Content area
Bayesian Conditional Probability (BayCoN) plots provide an empirical means to classify systematic errors in single-crystal diffraction experiments based on weighted residuals. Using a set of 314 structures from IUCrData, four recurring error types are identified and illustrated as follows: (i) incorrect standard uncertainties of observed intensities, (ii) intensity and significance cut-offs, (iii) cases where weak observed intensities exceed calculated ones, like with incomplete absorption correction, and (iv) cases where they are systematically smaller. Only eleven data sets showed uniform BayCoN plots. The analysis focuses on how the residual distribution reveals and categorizes systematic errors; the impact on model parameters and their uncertainties is not addressed in this work. The proposed classification is intended to assist crystallographers in improving experimental accuracy by recognizing characteristic residual patterns.
