There's been a lot of discussion on the ongoing balance/design issues in WoL with respect to how it goes. Many things I say here have been said before. However, I think that a synthesis of the issues into a cohesive problem identification hasn't occurred yet. That is my goal here.
Now, moving on to the actual content. In my view, the issues we see from the game stem from a fundamental design tension created by what I feel are the three central abilities of each race:
Terran: Stimpack + Medivac
Protoss: Warp Gate
Zerg: Inject Larva
For the purposes of this analysis mirror match-ups will be ignored, since I am looking at how these abilities play off of each other.
None of these abilities in and of themselves are imbalanced (Warp Gate is the closest), and I feel they are great representations of the dynamics and play styles of each race. Therefore, I don't think they should be changed lightly. However, they do create design problems and tensions that need to be solved in each match-up.
Let's look at these abilities in the context of each match-up.
The general design issue that Stimpack causes is that Terran MMM, when microed well with Stim, is incredibly cost efficient against Gateway units (TvP) and Ling/Bling (TvZ). While providing excellent drop play.
In TvP, Terran bio is already fairly efficient on a basic level against Protoss gateway units in a straight-up, a-move fight. The issue, of course, is that Forcefield negates a good amount of this efficiency. However, once Stimpack/Medivac enters the scene, we have the following dynamic:
• Drop play is extremely powerful because small numbers of Terran units are far more efficient than small numbers of Protoss units. Thus, Protoss must either deflect the drop before it hits, or over-commit units to drop defense.
• Movement speed increase combined with Concussive Shell on Marauders makes retreating with Gateway units near impossible (Sentries pretty much are always the first unit to die, and even if they don't, Terrans can often run around the few limited Forcefields left).
• Forcefield loses the vast majority of its effectiveness in blunting the efficiency of Terran bio, thus splash damage is required. This is for the following reasons:
1. Medivacs allow for high-ground vision, providing opportunity to snipe units on top of ramps - the best position for Forcefields.
2. Medivacs allow the Terran player to pick up Forcefielded units and drop them on the other side of the Forcefield, effectively saving them.
3. The movement speed increase makes it very difficult to Forcefield the Terran army in larger engagements, as cutting off too much of the army results in significant damage to the Protoss Gateway ball, and cutting off too little of the army effectively wastes Sentry energy.
Remember that these points apply strictly to MMM vs. Gateway ball engagements (Immortals are grandfathered in because they were once Gateway units), and not engagements involving Colossi, Templar, Ghosts, or Vikings.
Warp Gate Research:
The primary design issue created by Warp Gate Research is the removal of defender's advantage. Let’s take a look at this concept as a whole. Defender’s advantage has the following components:
1. Shorter reinforcement distance for the defender, thus generating localized supply/unit superiority for a short time.
The positioning advantage is sometimes not relevant or can be overcome by having a stronger unit composition. For example, in PvT, the Terran player might be bunkered up on top of a ramp with SCVs pulled for repair, but if he has no Vikings (just MMM) and the Protoss player has 3 Colossi and some Sentries in the gateway ball, it’s going to be very difficult for the Terran player to hold the push. The Colossi allow the Protoss player to overcome the positioning disadvantage.
The advantage that normally does not go away is the reinforcement distance. This allows units that normally “trade” to have different interactions based on where the battle occurs. This is, in fact, most evident in ZvZ. In pure Roach v. Roach battles (I believe Roaches trade evenly with other Roaches in a straight up fight, but I could be wrong), you often see Roach attacks pushed back by waves of reinforcing Roaches due to the fights occurring closer to the defending player’s hatcheries. This localized unit superiority allows for otherwise relatively “even” units to establish temporary advantages due to unit number. This advantage is absolutely crucial to keeping units of relatively same combat power.
Warp Gate removes this advantage. In fact, in many cases, it can make the Protoss reinforcement distance SHORTER than the defender’s reinforcement distance (e.g. Warp Prism with main army allows units to be warped directly into the fight). Thus, Protoss units need to be weakened to return Defender’s Advantage to the opposing player.
The problem with is that when battles happen in neutral or Protoss territory the Protoss player is disadvantaged due to the cost-inefficiency of their units. The Protoss player’s own defender’s advantage is often negated because Zealots and Stalkers trade so badly against similar costs of units of both Zerg and Terran.
The only way Protoss maintains genuine defender’s advantage is through the Sentry (specifically Forcefield). Of course, this causes issues because of the offensive use of Sentries, but that’s an issue for another time. The core point is that the issue of defender’s advantage fundamentally alters the profile of Gateway units.
All in all, both of these points translate into a massive practical efficiency advantage for Terran in the MMM vs. Gateway units battle. Even the Protoss "upgrades" (Charge and Blink) don't greatly affect this dynamic.
Because of this inefficiency, the Protoss "tech" units (Colossi and High Templar) need to be extremely strong to compensate for the weakness of the supporting gateway units, otherwise Protoss would be completely incapable of trading efficiently against MMM. This, however, creates for the very binary engagements we often see in TvP: Colossi survive, Terran gets crushed, Colossi die, Protoss gets crushed.
The Ghost vs. High Templar battle is a little more interesting, but Protoss invariably loses far more High Templar than Terran does Ghosts. This is only mitigated by the fact that 2-3 good storms are often all it takes to crush a Terran bio army. Once again, this creates a very binary situation which is not very fun to watch. Occasionally you see awesome battles like MKP vs. STParting in the GSTL finals they played, but those games are the exception, and not the rule.
The holistic design problem here is that particularly at lower levels, the match-up is heavily imbalanced one way for significant periods of the game, due to the way the Protoss tech transitions work. At higher levels, Terran players have the multitasking to handle the Protoss tech, but at lower levels, you often get very binary battles one way or the other.
The core issue generated by the interaction of Stimpack and Warp Gate Research is the inefficiency of Protoss Gateway units against Terran Barracks units. This inefficiency pervades through the entire match-up and is, I feel, the root of pretty much every complaint in TvP. If Blizzard really wants to fix the game, this is one of the areas they must address. The efficiency gap between Protoss Gateway units and MMM must be narrowed (but not eliminated).
Stimpack is nowhere near the central problem in this match-up, however, the situation it creates is important to understanding the issue, so I will talk about it.
Of the two central issues Stimpack generates against Protoss, only one applies. Zerg can defend against drops much better (Baneling vs. Marine drops, lings vs. Marauder drops). However, the cost-efficiency of well-microed MMM vs. Ling/Bling is still fairly staggering. Zerg tends to perform well in engagements of Ling/Bling vs. MMM only in the following situations:
1. MMM cannot micro (surround/Terran player caught off guard)
2. Terran player micros badly.
3. Zerg has heavily out-produced Terran
We’ll get to number 3 later, but why do I not include flanking? Because flanking is a method to achieve goal #1. By attacking from multiple angles, you make it difficult for the Terran player to micro.
Note that the Zerg player has very limited control over two of these factors. He has no control over how good Terran micro/splitting is, and he has no control over whether or not the Terran player is ready for the attack. The best a Zerg player can do is attempt to set up flanks before a-moving into an army, or engage with units like Mutalisks first to attempt to pull forces out of/into position. In general, the Terran player has far more control over the effectiveness of both armies in the engagement, which is not a desirable design position to be in.
Siege Tanks only compound this issue. Close to the Terran base, flanks become non-optional. Thus, well-placed Tanks make it very difficult to break defensive Terran positions with Ling/Bling centered compositions. Even the effectiveness of pre-splitting Banelings is reduced by the fact that they often have to
• Run into chokes
• Run into layered/staggered Siege Tanks
• Burn through buildings
Thus, particularly when attempting to break defensive positions close to the Terran bases, Ling/Bling relies almost completely on the Terran player making a mistake. Lings in particular have a huge mitigating factor – their speed – which makes this almost acceptable, but from a design standpoint it is still a huge hole, albeit one that is greatly covered by the tactical strength of the Zergling as a unit.
The Infestor is the unit that ostensibly covers this hole. Fungal Growth’s root functionality allows Banelings in particular to engage much more cost-efficiently against Terran bio, and Infested Terran eggs allow for a method of absorbing Siege Tank fire before running in with very fast units (Zerglings). The issue with this solution is two-fold:
• Zerglings cover their own tactical liabilities very well with their speed, allowing them to exploit much smaller holes in defenses than other similar units
• Banelings, due to their low up-front damage and health but strong splash, often create very binary engagements – strong Baneling detonations are very efficient, whereas weak ones are very inefficient (due to the gas the Zerg player expends on Banelings; remember that 2 Banelings costs 100/50).
Thus, against Ling/Bling-based compositions, Stimpack creates the same issues as it does in TvP, but to a smaller extent because the Zergling covers many of its own offensive liabilities and the Baneling covers some of the defensive ones. Like I said before, Stimpack is not the primary problem with this match-up, but it is still important to note the situation it generates.
The primary issue created by Inject Larva is that it allows for “waves” of production, greatly enhancing the Zerg ability to mass specific units at any given point in time. This is particularly problematic in these situations:
• Production of Drones in the early and early-mid game.
• Re-maxing armies in the mid and late-game (also includes ability to tech switch quickly).
The Drone production is only problematic because it allows Zerg to establish saturation on 3-bases inordinately quickly. On 2-base, worker production is still going to be restricted by the actual mineral flow. However, when you reach the third base saturation, the limiting factor becomes the worker production mechanic. With Larva Inject, Zerg can establish 3-base saturation far faster than Protoss and Terran, thus having a substantially stronger economy.
This means that Protoss and Terran players are forced into situations where they must pressure Zergs early to force units to be produced and not Drones, thus controlling the Zerg player’s larva usage.
The second aspect of Inject Larva that is crucial is the re-max. The ability to pool a large amount of larva and immediately utilize it to build large numbers of units means that if a trade occurs, the Zerg player will establish a substantial supply advantage due to his ability to reproduce his army immediately. In addition, the army can be of whatever tech the Zerg player chooses, since all units come out of the hatchery.
In TvZ, the tech switch is particularly relevant since Terran players cannot tech switch easily and require different compositions to deal with Broodlords or Ultralisks. The issue that Inject Larva creates is that it makes Zerg a very strong defensive race in the early and mid-game, as Zerg players can drone extremely hard (establishing a strong economy), only producing units when absolutely necessary to deflect an attack. This is further exacerbated by the Queen serving as both the injector and a reasonable defensive unit, particularly in TvZ.
Inject also allows a Zerg player to tech switch easily, inordinately punishing Terran players for smaller mistakes in army composition (too many Vikings? Oops, how ‘bout 15 Ultralisks suddenly showing up?)
The raw-remaxing power of Zerg can often be seen in mid-game trades, where Terran does not have the infrastructure to reproduce 80 supply of army immediately, but Zerg does, thus allowing the Zerg player to push the Terran player around the map while the Terran waits for the 2-3 production cycles it takes to re-produce the army.
Inject forces the enemy into an aggressive stance against Zerg, as outside of this stance the Zerg player will have both a stronger economy and stronger production. This is bad design, because at no point in the game is the Zerg player forced to attack strategically. He can only be forced to attack via a tactical victory that limits him in one area (generally economy via damage to drone line). Otherwise, a Zerg player can simply sit back and hit the “D” button.
In summary, TvZ is a match-up dictated by Terran aggression, because Inject forces it to be so. However, the Zerg ability to mass specific units at specific times often allows an astute player to over-produce units in the short space he needs to and crush attacks by the Terran player through superior unit numbers or, in the case of Infestors, superior units in general. This gives the Zerg player long periods to go and establish economy, since Terran does not have the infrastructure and economy to produce large armies quickly until the late game.
This production mechanic is particularly troublesome with Zerglings because “safety” Zerglings produced are never useless. In the event they aren’t needed for defense, Lings can be turned into Banelings and set as land mines, or they can be used to counterattack and backstab, often dealing massive damage to SCV lines.
What happens in this match-up is the strengths of Inject railroad the match-up into a specific structure – Terran must attack, via aggression or push – consistently throughout the game to hold back the Zerg ability to drone. However, in doing so, Terran is vulnerable to Zerg over-producing units (via the same mechanic that lets them produce tons of drones) in a short time period to deal massive damage to either the army or economy of the Terran player. The only reason this match-up hasn’t devolved into a complete cluster!@#$ is the fact that Terran has such strong harass options available – Hellion, Marine-Medivac Drop, Banshee – and is thus actually able to hold up constant aggression during the game. Even so, the design issue is present.
Efficiently trading and turtling while also harassing the Zerg player works at the highest level, and is particularly efficient at lower levels where Zerg players don’t have the APM to Inject, Creep Spread, attack, and build units/structures. Thus, Terrans who focus their attentions on efficient harass while sitting defensively behind tanks and turrets have an easier time with straight ling/bling compositions.
The Infestor and Broodlord are the units designed to break this method of Terran play. However at lower levels, players don’t have the macro efficiency to construct full-fledged Broodlord armies against this style of Terran play, whereas at higher levels, the incredible efficiency of the Infestor as both a defensive and offensive unit causes issues whenever a Terran player decides to leave his base.
At the highest level, this match-up isn’t unbalanced from a design perspective because Terran actually has late game inevitability. The Raven/BC/Ghost Army/Tank/Viking is actually very good against Infestor/BL/Corrupter. The problem is that it’s very hard get up, and it’s vulnerable to Zergling counters. However, Terran does have ways of dealing with the Infestor/BL composition; they just need to be buffed a little bit.
I’ve already talked about the basic construction of each of the key abilities, so I’ll just talk about how they interact in this last section.
Warp Gate Research’s primary functionality is, in fact, almost completely negated by the standard Protoss opener – Forge Fast Expand. By the time Warp Gate is available, Zergs can establish a strong enough production and economy line to negate the advantages the ability confers via sheer numbers.
The issues present with Inject, however, are further exacerbated; primarily the droning issue. Forge Fast Expand essentially cedes all ability to be aggressive for the first 6 minutes, by which time a Zerg player can establish a 3-base economy. In addition to this, the Protoss player cedes map control by not building units, and thus has no clue what the Zerg player is up to. Even a single wave of units can be utterly devastating if the defensive Zealot is taken out and a Sentry is not ready to take its place. Without the ability to see what is happening on the map, the Protoss player is completely at the mercy of the Zerg player.
The problem is that a 3-base economy catapults the Zerg player into the mid-late game essentially from the start. On 3-bases a Zerg can sustain large waves of production which the Protoss can’t on 2. This translates into a massive supply disparity which forces one of two reactions from the Protoss player:
2. Play defensively until a third is established and mining
The problem with #1 is that all-ins are strategically weak and rely on the opponent being unprepared. Even with the existence of the immo-sentry all-in, forcing such play is still bad design and doesn’t indicate a balanced match-up. It’s not any sort of actual thinking or play going on when the Protoss player is railroaded into a specific build if they want to win.
The problem with #2 is that it gives the Zerg player free reign to drone, thus giving them a huge economic lead while still allowing the Zerg player to do damage via harassment (Mutalisk and Zergling are the biggest culprits here). This severely reduces the impact of Protoss tech options. Mid-game 2-3 Colossi armies can easily be crushed by a mass wave of Corrupters in addition to the standard ling-roach whereas Templar are generally much less effective against Roaches, due to the high health of the unit.
The Sentry is what’s holding this together. What’s going on here is the opposite thing that happens in TvZ and the necessitation of Fungal and Infested Terran. Forcefield needs to be as strong as it is because of the early and mid-game aggression Zerg is capable of doing against Protoss. Because Zerg production can ramp up as fast as it can, Protoss needs a way to engage small numbers of units on small numbers of units. Otherwise, Roach Max would just win every PvZ ever. So when you complain about Forcefield, it’s really there because Warp Gate interacts with Terran and Inject interacts with Protoss the way they do. The most elegant design solution to both problems is the Sentry. It is in fact, a quite excellent solution, but doesn’t actually deal with the underlying design problem.
All in all, the central tensions generated by the core mechanics of each race define the way each of the match-ups plays out. In order for StarCraft II to be a genuinely fun and balanced game, these tensions need to be dealt with in an exciting manner while not compromising the creative vision of each race. Thus, I feel that the two design goals should be as follows:
• Lower the production ramping capability of Inject Larva, thus forcing the Zerg player to build a mix of drones and units more frequently while also investing in extra hatcheries (Zerg invests, by far, the least in infrastructure of the three races, further compounding the production issues created by Inject).
• Raise the straight-up combat efficiency of Gateway balls against both MMM and Roach/Ling while not ruining defender’s advantage in the early game.
If you want specific balance suggestions (and I’m by no means sure this is a good idea), this is what I would do:
• Lower Inject Larva production such that 2x Hatchery + 1x Queen per base is equivalent to 4 Hatchery + 3 Queen on 3 base (a frequent production composition today). This can either be done through Inject time, Larva Number, Inject Cost, or all of the above.
• Change Zealot damage to 7 (+1 vs. Light, +1 vs. Biological) with +1/+0/+0 per upgrade [Zealots do 9 damage per hit vs. units that are both Light and Biological]
• Raise Zealot base move speed to 2.4
• Increase Stalker Damage, Fire Rate, Health or all of the above.
• One of the following:
o Warp Gate Research now requires Twilight Council combined potentially with swapping Sentry to Robo and Observer to Cy Core.
o “Warp Crystal” Pylon upgrade – Allows units to be warped in within the upgraded Pylon’s power radius. Upgrade should cost both minerals and gas (100/50 would be a good start), and have a reasonable construction time (say 20 Blizzard seconds). This could be combined with further buffs to Zealot/Stalker or changes with Sentry tech as well.
• Leave Stimpack untouched
Anyway, those are my thoughts. I hope to get some good discussion.