Regression diagnostics meets forecast evaluation: conditional calibration, reliability diagrams, and coefficient of determination

A common principle in model diagnostics and forecast evaluation is that fitted or predicted distributions ought to be reliable, ideally in the sense of auto-calibration, where the outcome is a random draw from the posited distribution. For binary responses, auto-calibration is the universal concept of reliability. For real-valued outcomes, a general theory of calibration has been elusive, despite a recent surge of interest in distributional regression and machine learning. We develop a framework rooted in probability theory, which gives rise to hierarchies of calibration, and applies to both predictive distributions and stand-alone point forecasts. In a nutshell, a prediction is conditionally T-calibrated if it can be taken at face value in terms of an identifiable functional T. We introduce population versions of T-reliability diagrams and revisit a score decomposition into measures of miscalibration, discrimination, and uncertainty. In empirical settings, stable and efficient estimators of T-reliability diagrams and score components arise via nonparametric isotonic regression and the pool-adjacent-violators algorithm. For in-sample model diagnostics, we propose a universal coefficient of determination that nests and reinterprets the classical R2 in least squares regression and its natural analog R1 in quantile regression, yet applies to T-regression in general.