Anti-Snowball Design

Uncontrollable snowballing in RTS games is not fun. Nobody enjoys being set back by a small margin and being destined to gradually lose. The mechanics of RTS games, such as resource acquisition and army amassing, naturally lead to snowballing.

However, successful RTS franchises tend to put checks on the amount players can snowball. Instead, excessive snowballing is a problem that can be found in poorly designed RTS. This essay will discuss how various RTS franchises handle anti-snowballing to help keep the game fun for all players. The intention isn’t to negate the benefits of winning, but to reduce it to a more manageable level.

Some anti-snowball methods are intrinsic to a game’s design and balance, while others are engineered. This distinction can be referred to as anti-snowball mechanics compared to anti-snowball features. These are sometimes referred to as “comeback mechanics” but I think that’s not quite right. The inherent anti-snowball mechanics mostly stem from diminishing returns on the effectiveness of advantages such as army sizes and economy. Comeback features include army upkeep and retreat.

Non-linear power curves of army size advantages reduce the snowball effect. This is often to do with the level of attention required to utilize units. Two Sherman Tanks are not twice as effective as one because tanks require so much precision of movement to avoid being caught by enemy anti-tank guns. While two is obviously better than one, it is with diminishing returns. Likewise, utilizing heroes and spell casters in WarCraft 3 requires so much actions-per-minute.

This effect is more subtle in StarCraft 2. Stalkers have a large model size relative to their attack range, lowering the damage output of a blob of Stalkers unless they concave. Terran mirrors are not decided by which Marine blob is larger, but by which player can get their Siege Tanks in the best position. Diminishing returns of army size can be caused by attention needed to control units, inherit penalties of a unit being massed or having strong counters. 

Good implementation of economic progression can create vulnerabilities in the winning player. As a StarCraft player gains more bases over their opponent, they will have more territory to defend. They will be vulnerable to harassment as their forces will either be spread thin or have gaps in defences. The maximum base saturation of Minerals and Vespene Gas is crucial because it forces players to expand onto the map.

Command & Conquer 3 has no hard saturation limits per base, but over-saturating Harvesters causes diminishing returns as the Tiberium Fields rapidly deplete. Regardless of how its implemented, economic progression should be vulnerable to the enemy and capable of being reset. This can take the form of weak worker units or territory points that can be captured.

I’m not a fan of how Dawn of War 1 does its economy; Listening Posts are tough and well-armed while Power Generators are safe in the base, and both resources have permanent, global upgrades. It’s not surprising that Dawn of War 1 can suffer from the snowball effect. Command & Conquer: Generals has the same problem in the late game with resource buildings like Black Markets being built safely in the main base. Ashes of the Singularity suffers from its region-specific upgrades like Amplifiers and Orbital Abilities that are impossible to destroy or remove. The losing player should have the potential to strike back a deadly blow at their opponent.

Population limits also reduce snowballing. A maximum population limit puts a hard cap on how big a player’s army can grow, forcing them to attack before they lose the advantage. The defending player will typically benefit in positioning and trade better, often counteracting the army size disparity. The slow build speeds of StarCraft necessitate building excess production structures to rapidly replace lost units, reducing the advantage of having a large resource bank.

Population buildings like Supply Depots or houses also help anti-snowball. The loser of a fight can immediately replace their lost army, but the victor has to construct additional Pylons before further growing their forces (or workers). Seemingly unrelated mechanics like selling structures can also have a role in anti-snowball. If you’re about to lose a forward base in Command & Conquer, you can sell off structures to receive a 50% refund and lessen the setback. As one player gains the majority of map control in Company of Heroes, their opponent will have shorter distances to reinforce from the base and return to the frontlines. 

Anti-snowball mechanics are inherent because they’re just a part of the game and probably weren’t designed with anti-snowballing in mind. However, some features are designed to explicitly address snowballing. Most notable is upkeep in games like Company of Heroes where players gain less income as their forces grow. Upkeep passively helps the players with smaller armies catch up. The retreat mechanic means players can lose battles without having their forces annihilated.

As a result, battle outcomes primarily determine territory shifts, and the loser can quickly regroup and counter-attack. Company of Heroes has many strong anti-snowball mechanics and features which are necessary to balance out its pro-snowball mechanics. Units gain veterancy from combat, acquiring bonuses to stats or unlocking new abilities. In essence, veterancy gives further bonuses to the player who is already winning. That may sound problematic, but it’s worth it because veterancy is such a gratifying feature. Tanks in Company of Heroes also tend to snowball players because regular infantry units have no anti-tank damage. Snowball mechanics are not a bad thing because they can be really fun to benefit from. However, the more snowball mechanics an RTS has, the more it needs to be offset by anti-snowball mechanics or features.

Anti-snowball features should be subtle so that the winning player does not feel they are being punished. Players in Company of Heroes are not thinking about upkeep drain; it just happens passively in the background. Compare this to WarCraft 3 where upkeep hits sudden, sharp thresholds. WarCraft 3 enthusiasts may be upset with me here, but I hate this system because it feels punishing.

On the one hand, the thresholds add strategic depth because you may choose to maximise income by staying below the threshold. On the other hand, it discourages you from the fun act of growing your army, and its sudden alerts can create a fear of missing out. Well designed anti-snowball features might punish the player for having an advantage, but it shouldn’t communicate it so blatantly.

Mechanics creating a defender’s advantage prevents tiny gains in the early game from suddenly being game ending. All RTS have some inherent defender’s advantage due to reinforce timings, but often more are required. A rush of 5 Marines against a defending player with 3 Marines will not end a game of StarCraft. The defender has the high ground advantage in their base, and they could pull workers from Mineral lines to engage.

As the game draws out, the 2 Marine advantage becomes negligible when comparing 52 to 50 Marines. Command & Conquer 3 has a different defender’s advantage with repair drones on War Factories. Defender’s advantage should scale off past the early game so that it doesn’t prevent one player from ending the game when they have clearly won. The speed of Command & Conquer 3’s repair drones and the damage of SCV’s becomes insignificant in mid-game fights, while StarCraft 2’s high ground advantage is negated by having vision.

Different victory conditions present comeback opportunities. Being able to win in Company of Heroes through Victory Points means the toughest army does not always come out on top. Being able to take, hold and stall regions on the map tends to require a different army composition than one designed for driving into a base and blowing everything up. It also means one player might have an advantage in army size, but still has to remain vigilant because they have a disadvantage in Victory Points.

The elegance of Victory Points in Company of Heroes is how territory capture interacts with mechanics such as suppression and smoke shells. Command & Conquer: Rivals has even more comeback potential in its missile victory condition. The losing player need only steal the launch pads at a crucial moment to redirect its launch destination.

RTS gameplay of gaining resources and growing armies inherently contain snowballing. As a result, RTS games need to implement anti-snowball mechanics and features, whether they be a natural consequence of its design or deliberately engineered. Successful RTS players should gain advantages, but the effects should be minimized to prevent snowballing out of control. Many RTS games have diminishing returns of larger armies due to the required precision of control.

Army size scaling is also limited by stats such as attack range and having strong counters. Economic progression should necessitate expanding out onto the map, creating vulnerabilities to harass. The methods of gaining resources should be at risk of being destroyed or reset, such as weak worker units rather than global upgrades. Population limits and other various features like selling buildings can all contribute to slowing down the rate of snowballing.

Engineered anti-snowball features like upkeep may be necessary if a game contains pro-snowballing mechanics such as veterancy. Good anti-snowball design should be subtle and not make players consciously feel punished for winning. Creating some kind of defender’s advantage prevents early leads from being game-ending, but should scale off in the mid-game. It’s not fun when small errors gradually and inevitably cause a loss.