Financial institutions need a conceptually sound and easy-to-implement approach for analyzing model performance. Matthias Arnsdorf and Mante Zelvyte offer a new framework for counterparty risk model backtesting, based on Bayesian methods, and compare their approach with more conventional strategies.