Haavelmo’s appeal for the probability approach is based on his firm conviction that probability modelling by means of classical statistics is the only systematic way to tackle ubiquitous uncertainty during empirical verification of economic theories using passively observed data from an open world environment. This chapter delves into the main aspects of the uncertainty he has identified in his Chapter 1 and finds that the primary aspect is in fact model uncertainty in econometric research. However, the task of handling model uncertainty goes well beyond the framework of classical statistics based on probability models.