The beautiful dataset

Natural experiments, high stakes, expert subjects, knowledge of the precise objectives and exact rules, clean measurement, observability of strategies, incentives, actions and consequences, large datasets, exogenous rule changes, distinct social effects, and no Hawthorne effects. These and other desirable attributes for empirical work are found in sports settings. In sports, features that often characterize either the lab or the field are found simultaneously. Sports can offer the best of both worlds. Reluctance to recognize these advantages reflects a misunderstanding of the virtues of sports data, and this reluctance has discouraged the study of these settings and slowed down the production of knowledge. This survey reviews literature that shows how sports settings have made it possible to implement the first successful test, or the best test to date, of various models and hypotheses, and to discover new phenomena. It is not a question of what economics can do for sports, but what sports can do for economics.