the mid game, Infestors are still the best choice by far.

And when you are building large amounts of Swarmhosts, you have to end the game with that composition due to their cost and the fact they don't synergize with Broods or Ultralisks (you won't be able to protect your Broods like Infestors can, and Ultralisks will just get blocked by the Locusts and Fungal won't be there to hold units down for Ultralisks) so they will be best used for all-ins. Thus in many ways, they function like Hydras do in PvZ, good for some all-in timings, but not viable for late game play.

And so Zerg can build a bunch of them and hit a timing with Corrupters and Hydras or Infestors or whatever, but the Protoss player doesn't really need to do much different than standard play, teching up to Colossus and building a few for the lategame deathball (note that unlike the Swarm Host the Colossus works well in the mid-game and then transitions well as part of the late game deathball). It is these Colossus that allow you to defeat Swarm Hosts. So you simply need to scout your opponent, realize they are heading for Swarmhosts, and cut your tech and produce a lot of units. So you hold them the same way you'd hold any other all-in. The Swarmhost then just provides another all-in opportunity for Zerg in PvZ, rather than a new unique and interesting style of play.

I haven't had many issues with Swarmhosts as a Protoss player. I lose to them when I don't scout for them or prepare for them properly as I should. I have been playing versus Masters and GM Zergs lately.