What Is Game Balance? Keeping Fairness and Meaningful Choice
From komi in Go, first-move advantage in chess, and StarCraft's three races to rock-paper-scissors design and buffs and nerfs, game balance explained.
Whenever patch notes go up, the community always gets loud. "That character finally got nerfed." "Why did only the weapon I use get weaker?" "This patch ruined the game." Game balance is one of the topics players react to most sensitively, and for developers it's homework that never ends.
But what exactly does "good balance" mean? Is balance good if every character has the same stats? In this post, we go through the core concepts of game balance (fairness, meaningful choice, dominant strategies, asymmetric design, and feedback loops) one by one, and use examples ranging from Go and chess to StarCraft to explain how developers balance their games.
The Two Goals of Balance
To sum up game balance in one sentence, it's "a state that is fair while offering several choices worth picking." The two goals are connected, but they're not the same thing.
- Fairness: The reasons players win or lose should mainly be skill and judgment. Starting conditions or luck should not decide the outcome in advance.
- Variety of choice: There should be multiple viable strategies, characters, and weapons. If there's only one right answer, the game becomes a task of executing that answer.
There's a common misconception here. It might seem that making every element equally powerful would produce perfect balance, but in reality it doesn't. If everything is equally strong, it makes no difference what you choose, so the choice itself loses meaning. Good balance aims not for "equal power" but for "different but similar value."
Meaningful Choices and Dominant Strategies
The most important concept in discussions of balance is "meaningful choice." If none of the options has any effect, or conversely one is always the best, the choice is meaningless. It's like choosing a die face when 6 is always best: there's no reason to think about it.
A strategy that always produces the best result like this is called a "dominant strategy." A dominant strategy is the most harmful form of meaningless choice for a game. If one character or one build is too strong, people use only that, and the dozens of other options effectively disappear. A large part of balance work is finding and eliminating such dominant strategies.
A typical way to prevent dominant strategies is intransitivity, or simply put, a rock-paper-scissors structure. If A beats B and B beats C, you'd normally expect A to beat C too, but if you make C beat A as in rock-paper-scissors, no absolute strongest option emerges. This is why many strategy games build these counter relationships between unit types. However, rather than simply declaring "C beats A," it's easier for players to accept when counters arise naturally from differences in attributes such as range, speed, and cost.
Symmetry and Asymmetry: Go, Chess, Janggi, and StarCraft
Giving every player the same conditions is the simplest way to balance. A fully symmetrical game like Pong is fair in itself. But games where players take turns can't be fully symmetrical, because the side that moves first is inevitably at an advantage or disadvantage.
In chess, White moves first. Statistics accumulated since 1851 show that White consistently wins slightly more often, with a winning percentage generally between 52% and 56%.
Go solved this problem with "komi." Since Black, who moves first, has the advantage, White is given extra points. In Japan, komi was once around 2.5 points before rising to 4.5 and then 5.5, and the Nihon Ki-in raised komi to 6.5 points in 2002, citing that Black's winning rate under the previous rule was 51.855%. Today the standard komi is 6.5 points under Korean and Japanese rules and 7.5 points under Chinese rules. Komi is set in half-point units to reduce draws. Given that even a centuries-old game has revised its numbers based on statistics, Go's komi deserves to be called the oldest "balance patch."
Janggi (Korean chess) has an interesting mechanism too. In janggi, the positions of the horse and elephant can be swapped, so the side that sets up later can respond after seeing the opponent's setup. That's why in tournaments Han places its pieces first, Cho places them next, and Cho makes the first move. In effect, the setup order and the first move are split between the two sides.
The representative example of asymmetric design is StarCraft (1998). The three races, Terran, Zerg, and Protoss, differ in units, how they build structures, and how they are played. Protoss pursues quality with expensive, powerful units; Zerg pursues quantity with cheap, fast units; and Terran pursues flexibility in between. Even so, the goal was to match overall strength so that no race is inherently advantaged. Asymmetric games are far harder to balance than symmetric ones, but they produce that much more variety in matches and strategies.
Mechanisms Against Snowballing: Feedback Loops
Balance isn't just about starting conditions. How the gap changes as the game progresses also matters.
- Positive feedback: A structure where the leading side grows stronger. Taking territory increases resources, and more resources let you take more territory. If it's too strong, the match is decided in the first few minutes and the rest is just confirmation.
- Negative feedback: A structure where the leading side bears a burden. Examples include upkeep in real-time strategy games, where more units means more tax, or soccer, where the deeper you push into the opponent's half, the more defenders there are. If it's too strong, draws become frequent and matches drag on.
Most games mix the two. The level structure of RPGs is a good example. As you grow stronger, you defeat stronger enemies and earn experience faster, but the experience needed for the next level also increases, so growth stays roughly steady instead of exploding. In strategy games, newly conquered territory having lower production than your home base or a risk of rebellion is also a mechanism that slows down a leading player's snowball.
Buffs and Nerfs: The Hands That Fix Balance
The most common way to adjust balance after launch is with "buffs" and "nerfs." A buff makes an element stronger, and a nerf makes it weaker. The term "nerf" first took hold in the MMORPG Ultima Online, borrowed from the Nerf toy brand that shoots soft foam darts.
There are a few principles often used when making adjustments.
| Principle | Description |
|---|---|
| Bring down the top | Overpowered elements make others useless, so fix them first |
| Pull up the bottom | Raise elements nobody uses until they become viable |
| Buffs over nerfs | Players generally react better to buffs than to nerfs |
| Indirect adjustment | Adjust by changing opposing elements or surrounding rules rather than touching the target directly |
| Bans as a last resort | Avoid banning specific elements in tournaments whenever possible |
Nerfs require caution because players' time and emotions are at stake. When an item that took a long time to obtain suddenly becomes weaker, that effort can feel wasted. In online games where items are bought and sold, a single nerf can even shake the in-game market prices significantly.
Perfect Balance Isn't the Goal
Interestingly, balance researchers say a little imbalance can actually be good. When a strategy becomes popular, players find ways to beat it, and that counter becomes popular in turn, so the "meta" moves on its own. In a perfectly balanced game, only proven strategies may get repeated, and only top players may be able to create new ones.
That's why many developers let the meta run on its own and intervene only when a single strategy dominates for too long. Another important thing is keeping appearance and numbers consistent. If a giant dragon is weaker than a small goblin, players will feel something is off even if the numbers are balanced.
Balance Work at a Small Studio
Small teams have a hard time collecting vast amounts of data the way big games do. The methods we use are simple.
- Start with a spreadsheet. Put the cost and effect of every unit and weapon in one table, and first look for anything whose cost-effectiveness stands out.
- Test the extremes. Deliberately play using only one thing over and over; if you can get all the way to the end with it, there's a dominant strategy.
- Change one thing at a time. If you change several values at once, you can't tell what made the difference.
- Watch out for situationally strong elements. If something is only useful in specific situations but costs the same as everything else, it's especially hard to balance.
Action Janggi is built on the piece relationships of traditional janggi, so even when we add new control elements, we keep checking that the value balance between the original pieces doesn't break down. Emergency tools like the nuke in GRID DEFENSE: GRID WAR must keep their place as "a tool that turns the tide once in a crisis," so that the fun of thinking about which turret to place where in normal play doesn't disappear.
Closing Thoughts
Balance looks like a matter of tuning numbers, but in reality it's about protecting choices. A well-balanced game is one where players can think, "Should I try something different this time?" and, after losing, say, "Next time, a different strategy."
