Maybe We're Looking At Deathballs Backwards - Page 5

Now, speaking generally, as the size of your army increases past a certain point, additional units begin to give diminishing returns. This is especially true for units with short ranges (particularly melee units). However it also has a significant effect on powerful units and positional units. Adding another unit to the big force yields less utility than putting that unit somewhere else.

SC2 is firmly stuck in accelerating-returns-land due to the small numbers of expensive units. Colossi are a paradigmatic case. Two colossi is MUCH more than twice as strong as just one colossus. Against units that the colossus can actually attack (discounting units like vikings, corruptors, void rays for the moment) they actually never stop having this kind of wildly accelerating returns. Twenty colossi is 120 supply- ridiculously impractical for a real game- and they STILL will never be less effective than multiple smaller groups of colossi. It's only twenty actual units with twenty distinct chunks of HP, and each one is so expensive you can't afford to send it off alone. Sending a colossus out to fight by itself is a waste of resources. Not so for smaller, cheaper units.

Marines (often considered a deathball-y unit in SC2) are actually quite different. They are 1 supply, and an expendable 50 minerals. You can actually send, say, eight, or sixteen marines out on their own (usually by dropship for mobility) and you gain strength by doing so. Splitting the marines in a strategic sense across the map makes marines more effective. However a big army will kill such a small group with zero casualties. So in order to fight a big group and actually win, you have to ball up yourself. Herein lies the reason deathballs are dominant- one player is incentivized to use them, and forces the other to do the same (Prisoner's Dilemma). If they don't, their small forces get destroyed for free, or with minimal damage to the deathball.

However, suppose the units are smaller and cheaper, which allows the tiny squads to meaningfully engage a deathball, and do damage without being destroyed for free. They still lose of course, but they inflict *casualties*. This happens when small units are used by both sides. Not with marines fighting colossi or thors or anything huge and expensive, but with marines fighting BW hydralisks. A big army of 1 supply hydralisks vs a small force of marines will win, and sustain casualties. The marines are fighting efficiently by splitting up. Colossi and Broodlords win, and sustain no casualties primarily because of their size and cost, and also because of their mechanics.

There is no compelling gameplay reason apart from "coolness" or gimmick-factor why bigger units would be desirable. Thors, for example, are not a better unit for being stronger and more expensive. It only reduces the number of options a player has for building and controlling that unit. One thor versus three goliaths, for example. What if I only want one goliath, and don't build the other two? Or three goliaths can be in three different places, controlled separately.

More relevant to deathballs, smaller units die more often. A large army of goliaths will suffer casualties for less damage taken than the same army of thors. Each goliath is easier to kill. A smaller enemy army can inflict meaningful, lasting damage that requires resources and production time to recover.

A big army of thors is strong only by relying on its weight. There is no finesse there, only the brute force of HP and DPS. And the enemy is going to need a big enough hammer in order to actually do any lasting damage, and kill even a single thor. You need to do at least 400 damage to even kill one. Whereas if that army was composed of 125 HP goliaths, then a much smaller force becomes much more effective, and can chip away at the goliath army much more easily.

Blizzard seems to like armies clashing, and dislikes positional play as it seems "boring." However the very idea of having a "big battle" presupposes an intent to incentivize deathball play. They focus on units' counter relationships and how compositions interact rather than on how fairly bland units fight one another in uneven distribution.

Suppose for the sake of argument that there is only one type of unit in the entire game. Obviously this unit's properties will have a massive impact on how the game plays.

Zerglings? Mad aggressive attacks to try and end the game immediately. Lots of zergling-on-zergling battles in the middle of the map. It plays fast, and perhaps is exciting, but ultimately there's not a lot more to be said.

Marines are much the same story. Whoever has more in one particular location will win locally. And marines have much stronger accelerating returns than zerglings due to their ranged attack, high dps, and low HP. The game will snowball quickly in favor of the player with more surviving marines early. The long and the short of it is that direct combat units, or "A-move units" are really quite boring if they're all you've got. They kill stuff provided you brought enough hardware, and they lose if you didn't.

Positional units don't follow this rule. An objectively weaker force can defeat a stronger force when well positioned. Unfortunately this dynamic is basically gone from SC2 (except the neutered tank). But in BW there were lurkers, dark swarm, reavers, arbiters, and other ways to get a local advantage for cheap. Put a few lurkers under a dark swarm, and they can bring the entire Russian Army up that ramp and it's not going to matter.

So if we do our one-unit thought experiment with a unit like a Lurker (assuming detection is available somehow) then we actually have a dynamic game. Where players put their lurkers matters. Covering one ramp might leave another ramp exposed. Sending a big force of lurkers forward into just a few lurkers already in position can be a big mistake (not so for marines).

Varying degrees of local advantage, the presence of strong positional units like lurkers, and the presence of force multiplier abilities and effects like dark swarm all act to make big armies acting in concert less effective. They can actually get stalled, or even defeated, by smaller armies used excellently.

Furthermore, micro depends on a player's ability to control their units. The most obvious way to encourage micro is to simply give a player more units to control. Positioning becomes more relevant when there are more pieces on the board, as the main consideration is pieces' positions relative to each other.

Increasing the siege tank's supply cost to 3 is the biggest single nerf to the siege tank that Blizzard has made. A maxed army of siege tanks is now 33% smaller than a Brood War equivalent. Its damage was nerfed considerably, but if the tank were still 2 supply it might be feasible to use the tank in numbers.

This is especially important for a positional unit like the siege tank. At 3 supply you need every single tank you can possible build to be involved in a single fight simultaneously in order to even be effective. If you have a strong economy of about 80 workers, that leaves you only enough room for (at most) 40 tanks if you build literally nothing but scv's and tanks. If tanks were 2 supply, with the same worker count you would max out at 60 tanks. That is a tremendous difference in absolute army strength, resource cost notwithstanding.

With 60 tanks you can put 60 guns in 60 different places, rather than just 40. Mechanically, sieging and unsieging more tanks and controlling their positioning well is a greater challenge. Building that many tanks requires more economy, more production facilities, more time, and better macro, but yields a stronger army as a product of that additional macro skill. And the army is larger and more difficult to precisely manage, and much stronger when correctly micromanaged.