http://us.battle.net/sc2/en/forum/topic/16201180104?page=1#0
Disclaimer: This thread will be updated as I test the new balance test map more. It is likely the balance test map/mod will be updated and therefore it is ideal to make valid suggestions to make each test as worthwhile as possible.
These are the thoughts for the version 1 of the balance patch for swarmhosts.
Hello.
I have had the pleasure of playing a couple of games on the new mod as well as see the series played last night on Nathanias stream. I will use these games as a point of reference.
Note: Any games played throughout the balance testing are not necessarily representing the best, fastest or smartest plays but instead seeks to reveal the key aspects of the new design/balance.
_____________________
These are my suggestions for an update to the balance test map.
- Allow the swarmhost to move while burrowed. My suggestion is 2.25.
- This may require a burrow-movement animation.
- Make locusts automatically attack after swooping down to the rally-point on ground OR make locusts automatically swoop down once a nearby enemy unit is in range, not a building, and then attack. Currently there is a bug where they swoop down to a building and not attacking - which is a huge problem
- Decrease/increase the cost from 100/200 to 150/150.
- This may change depending on goal of compositionary design. 200 minerals and 100 gas is an option, so is 150/100 or even 100/100. I think 150/150 is the best point to start with a resource cost buff.
_____________________
I think it is a mistake to buff locust health or damage. Currently, 65 hp x 2 from 5-6 swarmhosts is enough to kill an entire workerline if undefended or unpulled, and the current 0.6 attack speed with x2 12 damage makes them even able to snipe a town center (Nexus, Command center, hatchery.) from a single wave. That is mathemathically 800-1000 gas (after suggestions added, 600-800 gas) and likewise mineral cost being able to "freely" kill a base town center. That is an extremely cheap harass tool at highest effectiveness. Theoretically, as long as swarmhosts roam without interception, this is the best harass tool in the game. I therefore think buffing these values on locusts would not be as ideal as buffing resources that promotes zerg compositions rather than the single unit itself and would not only make swarmhosts "stronger" for harass, they will become stronger in deathballs again which is countering our intention of our unit redesign.
I am open to buff the swarmhost however, not locusts. Increasing HP from 160 to 220 or base armor by 1 or 2 points is reasonable now considering the cost. However, doing this may require NOT to change resource costs. I also want to note that if swarmhosts are attacked and is dying it is most likely because of lack of compositionary supportive defence from zergs rather than the swarmhost needing armor og health points. Below is an example of how succesful swarmhost harass attempts work.
http://gyazo.com/49bd70f94b16dbbeeed385ceab5f9f49
http://gyazo.com/61c8cf06f06ad3c4c75f58878e86471c
In these two screenshots we see 7 swarmhosts, only 4 actually being used to send locusts (8 locusts) because the rest are on cooldown. This could easily be a scenario around the 12-13 minute mark instead.
After that, the locusts flies into the protoss main base and even with a full chargelot warpin, the locusts would, if microed, still be able to kill a lot of probes. This is considering the zerg ability to still attack other places and, perhaps, have two groups of swarmhosts to attack. This is why we want to increase the compositionary potential of zerg rather than buff the swarmhosts.
So what exactly should we do to the new swarmhost?
Right now people seem to be mostly concerned with PVZ since there are two major problems with the new swarmhost and playing versus protoss:
1. It is too weak against protoss deathballs
2. It is too expensive to make compositions, to actually perform harass or to defend in general.
I have already seen very good aspects of the new swarmhost and I see great intention when using it, but all these good things has to play around the viability in the game. Now, note that the balance patch has JUST been released and people are still experimenting, but I want to note a few things as a guidelines to how it "must" be used.
1. It is opened in midgame to have most harass efficiency.
2. It has 200 gas cost so it empties many possibilities for "supportive unit" compositions and might also be easily scoutable due to this fact.
3. The new swarmhost must go on a hyper offensive since using it´s ability at home, blindly, will be a huge risk to especially protoss timings.
Let us start with 1: It must be opened early to have more harass efficiency. Why?
This is because the swarmhost role is an harass unit. This is best defined that the unit contributes very little to deathball armies. It is similar to saying that a reaper is a harass unit that will be useless to open at the 12 minute mark in any terran matchup. This forces the zerg to move out almost instantaeneously and reveal the strategy, making very simple defence mechanics for the defending protoss player whom is not worried about mutalisks or the usual lategame zerg strength per se. I see this as a problem.
Protoss does not have to change build orders or compositions to play against this, in fact, protoss is opening a "remotely strong midgame army" since the main unit defining the composition is blink stalkers - which actually happens to be one of the best compositions to have against the harassing locusts, obviously because they can shoot air, but also because of the strong mobility.
If we look at:
http://www.twitch.tv/nathanias/v/3736411?t=1h20m01s
- at 1 hour, 20 minutes and 1 second in the game between goswser and minigun, we see that our zerg player feels almost forced to make early units due to the heavy gas cost on swarmhost and the fact our protoss player still can base his midgame opener around sentries. This means that in order to get the swarmhost for the midgame position you want for harass optimally, you almost have to cut on the early unit investments and instead have the swarmhosts as close to your opponent at perhaps latest 11 minute. This exposes you heavily to allins and timings.
Personally I do not believe with the current 200 gas cost that it is possible for zerg to make anything but roaches and lings (And statics + queens) while getting out swarmhosts, and as soon as the swarmhosts are out, blink will be finished since the timings match only in a window of perhaps 30-60 seconds, blink getting out faster if chronoboosted. It is possible to get hydralisks, but protoss compositions should outmatch that due to the expensiveness of swarmhosts. This is rather "obvious" from the current games, but this statement is not necessarily true, however highly likely.
If we imagine that the TVP blinkstalker allin from one base and two bases were the headline of this balance patch test, it would be similar to seeing terran having stim and combat shields and concussive shells everytime a protoss tries to attack with blink stalkers for the first time. Now, this weakness for the attacking protoss player might be a slight overexaggeration to how we see zergs initiate an attack on protoss players with the new swarmhost, but I think there is an automatic off-timing since the emphasis of the zerg strategy is the swarmhost, basically, solely because of the 200 gas cost at the time being.
I would like to see 10-12 swarmhosts being opened around the 12 minute mark supported even by hydralisks. This could theoretically be possible since this is not a unit you necessarily are supposed to MASS like you do with mutalisks, making gas totally exhausted only making mutalisks, but a reasonable change of cost to 150/150 may invite zergs to do compositions while also not being able to just open 30 swarmhosts and walk in and kill a base without a player can do anything about it. In examples shown, 3-4 swarmhosts is also more than enough to cause havoc if the defending player is unaware of the threat. Maybe 1 swarmhost for each mineral line while just playing standard roach hydra viper? I refer to the "endless possibilities" again.
I also believe the suggestions I have made would be more than fine for TVZ. The iaguz versus petraeus game on nathanias stream as linked is a good example. However, I can not go into detail about this yet without much more thorough testing since my main source of knowledge is ZVP.
This accounts for the reasoning behind changing the resource costs as mentioned in point 1, 2 and 3. First, to address the issues both with the new swarmhost´s ability to harass but also the compositionary design followed by opening with swarmhost. If their harass and composition-design problems are fixed, the swarmhost strength as a unit versus lategame protoss SHOULD be more or less irrelevant.
About the burrow change, I believe it is ideal for swarmhosts to be able to move while burrowed - and spawn locusts while burrowed - not only does this contribute to their harass effect in lategame if a player lacks detection on the map (similar to an invisible ghost walking past units to make nukes) but also actually makes burrow useful. You would be able to escape scans or units without the need to burrow or unburrow. Also, due to the mobility of the new swarmhost, spending time burrowing and unburrowing seems to be worse than just leaving them unburrowed.
The second game of minigun vs goswser can be found here.
http://www.twitch.tv/nathanias/v/3736411?t=1h39m01s
A double melee/armor upgrade, plus hive, plus +2 melee attack and speedlings with a 4th base barely gets 10 swarmhosts out by the 13 minute mark. (with more or less better macro) This is a crucial point where protoss would simply be able to do a 3 base blinkstalker colossus allin and, likely, just automatically win. I believe goswser tried to make a lategame macro opener through melee upgrades to transition into ultralisks - a definitive and smart strategy considering the open map, but what goswser was surprised by is the fact that swarmhosts empties his gas _twice as fast_ as mutalisks does - and if not mutalisks was a unit that is expensive. In the production tab, goswer had to skip on +2 melee armor likely because of the fact he realized he was getting far off timing with the swarmhosts.
Then later, at 1:40:05, we see why burrowed movement could be ideal for the new swarmhosts. Other zerg units often do other things like multipronged attacks while the new swarmhosts uses it´s ability to flank your opponents base. Now, goswser used the locusts to attack the army - arguably not how it is supposed to be used honestly - and then his swarmhosts were left vulnerable. Protoss can have detection, maybe they can not - the point is that the expensiveness and lack of anti-deathball strength means you can not support them as easily as you have been used to. I definitely believe burrow movement can help zerg players do these kinds of "naked" swarmhost movement on the map and be safer against for example phoenix pickups or blinks. Basicaly losing the swarmhosts here costs goswser the game.
Even if swarmhosts had more health points or armor, they would still die every time in this situation at bad trading efficiency. This is why I think the best option for blizzard now is to change resource cost and perhaps burrow utility.
As mentioned, buffing locusts any more would probably hurt the intention of the new swarmhost and make it both too powerful versus armies and as an harass tool. The following example is of a game I played showing what a reasonable number of locusts can do if you are busy defending some other place:
http://gyazo.com/da3803bcb1d8b572fc806109f9319034
http://gyazo.com/e8cb5fe45e3e615d7cca636148679862
At first we see how zerg attacks while sending just 4 swarmhost worth of locusts into my main. Note: This game has been going on for some time and I, protoss, has been behind because of an unexpected queen swarmhost attack. - but just slightly behind, arguably.
Just moments later, while I was busy defending against queens with voidrays, this happens in a matter of seconds.
http://gyazo.com/0d7eada51f6f51a895e9d99aaebe71f7
And I want to emphasize This is just 4 swarmhosts.
Imagine if a slight decrease in cost of swarmhost would do while zerg does not make any more than just 3, 4 maybe 5 swarmhosts and then starts an ordinary, non-swarmhost oriented playstyle. Maybe infestors? queens? Statics? Ultralisks? Upgraded range or melee units?
Or even mutalisk.
The zerg used in this game the nydus network to quickly find positions to swoop locusts in to attack. Meanwhile, the zerg managed to bank mutalisks and, together, mutalisk and swarmhosts could be ideal, if resources are not too expensive. You can probably not mass swarmhosts... maybe massing mutalisk can compensate? Remember, protoss will likely need a counter-attack with blinkstlaker colossus or stargate phoenix and phoenix range to deal with mutas. The expense for protoss could arguably be even, depending on game, with the expense zerg has to invest with mutalisk swarmhost.
And what would not be cooler to see both mutalisks and swarmhosts in play?
http://gyazo.com/05682b650649d72fb5ab0db10af5a424
http://gyazo.com/05682b650649d72fb5ab0db10af5a424
My zealot just BARELY kills the nydus with some swarmhosts at the exact moment the zerg attacks with mutalisks. Imagine if this nydus network actually had not died, perhaps if it was spawned at the highground where I would not know about it?
A mutalisk + swarmhost combination, if zerg can manage, would be an overwhelming and cool harass tool that I would really like to test if truly is a good playstyle. However, the expense in gas is surely a problem, but any way for zerg to deflect damage and buy time with tier 1 units or likewise could make many possibilities with the new swarmhost come true.
I definitely want the swarmhost to have a slight cost reduction to 150/150 and perhaps allowed burrow movement to really improve the quality of further tests. I will edit this thread with more or better examples with other high-caliber, top gm zerg players. For ZVT, when time comes, I will likely make a seperate thread.
Hey wait, what about lower levels? How are they affected by swarmhosts?
I would first say I believe lower level players can answer this question better than I can, but I think the defending players versus swarmhosts will have more trouble than the swarmhost user. Just like with oracle, widowmine drops, allins, the new swarmhost will more "freely" be able to harass lower-level players and I think they will do more damage than in higher level games. The new swarmhost is "arguably easier" to use than the old, but with the various bugs and stuff this is not final to say.
Never the less, I think my suggestions will be great for lower level players as well - especially allowing burrow movement. And the new swarm host will definitely be better than the old for new players as well since the new swarmhost probably is more exciting to play with and against.
I hope we can see the balance patch hit the live game as soon as possible, but this definitely requires some testing to ensure "balance".
Thank you.