Title: Frequent or Systematic Changes? discussion on "Detecting possibly frequent change-points: Wild Binary Segmentation 2 and steepest-drop model selection."

Abstract: We discuss Fryzlewicz's (2020) that proposes WBS2.SDLL approach to detect possibly frequent changes in mean of a series. Our focus is on the potential issues related to the model misspecification. We present some numerical examples such as the self-exciting threshold autoregression and the unit root process, that can be confused as a frequent change-points model.