The debate around predictive power in theories is discussed with the example of inflationary cosmology, as some argue that if anything can happen, there is no predictive power, while others claim predicting probabilities is what theories do.