The paper by Gneiting, Balabdaoui, and Raftery discusses probabilistic forecasts, calibration, and sharpness. It introduces a diagnostic approach to evaluate predictive performance based on the concept of maximizing sharpness while ensuring calibration. Calibration refers to the statistical consistency between forecasts and observations, while sharpness refers to the concentration of predictive distributions. The authors propose tools such as the probability integral transform (PIT) histogram, marginal calibration plots, sharpness diagrams, and proper scoring rules to assess calibration and sharpness. They illustrate these tools using probabilistic forecasts of wind speed at the Stateline wind energy center in the U.S. Pacific Northwest. The diagnostic approach is shown to provide a general, nonparametric alternative to information criteria for model diagnostics and selection. The paper emphasizes the importance of routine assessments of sharpness in evaluating predictive performance. It also discusses different modes of calibration, including probabilistic, exceedance, and marginal calibration, and provides examples of forecasters with varying degrees of calibration and sharpness. The authors conclude that calibration and sharpness are logically independent and may occur in any combination. The paper highlights the need for proper scoring rules and graphical displays to assess forecast performance.The paper by Gneiting, Balabdaoui, and Raftery discusses probabilistic forecasts, calibration, and sharpness. It introduces a diagnostic approach to evaluate predictive performance based on the concept of maximizing sharpness while ensuring calibration. Calibration refers to the statistical consistency between forecasts and observations, while sharpness refers to the concentration of predictive distributions. The authors propose tools such as the probability integral transform (PIT) histogram, marginal calibration plots, sharpness diagrams, and proper scoring rules to assess calibration and sharpness. They illustrate these tools using probabilistic forecasts of wind speed at the Stateline wind energy center in the U.S. Pacific Northwest. The diagnostic approach is shown to provide a general, nonparametric alternative to information criteria for model diagnostics and selection. The paper emphasizes the importance of routine assessments of sharpness in evaluating predictive performance. It also discusses different modes of calibration, including probabilistic, exceedance, and marginal calibration, and provides examples of forecasters with varying degrees of calibration and sharpness. The authors conclude that calibration and sharpness are logically independent and may occur in any combination. The paper highlights the need for proper scoring rules and graphical displays to assess forecast performance.