After some amount of playing and studying video games, I’ve concluded games need certain ingredients. I will try and capture these in the following paragraphs. What’s fascinating is how the umbrella of uncertainty encapsulates these elements.

What is a Game?

Before we dive in, let’s define “game” for the purposes of this document. There are many definitions as it turns out. Let’s assume that a game is a play activity based on a set of rules that result in winning and losing states. Playing (or spectating) the game presents opportunities to learn how to change states.

Games are often defined as “artificial” or “non-productive” but to me, this is far from the truth. Games are all about learning and these learnings can be applied to real problems. As humans, we make lots of decisions everyday. It’s what we do. And lo, even the most basic game contains the element of learning how to make decisions. This includes not playing the game itself!

A game is a series of interesting choices.

— Sid Meier (definitely misquoted)

Risk and Reward

Games need to present choices (do I attack? do I invest? etc.) and one of the best ways to make choices interesting is to introduce risk and reward. Masahiro Sakurai talks about this often on his YouTube channel. I recommend this video in particular. The core idea is the player chooses to risk valuable resources in the hopes of attaining more valuable resources.

It’s important to note that this is voluntary. The player has to choose to take the risk. If there’s no choice i.e. the player is forced to always take or ignore risks, this becomes less interesting.

Is this ingredient plain old gambling? Not quite, unless it’s just a game of pure luck. Games often test a player’s skills and strategic thinking so it’s wagering on one’s ability. This means the outcome must be uncertain when the game tests the player. If the player always wins, there is no risk. If the player always loses, there is no reward. Avoid both ends of the spectrum and fluctuate in the middle if possible. Good systems allow the player to decide the stakes and also grant rewards in proportion.

Predict and Punish

Whether fighting a human player or one driven by artificial intelligence, an essential element of gameplay is reading the opponent, sometimes called Yomi. David Sirlin talks about this in his book, Playing to Win; a relevant chapter is here. It goes over having rules that are complex enough that it is not easy to predict what an opponent will do. This depth provides uncertainty and with it, the thrills of anticipating your opponent. However, having different layers of Yomi or being able to read the opponent is not enough. The game must provide the opportunity to punish the opponent if they were correctly read.

Fighting games present many opportunities to predict and punish. But let’s consider a different genre. In most shooters, AI-controlled grunts will fire at the player on sight. This behavior is a given so predicting it is pointless. The designer might compensate for this by adding overwhelming numbers. But this needs other mechanics to allow for interesting choices. Consider a modification to this scenario. If the system delayed the behavior at random and implemented a tell before execution, Yomi is now possible. With this, it is now necessary to have a mechanism for punishment. This could be as trivial as a numeric bonus for the player. Or it could be as complicated as a separate input action that grants debuffs to all enemies. Without such punishment mechanics, it renders Yomi useless. If there’s no difference between shooting a grunt during its tell or at any other time, then there’s no meaningful decision-making to be had.

Prevent Nash Equilibrium

Nash Equilibrium is a Game Theory concept wherein a player does not need to deviate from a strategy. It’s like a game with a hundred moves but one works so well in all situations that there’s no reason to switch to another. While trying to get the best payoff for the least effort, many players will approach a game’s Nash Equilibrium. Regardless of what the game dishes out, sticking to a single strategy does the trick.

Experimenting and discovering effective strategies can be fun. But designers should strive to prevent comfortable unbeatable techniques. These create a game without uncertainty which can lead to boredom. With the example of a game with a hundred moves, one can prevent Nash Equilibrium by ensuring there are situations where some moves are significantly better than others. Adding advantages or disadvantages to each move can entice changing things up.

Interconnected Systems

Many seek to design game systems that interact with each other. It leads to emergence i.e. valid player actions that weren’t explicitly designed. Beyond emergence, interconnected systems can enable meaningful decisions.

Consider a flashlight mechanic in a game with dark environments. The flashlight utilizes the game’s lighting system and illuminates the level. As such, the player will always use the flashlight to navigate dark rooms without error. This design does not lend itself to strong decision making. One may as well save a user input button and automate the flashlight based on visibility. But, if the flashlight interacted with the AI system to alert enemies, we now have a better decision to make. The more systems the flashlight can interact with, the more interesting a tool it can become.

The above is a trivial illustration. The weather in Breath of the Wild is a complex example of a system that interacts well with others. For example, the rain affects surfaces (climbing) and extinguishable objects (cooking, combat). It forces the player to rethink how they play and make new decisions.

Systems that play well with each other can entice exploration and discovery. If taught a few basics, the player can intuit more techniques through experimentation.

Conclusion

Games are more fun when their outcomes aren’t predetermined. As such, uncertainty is important to game systems. This is not the same as instability or buggy behavior. When encountered, the game must communicate any variance as the predictable result of its machinations. The elements listed above use uncertainty to ensure some thinking is necessary. It could be risk evaluation, visualizing possibilities, formulating strategies or understanding complex systems.