Counterplay is King in PvP Balance

I've recently been taken in by Dungeonborne, a fantasy loot extraction game where you pick a combat class, then try to escape an area populated with hostile creatures and other players. If you die during a match, you lose any gear you entered with, plus any loot you picked up along the way. With these stakes, and PvP as a key aspect of the game, class balance is a top concern of players. Examining this game has kicked off a lot of thoughts about how to balance asymmetric PvP, mistakes I often see, and lessons I've learned over the years.

 

Metrics Can Be a Trap

Let's start with a contrived example. We've made the game Hardcore Craft Battles. It's a 1v1 PvP brawler where you queue up as one of three classes: Rock, Paper, or Scissors. You match against another player and fight to the death. Let's say every class has equal representation, so you have an equal chance to match against any other class. In our game, Rock always wins against Scissors, which always wins against Paper, which always wins against Rock. Mirror matches (like Rock vs Rock) should definitionally have a 50% win rate for that class (this isn't Brawl, after all). 

If we gather statistics about matches and look at the overall win rates of each class, we'll see they all sit at an even 50%. Seems perfect! We want matches to be even. Clearly though, the player experience will be horrible. As a lifelong Rock main, I'll win roughly 50% of my matches against other Rocks, 100% of my Scissor matches, and get stomped in all my Paper fights. Aside from potentially fun Rock encounters, most of my matches are decided the second I see my opponent. In other words, they're totally uninteresting.

Despite being a contrived example, we can imagine a game that feels similar, where certain classes or compositions hard counter others. If you have circular counters, the game can look balanced on paper, but feel incredibly unbalanced to actually play. Player perception and experience is not only important, it is often one of the most important aspects of game balance.

 

Counterplay

The key aspect missing from Hardcore Craft Battles is counterplay. Let's clarify. Counterplay is the capacity to respond to an action your opponent takes in a way that mitigates or counters that action. To go deeper, I think there are two key types of counterplay: gameplay counters and strategic counters. 

A gameplay counter is an action you can take during a fight. These include the ability to parry an attack in a fighting game or build advantageous units in an RTS. Strategic counters are actions you take before the fight. Examples are adding cards in a deckbuilder to deal with common strategies you'll encounter, or equipping elemental resist gear before you fight a mage. Strategic counters are specifically decisions made that cannot be changed once a fight has begun.

A twist on the original problem of Hardcore Craft Battles is that soft class counters are a type of strategic counterplay. This is important in games like Dota 2, where character picks and counterpicks is a science unto itself, but it's critical that these counterpicks provide an advantage, not the win.

Cursed Balance Problems

A mistake I made early in my career was believing there was always a set of numbers I could apply to a game where everything would be "balanced" -- that some combination of damage values, cooldown durations, health amounts, etc exists where every class and strategy will be even in power, and it was my job to hone in on those numbers. In case it needs to be said, this belief is deeply flawed.

There are two ideas worth examining here: the idea of cursed design problems, followed by Goodhart's Law. A cursed design problem is one in which you're trying to accomplish multiple incompatible goals. Here, it is the idea of having both asymmetric PvP and a perfectly balanced game (when every class or strategy has an equal win rate). These goals are at odds. You must either give up asymmetry, or surrender the idea that you should chase perfect balance. If you do not, you will instead surrender to Goodhart's Law, where the metric has become the goal. The most efficient way to balance two asymmetric things is to make them similar, inevitably leading to the over-homogenization of gameplay. In other words, you can make your game symmetrical from the start, or by chasing perfect balance you will slowly make it symmetrical, giving up what makes asymmetry interesting to play.

 

What Do We Do?

When changing the numbers alone isn't good enough, it's time to change the design. This might be the design of a particular mechanic, item, class, or system, but when the design must change it's a good time to look for areas of counterplay.

Here are a handful of approaches I've used in the past:

1. Write out advantages and disadvantages a particular piece of gameplay should have. A classic example might be ensuring a ranged class is strong at a distance but vulnerable up close. How you accomplish this is highly dependent on your game, but laying out intentions is a great way to help focus your efforts. If an element of the game is too strong, emphasize its intended weaknesses, or provide more ways to mitigate its strongest aspects. If too weak, bolster its strengths, address counters that are too potent, or buff the areas it should excel in. Establishing where the lanes are can help avoid overly chaotic changes.
   
   
2. A variation of the first approach is adjusting windows of opportunity: where weaknesses are exaggerated, strengths are diminished, or even when the opponent's counterplay options are limited. These include moments when abilities are on cooldown, resources have been denied, or after an opponent whiffs their heavy attack.
   
   
3. A great holistic approach can be providing consistent defensive strategies, while having a wide variety of offensive capabilities. This is covered extremely well in this post about Guilty Gear. By providing generic, flexible defensive options, you can be sure that counterplay is consistent across your asymmetric designs. This allows more focus on the offensive aspects and ensures they can be properly countered by skillful use of the defensive suite. Note that these defensive options can also be asymmetric. In Starcraft, every race has the ability to reveal stealthed units, but execution for each is varied.
   
   
4. Avoid problems learned by other games. Monitor the patch notes of competitive games over time and analyze how they deal with particular issues and the community response to them, both immediately after the change and after giving it time to settle. Build up a library of problematic gameplay patterns to be aware of. Some examples:

• Easy access to hard CC and burst (a character that can both stun on demand and deal fast, heavy damage before the stun is finished). Areas of counterplay can be awareness of when the CC is coming, allowing the CC to be blocked, dodged, etc, or ways to escape the CC.
  
  
• Silent power. This may warrant a post on its own some day, but I define silent power as anything that adds strength in a way that isn't obvious. A common example is a character quietly granting bonus stats to allies around them or having a power spike that isn't clearly communicated to the opponent. If a player doesn't know about it, they can't be expected to react appropriately. Typically these are solved with more obvious audiovisual cues, but may require a deeper look at when and how the power is applied.
  
  
• Any situation where the best counter to a strategy is the same strategy. This creates a balance ouroboros where the moment that strategy has any dominance, it will threaten to overwhelm the game. If the primary response to camping in a shooter is to never peek walls, you will quickly have a game where nobody wants to move.
  

The most important lesson in asymmetric balance is this: you are succeeding when the game feels balanced. Metrics are important to verify where to focus, but as a player, I want to know that there was something I could have done to mitigate a loss. It is important not only for counterplay to exist, it should be discoverable, if not obvious. If the proper reaction is too niche or difficult to employ, it may not be serving its purpose well. Seeking out areas to add clarity and counterplay in an imbalanced game is a great starting place for healthy PvP.