In Chapter 2, we looked at ways in which a variable can be estimated based on an independent variable. The dependent variables we were estimating were continuous (sales of ice cream, weight of baby). However, in most cases, we need to be forecasting or predicting for discrete variables—for example, whether a customer will churn or not, or whether a match will be won or not. These are the events that do not have a lot of distinct values. They have only a 1 or a 0—whether an event has happened or not.