Sunday, February 2, 2014

Balance Beyond Fairness

In which Max argues that the way many designers think of balancing stands in their way; balancing shouldn't be though of as "making your game fair." Balancing should be understood as "making sure your players' choices are interesting."

What Balancing Isn't
Balancing is usually understood as the process of tweaking numbers to make a game fair.  This often entails getting rid of 'dominant strategies' (strategies that, when chosen, make the player win more than their fair share of the time), and making sure no components are tremendously better than others.  Balancing also generally includes tweaks like making sure the first player doesn't have an advantage over the others.

This is a very limited view of the role of balancing for one major reason: fairness for fairness' sake is not necessarily desirable*.  It's easy to imagine a completely 'fair' game that is no fun at all, for example: "All players roll a die.  Whoever rolls highest wins."  This 'game' is completely fair, and consequently completely boring.  This concept can be extended to exception based components as well - if all the cards in Magic were the same (or similar in function), the player wouldn't have any meaningful choices and the game would suck!

What Is Balancing?
There have been a few pieces written in the community recently on strategies for balancing games, with a focus on achieving fairness.  While achieving fairness is a result of balance, balance is not about avoiding giving one player an edge over the others.  The balance that the term refers to is not balancing the players' likelihood to win, nor is it balancing the power of cards: it's about balancing the choices the players make.  At their core, strategy games are about tensions in choice: I want to do A or B, but I can't do both.  We balance these decisions to make sure that players always have one of these tensions on their turns - if all of these choices are imbalanced, then the tension vanishes.  Balancing is the act of tweaking numbers (resources, probabilities, options, etc.) to ensure that players always have at least two interesting and equally appealing choices during their turns.

As I mentioned, this understanding of balancing still achieves fairness.  For example, imagine players are drafting cards. If any of those cards is much better than any other, then the player has no meaningful choice.

How Is This Understanding Practically Different?
Take, for example, a Magic draft.  If the Magic R&D team had understood 'balance' to mean "making all strategies fair," then they would design each of the five colors of Magic in any given set to be roughly equal choices for your deck while drafting.  Understanding balance to mean "balancing player choices," however, could allow the Magic design team to make one color generally underpowered with a few really great individual cards.  Thinking about this design from a choice-balancing framework, this could easily result in fun choices: imagine a player opening their first pack to find a mediocre red card (but knowing that red is a strong color for draft in this set), and a really strong white card (but knowing that white is a relatively weak color for draft in this set).  This could definitely be a fun and meaningful choice for the player, and one that I would called balanced (assuming one option wasn't much better than the other), without forcing the strategies (colors) or individual components (cards) to be equal or "fair."

The goal with your balancing is to think through the choice points players encounter during the game, and to average 2-3 meaningful choices at each point.  One option (no choices) is occasionally alright, especially at the start of the game when more choices could be overwhelming.  More than three options can also be okay, especially in games with high replayability, or late in the game (basically, the more choices there are, the more players have to know what they are doing in order to not be overwhelmed).

The concept of meaningful choices (or not false choices) is a relevant one here.  Often times players will ostensibly have many choices, but only several will be meaningful.  Revisiting the Magic draft example, if I've committed to making a red green deck, a pack of cards from which I must choose might have 15 cards total, but usually only 6 or so of them will be in my colors (allowing me to ignore the false choices of the white, black, and blue cards).  Of those 6 maybe 2 to 4 will be decent choices, and after that it's up to me to make the decision.  Having these false choices is not bad design, as they actually help make the player experience easier while allowing for plenty of replayability.  However, designers must realize two things: first, the average 2-3 choices at each decision point is counted after omitting false choices, and second, false choices that look like meaningful choices can confuse new players.

Using This Framework
About a month back Seth Jaffee wrote an excellent article on concrete steps to follow in order to balance your game (using the traditional definition of 'balance').  While this was a pretty cool walkthrough, it's missing the single most important step to balancing - the step that follows from understanding balancing as ensuring tension in player choices.  As a designer, you must first identify your game's decision points and the potential decisions the players will be making at those points.  Only then can you know how to assign values and probabilities to guarantee 2-3 meaningful choices at each juncture.

As an example, I am working on a game with a "veto drafting" mechanic. Each turn, the player chooses one card from a row on the table, and asks the other players whether she can have that card.  Each other player has a chance to deny that player the card by paying her a flat number of victory points.  If they don't, she gets the card.  If they do, she gets to choose a different card.  "Which of the cards on the table do I ask my opponents if I can have?" is the biggest decision point in the game.  The player's choices at that point are:

  1. "Do I ask for my first choice card, assuming my opponents will give it to me?"
  2. "Do I ask for my second choice card hoping that an opponent will pay me to veto the choice, and then I can choose my first choice?  Even if they don't veto, I get something alright."
  3. "Do I ask for something I know someone else wants, but I do not, in order to force them to pay me VP, and then I get my first choice? If they don't veto, I get stuck with something crappy"
  4. "Do I play a card from my hand?" 
The goal is to have the player split between 2 or 3 of these choices every turn, with the occasional 1-choice turn or 4-choice turn.  Once a designer understands her players' decisions at their decision points, she can analyze the game to determine which game elements or numbers could make some of the possible decisions false choices.  In my game, here are some numbers in the game that can imbalance some of the choices above, reducing player agency:

  • Average victory points per card - each card drafted also provides the player victory points. The more victory points a single card can score, the easier it will be for an opponent to veto a choice, and consequently the less likely a player is to choose option 1 above.  It would be very risky.
  • Number of suits in the scoring scheme - the players are trying to collect cards with the same suit in order to score more points.  The more suits in the scoring scheme, the more each card will be worth to a player who is investing in gathering that suit, and the less it will be worth to another player who isn't.  The higher this difference between value for a player who is collecting those types of cards and the player who isn't, the less likely a player is to choose option 3 above, because getting stuck with a card of a suit you're not collecting is very bad.
  • Victory point cost to veto - the higher it costs for a player to veto an opponent's initial choice, the less likely a player is to veto, and consequently the less likely a player is to choose options 2 or 3 above, since in both of those cases the player has a high chance of getting left with something that isn't their first choice.
  • Number of cards in hand - since players can sometimes play cards directly from their hands, the more cards a player has in her hand, the more likely she is to have a defined strategy that she doesn't want to step outside of, so the less likely she is to choose option 3, above, and the more likely she is to choose option 4.
The core balance goal of this game is to make vetoing an attractive enough option that players must always take it into account, but not so attractive that players can always count on their first choice being vetoed.  With that in mind, I can now apply Jaffee's 5 steps for balancing game elements.

So next time you're ready to balance a game, remember: when balancing, you don't want your game elements to be "fair"** first and foremost.  Your primary goal should be to make sure the numbers in your game work to afford the players with an interesting choice every turn and beyond. 

*Fairness is not usually bad, but it can get in the way of more important design, and it can definitely hurt when misused.

**It's worth noting that not all forms of fairness will result from my framework of balance as ensuring interesting player choice.  For example, just making sure that players always have interesting choices doesn't necessarily mean that going first won't be a huge advantage.  This is where fairness for its own sake can be taken into account. 


Alex H said...

This was a very enjoyable read Max. Your veto mechanic presents an excellent point that viable options need to merely be available to create an interesting choice.

I like to think of object balance as less about equality and more about practicality in the environment of a game.

Ramenhotep said...

I totally agree, Alex! That's a great way to phrase it.

Larry said...

Hi Max, has your game (with the veto mechanic) been published yet? I'd like to reference how you implement the mechanic.