To drill or not to drill? It sounds simply enough – but in reality a wildcatter has more choices than just whether to drill or not. They typically have the option to conduct one of several tests intended to provide imperfect information on the likelihood of finding oil and how much. When a decision (Drill?) is preceded by a chance node (Test) that provides imperfect information, we say that the model contains learning. Test drilling for oil is a classic example in which a learning model can be employed to find the real value of an asset.