[MSL] Ro4 Day 2

On February 07 2011 14:19 hacklebeast wrote:
*SPOILER* I assume most people know how the Ro8 ended, but If you still want to be surprised, stop reading

With midterms done, I wanted to spend a little time breaking down the unknown territory that is Hive tech ZvZ (sorry Jaedong and Hydra fans, but this is going to be all about Zero and Great).
I. Trends
II. Unit roles
III. Application prediction
IV. afterthoughts

---

I Trends
I wanted to find some commonalities in Great and Zero’s hive games. Zero and Great both have played 4 such games (3-1 records) with great winning their head to head matchup. What is interesting is that in Greats games, he was the one who went for hive tech first (which from now on will be referred to as “the instigator” from now on because instigator is a cool word and because it is a lot easier to say than “the one who went for hive tech first”) only once, and that was the game he lost.
Here is a list of the games I looked at for reference
+ Show Spoiler +
Trend 1: The instigator always has an economic BO advantage
In all seven games, the instigator either had a 12 pool vs 9 pool or a 12 hatch vs 12 pool (one case of 12hatch vs 9 pool, but since he survived, he still had an economic advantage). This makes sense; the player with more money is able to get away with cutting mutalisks to get the nest+hive+greater spire. So the question is: can we see in instigation with the same BO, or even a BO loss? I don’t think so. The tech already sacrifices nearly 1000 resources, which is 5 mutalisks. Adding that you will need 2 spores per base, we are now down 7. We could compensate by only building spores, but then a control group of lings will win the game.

Trend 2: the instigator will attack with about 2 control groups of lings and fail.
This was personally the most surprising trend, and it happened in every game with the exception of Zero vs. Great. I also want to include the trend “the instigator will always have a later spire (and in most cases a lot later),” but its factors in here. When the instigator sees his BO win, the thought process goes something like this “I have the BO advantage, so I should be in a good spot. My enemy has started his lair before me, so I can either abandon the advantage and try to play catch up tech wise, or I can just attack now with my two hatcheries against his one.” This can work in other games and just win outright, but because all of these games didn’t end fast, we only see the failures. Our instigator is now severely behind in tech and is forced to get spores as the only form of tier 1 anti-air (ignoring hydralisks for now). He could tech up, power more drones (he still has the eco advantage) and eventually catch up, but the other option is to just accept that he won’t be able to get the same number of mutas for a while and find other ways to equalize.

Trend 3: an unsupported mutalisk force never beats a supported one
Specifically we are talking about queens and devourers. The “why” for this one is obvious, but it can be surprising just how much a couple of queens can compensate for. In the most extreme example 12 mutas, a queen and 2 devourers beat 30 mutalisks (with equal scourge numbers.) It also means that the not instigator is forced to either follow the instigator in tech, or try to win immediately.

Trend 4: The winner always has more bases.
The exceptions to this were Zero vs Hoejja and zero vs roro. Again why the person with more bases wins is obvious, but the question is not “why does the person with fewer bases lose,” but, “why does the winner need more bases to win?” And the answer is they don’t. We will talk a bit more about this later, but in most of the games the winner takes more bases to finish off a game they are winning already.

---

II Units
The rolls of zerg units in the is matchups are murky due to the lack of examples, so let’s try to add a bit of clarity.
drone
Starting off simple. The drone’s role only changes from a poor fighter vs zerglings in the early game to no combat use at all later on.
overlord
Again, a simple unit, but the overlord severs a lot of minor roles beyond its primary function. The overlord provides an easy and invincible scout in the early game (duh). In the mid game it serves as a meat shield in muta battles. While mutalisks will naturaly target other combat units, the splash damage does not. Every shot the overlord takes is a shot that a mutalisk didn’t take. If the overlord takes 150 damage, than that is at least one mutalisk saved, but as far as we know could be 3 mutalisks with 50 health each. As we move to the late game, overlords are needed for detection vs lurkers. They can also drop, but there was only one game with drops, so we’ll ignore this for now.
zergling
Oh the zergling. It does everything for zerg short of flying. It serves the high dps role of the marine, the protect the ranged unit role of the zealot, the minefield/tank fire drawing role of the zealot, and even the “oh shit I’m losing bad, let’s just send some guys into their base and hope they don’t notice” role of the DT. That’s not to mention that they cost 25 minerals each and can transform into a cloud that makes your units invincible to most others.
Specific to this MU, The zergling a los forces the opponent to play more cautiously. You can have map control with a large air force, but it’s a different kind of map control compared to a mech push. While the air force can easily counter a large scale attack, two packs of six zerglings each presents a unique problem. If they are sent to attack different locations, then the opponent is forced to either have already set up defenses (costly, susceptible to run bys and unusable if the enemy isn’t brought to it) or the air force must split (results can be catastrophic is half of the larger air force is attacked by the entire smaller one). The zerglings force caution, and can severely punish over aggression.
Mutalisk
The other staple of the MU. Oddly, there is little to say about such a fundamental unit. They harshly counter everything below it on the tech tree, and get harshly countered by almost everything above it. There fast, fragile (which can be compensated for through stacking), and have reasonable attack in sufficient numbers. But we all know this. As the game progressive, they lose their effectiveness as devourers, swarm, ensnare and plague dramatically reduce their ability.
scourge
Their role is similar to the mutalisk in that it shrinks rapidly, but even more so. Once the mutalisk numbers start to get above 25, we see “corsair syndrome” where the scourge just all die before they can attack. The numbers required to attack would not only not be cost effective, but also take super human APM to clone. After the battle moves to the ground, however, they can have a role similar to their role in ZvT: trying to snipe queens and guardians that stray a bit too far from the main force.
It is worth noting that in great’s game vs Jaedong, he used scourge to kill overlords. He may have been trying to set up the lurkers he was teching to, or he might have been trying to supply block JD so he could tech safely tech. Either way, its intriguing. Ultimately he lost, but, hey, it was against Jaedong. Who can blame him?
queen
Of all the units in Broodwar, the queen keeps me up the most at night thinking “there has to be a better way to use these.” Luckily, Zero shares my sentiments, and has been trying to answer the question of how to use them right for the past year. Since there is no holy world in the map pool, we only have to worry about 3 abilities: ensnare parasite, and broodling. The fact that melee units don’t bunch well means that ensnare only works on the mutalisk flocks, but oh is it powerful. Want to engage but have fewer mutalisks than your opponent? POW ensnare just effectively took out 18 percent of them (not to mention that they can’t muta micro any more) Won the battle, and now they are retreating? POW ensnare allows you to clean those game-changing six mutalisks up. Did he just reinforce his mutalisk force? POW ensnare allows for a clean escape. Is his muta micro too gosu for your scourge? POW now he can have all the apm in the world and your scourge will still kill him. Broodling has been suggested as an equivalent to irradiate, and that a late game ZvZ could see clouds of queens in similar fashion (see P.S.). I think that broodling can take on that role, but having a cloud of queens is overkill. That many queens is not necessary (a) because the zerg will never hit 200/200 like it can frequently in TvZ (b) due to the spread out nature of lings, lurkers lose a lot of effectiveness, and do not need to be feared to the same extent and (c) because we do not have the massive M&M ball, zerglings can pick off hanging defilers quite effectively. The only thing that the queens should be overly concerned about is the ultralisks, and the 2-3 that are left over from the air battle I think can be sufficient. The last ability, parasite, is a bit of a mystery to me. I (and dare I say no one who will ever read this) will know how helpful that scouting information can be to a progamer. Yes it provides info, but don’t progamers understand what is going on without that information? And don’t they have other scouting methods that don’t waste APM or energy? Zero did use a few of them against great, so there is some value. I guess I feel that they just aren’t worth it especially when considering that the zerg units are just so cheap. And if the queen finds an expensive unit, it would be better of just killing it.
Devourer
The devourer, like ensnare, turns unfortunate battles into victories. I encourage everyone who doesn’t understand the acid attack of the devourer to check the liquipedia page http://wiki.teamliquid.net/starcraft/Devourer (go ahead, no one will think any less of you ) The devourer is the last thing seen before the player switches to the ground, so their role is limited, but if the opponent does not either switch to the ground or get devoureres of their own, their fate is all but sealed.
Guardians
The Guardian has the same role as they do in ZvT: a quick way to throw away all of your gas and shortly after lose an unexpected change of pace that can potentially kill all workers and cripple the tech tree if not dealt with. I don’t like them personally, and would favor teching do defilers instead. I can almost see them like carriers, staying out of range of the hydraliks and slowly whittling down the numbers. The problem is that carriers don’t have to deal with the smaller range, lower health, lack of an air attack, swarms, and less speed (I think). Soulkey was the only one who used them (in the circuit breaker game), and I admit they did do a lot but I think that they are a mistake to use.
Hydralisk
Hydralisk are terrible. They do half damage to zerglings, half damage to mutalisks, have the potential to be utterly decimated by guardians (again see soulkey’s usage of how to do it right and great of how to counter it wrong) and are helpless to zerglings and ultralisks that are under swarm. Then why are they so utterly important? Because they are the only thing zerg has that can attack air while under the protection of the swarm. The ensnare, devourers, and plague do a lot to force the transition from air to ground, but having invincible units that can attack you makes it compulsory. But after the mutalisk flock is gone, there is no good reason to use them. Great begs to disagree, and produces them as his first hive unit every time. Its hard to argue with him, considering that the one hive game he lost was against Jaedong, and was killed before actually producing any units. But considering how little they do vs zerglings, and how swam makes them useless, I don’t understand his love for them.

EDIT: There has been some confusion on my stance on the hydralisk. Let me reiterate: THE HYDRALISK, WHEN COUPLED WITH SWARM, IS THE SINGLE MOST IMPORTANT FACTOR IN MOVING THE BATTLE FROM AIR TO GROUND. The distaste I express towards them is after the opponent has moved to an ultralurkerling style army. Hydralisks are bad at combating lurkers (think of them as expensive marines) and the fact that they take 8 shots to kill a zergling means they cant fight them either. I guess that they could take on an army of pure ultra rather cost effectively, but that is unrealistic. And this is all even before swarm. Once swarm hits, the hydralisks have no value in the ground battle. So the hydralisk is very valuable in the small window where you opponent has no ground army, but are not worth it after that. Great has the right idea of using the hydralisks to counter Soulkey's and zero's mutalisk clouds, but once each of them switches to a ground army, it is my belief that Great should stop hydralisk production, which he never does.
Lurker
The lurker is what I am least sure of in this MU. They can’t fight air, so we can only look at how they fight against ground units. Ultralisks just ignore them. Hydralisks just die, but hydralisks are terrible, and everything kills them as long as they don’t have swarm (and then just most things kill them). I question how good they are vs zerg lings. First thought is “attacking a lot of little units with splash? Only two hits kill one? They sound like a perfect counter! But we have to consider that zerglings do not bunch u-p the same way as marines do, and with good splits, they start to only hit 2-3 per attack. Also consider that they cost lots of gas that could be used on ultras or defilers. In the games they seem to do enough, but I’m just no convinced. In my eyes, I think that the ideal use of them is to morph them after you are done with the terrible hydralisks (I kinda like just adding “terrible in front of the hydralisks, so I will for the rest of this), then place them strategically with some defending bases and some with the main army, mostly to force the use of detectors. But as I said, they might be a staple to the ground army.
ultralisk
Not much to say on this front either. Ultarlisks tank damage so the lings don’t. Once the defiler/lurker tech is up they can be a nice addition to push the army over the top. The only game with ultras was zero vs great, and I found the ultra:ling ratio to be terrible, like 1:3. I guess it could have something to do with the abundance of lurkers, but I don’t think it is something to take terribly seriously.
defiler
And the defiler. Arguably the most powerful two spells in the game to have for the matchup (although maelstrom and either storm or irradiate would be rather scary), and consume to boot. The defilers (with the terrible hydralisk) take the mutas out of the sky, and the defilers allow defensive positions to be established that wouldn’t otherwise be possible. It is interesting that every time defiler tech is reached, there are always double mounds. It seems extreme at first, but considering its only an extra 100/100 to get consume and plague that quickly, it made me wonder why zerg doesn’t do it in other matchups. I then realized that plague isn’t that good against the other races.
The only other thing I want to bring up is what day[9] once said about how to properly plague in ZvZ. It was in one of his dailies if you want it in audio, and articulated better. He talked about how when you should plague isn’t before the engagement, or even during. If you engage then retreat, the enemy will pursue and stack up as tightly as they ever will (assuming were at the point where we have more than 1 control group so the stacking method is obsolete) and then you plague. I thought that was a cool idea, although I didn’t see it in the games.

---

III Application
A recap of how hive ZvZ should look:
One player gets an economic advantage, but has slow mutalisks. He defends himself by spring his natural and main. Instead of just making mutalisks, he decides to tech. The opponent fears the tech, and is forced to tech himself. The game moves into muta flock vs muta flock with queen and devourer support. One player realizes that they won’t win that kind of fight, and switches to defiler tech. Once that player gets a good number of defilers and terrible hydralisks, the other one will, in turn, be forced to move to the ground. Then we get ultralurkerling vs ultralurkerling, and we get all the strategic elements of TvT, without the half hour of posturing.
So, what was the point of everything I’ve written? It’s so we can look at trends to make some predictions. Before we jump into the games, I need to address a discrepancy. Those who have watched the games and those who picked up on the discrepancy while reading might ask, “wait, the way you described you “ideal” evolution of the game, and how you described the unit’s roles does not match up with great’s style. He won 3 of four and beat zero using the progression you described. What up with that?” Yes Zero uses the strategy I described (that makes it sound like I’m Zero’s mentor doesn’t it. I guess I should say “I described the strategy Zero uses”) and yes he was beaten by great and his terrible hydralisks. The reason that Great won is he treated it as a modern game of starcraft. I’m going to compare it to a starcraft game in 2003 between a veteran (like Boxer) and Great is like a young hot shot (like Iloveoov). Boxer is experimenting with a young game making new exciting unit compositions while Iloveoov says “fuck it, I’m just gonna get a bunch of bases and kill you”. And that’s what Great did. He remembered that this still is starcraft and macroing and expanding are just as (and probably more) important than the micro and unit composition. Zero did expand in his wins, but he did it against weaker competition, and less vigorously than Great did. (for the record that’s not supposed to be derogatory towards any of the four players mentioned I have re3pect for all of these guys, and Boxer and Iloveoov are my favorite two retired starcraft players)

So when we look at the maps, we are looking for 2 things.
(1) There is a long rush distance so that early ling pressure won’t win outright.
(2) There must be a super easy third gas or a super hard third gas to take *
(3) There are relatively few expansions
*Ok that sounds weird. If there is a super hard third to take (like say aztec) , zero can prevent it, and great will be forced into a 2 base game vs someone who understands the matchup better. If the third is super easy to take (like a double nat map), then zero will be able to secure it easily and will still be able to compete economically with great.

So, here we go:

Circuit Breaker Zero < Great
There has already been a hive game on this map, and that’s as good of an indicator as any that there might be another. That third is a bit more than a hop, skip , and a jump away from the natural, but is it unrealistically far? I think that if we go hive here, Great can find a way to secure it, and win the macro game.

Triathlon Zero < Great
The site of hive ZvZ game that they had before. This map does have the double expo I described earlier, but there are just so many expos. While Zero can get up to 3 bases, great can get 4 or 5. Great takes set 2

Dante’s Peak Zero>Great
The location of the third is similar to circuit breaker, but the difference is that you can’t choose one of two equally far away gases, and that is just difficult enough to say that Zero can stop it.

Benzene Zero>Great
The third is so far away here. Great is going to have a bear of a time securing that, especially considering he abandons air fast. Zero equalizes.

Circuit Breaker Zero>Great
I could copy paste my first one, or I could say that zero figures out what he did wrong and adjusts. Zero takes it with his awesome queen usage.

It should be said that there is no reason to think that there will be 5 hive games. So many things have to go right. First Zero (considering that great is reactant in all of his games) must want to instigate, then there must be a BO win for Zero, then great must decide that he doesn’t want to muta all in when he sees zero going for a hive (he probably won’t, its mostly the first two). But hey, this would be rather short if we just said that there will be no hive, and just flip a coin five times right?

---

IV Afterthoughts
While watching the games I thought of a few things that I would like to see in the MU.

1. Drops. Zero did one of these and it worked brilliantly, I’m not sure why it didn’t happen in other games. Cracklings have ridiculously high DPS, and they are dirt cheap. Couple that with the fact that Zerg bases are significantly smaller than terran or protoss ones (only 2 factories (hatcheries) and no depots). It can easily take out the workers (which is even scarier in ZvZ because you can’t produce drones and units at the same time) and can definitely destroy a tech tree. And it is just so cheap.

2. Burrow. I think that this has a lot of potential. It can save drones from a ling run by. It can hide defilers that get separated from the main army. It can set up ambushes (like either the terrible hydralisk ambushes on shuttles, or ambushes on mutalisk packs with defilers (like in the day[9] trick I talked about earlier). It can set up zerg scouts to see where large mutalisk flocks are. It can deny expansions. At the very least it will force your opponent to get overlord speed, which costs more than burrow did.

---

And that’s it. Thanks for reading
PS:+ Show Spoiler +