On Uncertainty In Games



In Les jeux et les hommes, Roger Caillois asserts that uncertainty in the outcome of a game is what makes it enjoyable for its players. Consider a game of Hide and Seek: if the game played out the same way every time, the players would eventually stop playing it because the outcome would no longer be unexpected, and as a result, the enjoyment would eventually (and rather drastically) fall off. Greg Costikyan, in his book Uncertainty in Games emphasizes that while Caillois’ assertion is true, it is limited in scope. Uncertainty, according to Costikyan, is a necessary condition but should not be limited only to the outcome. His view is reasonable- consider the popular game Temple Run; the outcome is always that the player will fall or run into something. However, players still enjoy the game because of the uncertainty of the scores they might get. 

Enjoyment is not a requirement of game systems in and of themselves, but is a necessary condition nevertheless for voluntary player engagement. Why this is so could be better explained through psychology. The neurotransmitter dopamine is predominantly responsible for decision making and pursuing pleasure. The interesting thing is that dopamine is not as dependent on the reward, as on the anticipation and prediction of receiving a reward.  Our brains are set up to exploit conditions where more rewards (food for example) could be obtained. If one is able to successfully predict where the next meal will come from, they would have a better chance at survival. Thus, the brain aims to understand why it got that unexpected reward. That is why games like chess are so engaging; A victory is like finding food, and the player then works backwards to figure out what they did that made them win. However, since the number of ways of winning in chess is near infinite, there is always some new path to understand, and thus, a constant stream of dopamine. 

This psychological understanding of our brains presents an interesting consideration: uncertainty is enjoyable only when it influences outcome. By outcome, I do not mean the binary “Win” or “Lose” condition, but a broader, player defined state which she deems as being important based on her own inferences. This could be a crucial part of the story in Dungeons and Dragons or the condition where the player is suddenly confronted by an enemy in Counter Strike. Uncertainty will keep the players coming back to understand how things work in the game and what they can do to maximize their rewards. This brings us to the big question:


How To Design Uncertainty Into The Game System

Having established that uncertainty plays an important role in player engagement, the next question is how to best design it for use it in games.  The kind of games I am going to focus on fall in the strategy genre. While uncertainty is also important for less dynamic systems, game re-playability is better afforded by strategy games which are more concentrated on emergence than static games (which, in contrast are more story oriented). Re-playability necessitates uncertainty. 

Uncertainty within the game should be followed by a reward or punishment (also a reward but negative) of some sort. When this reward should be given depends on the designer. The designer should be able to anticipate what constitutes an important state in the game- a state that the players treat as crucial for their success or failure. Uncertainty without a reward should make no difference to the player- the uncertainty of rain in our Hide and Seek game will not change anything unless it introduces new hiding spots or strategies. That sort of uncertainty can add aesthetic flavor but its influence on player gameplay engagement will be minimal if at all present. 

The player must be able to understand the role of uncertainty in the outcome of the game. Without a clear understanding of why the outcome was the way it was means that uncertainty was replaced by randomness and this randomness turns the game into a play activity. While this may be enjoyable for younger players, randomness will only bring frustration to someone who wasted hours trying to understand a system only to realize that there is no order to it.

A good way to design uncertainty is to embed it into the game system itself. Games that support emergence necessarily support uncertainty because of the complexity of the system. 


In Conlusion

It is a good idea to think deeply about uncertainty. Uncertainty is intricately linked to player enjoyment, which is what we, as game designers, strive for.  We must ensure that this uncertainty is transparent and comprehensible by our players otherwise including uncertainty would be a moot point.