Several methods for outlier detection in Atomic Displacement Parameters – B value is applied to one particular macromolecular structure. It is demonstrated that outliers in B values give good indication errors in the atomic models. In this particular example it is demonstrated that a hypothesis that atomic B values distribution is shifted inverse gamma distribution. Removal of outliers from the set of B values improves the estimation of the parameters of the distribution. Local outliers in B values indicate errors in atomic models: validity of the assumption that neighbouring atoms must have similar B values has been verified. It is expected that local and global outlier detection program and modelling of B values as inverse gamma distribution will help to select the reliable atomic models. We suggest that such outlier detection and modelling should be part of model building and refinement of macromolecular structures using crystallographic diffraction data and single particle cryo electron microscopy maps.