This chapter examines how to predict stock market crashes and other large market events as well as the limitations of forecasting, in particular in terms of the horizon of visibility and expected precision. Several case studies are presented in detail, with a careful count of successes and failures. After providing an overview of the nature of predictions, the chapter explains how to develop and interpret statistical tests of log-periodicity. It then considers the concept of an “antibubble,” using as an example the Japanese collapse from the beginning of 1990 to the present. It also describes the first guidelines for prediction, a hierarchy of prediction schemes that includes the simple power law, and the statistical significance of the forward predictions.