Principlev1
Decompose prediction accuracy into calibration (do your
Decompose prediction accuracy into calibration (do your probabilities match reality) and resolution (can you distinguish probable from improbable) to diagnose different types of forecasting errors.
Why This Is a Principle
This principle derives from Systematic Overconfidence Taxonomy (overconfidence has distinct forms) and provides a specific analytical framework (Brier score decomposition) for understanding where calibration breaks down. It's prescriptive about how to analyze your prediction data, not just that you should analyze it.