Bayesian prediction uses knowledge about the past to predict the future. For example, if a boy king started headlining more and more chapters in each novel, you might reasonably expect him to dominate even more in the coming books. If a towering knight had many chapters in the first three books, but her presence petered off in the fourth and fifth, you would not expect her to show up quite as much in the future. And of course, if a character died (or might die in a future novel), his or her odds of returning are quite low. This same method of making mathematical guesses helps your Roomba navigate your living room, and helps cognitive scientists understand human choice.