Picking when and how often a player will enter battles is a tricky balance between fun and engagement. Do it too often and they will get annoyed, too infrequently and they will get bored. Also, the battles need to be achievable and challenge the player at the same time. It is a complex paradox that, if planned wrong, can ruin your hard work. The best solution, when all is said and done, is to get your game play tested by as many people as possible and genuinely consider feedback, no matter how harsh.

Alongside random generation, of course, is predictive planning: if you have a story to tell, you also need to have a framework to replay that story over a period of time, balancing when you need a random event or picking up the next page in your book.

In this chapter, we covered random generation, what it means, and when to use it effectively; covered some very simple uses of randomness; built a battle scene and planned for expansion; covered basic AI concepts and State Engines; and enabled...