“…To remedy this, several authors have suggested a randomized PIT. Specifically, if P is the predictive distribution, x ∼ P is the observed count, and v is standard uniform and independent of x , then where we define P −1 = 0, is standard uniform (Smith, 1985; Frühwirth‐Schnatter, 1996; Liesenfeld, Nolte, and Pohlmeier, 2006; Brockwell, 2007). For time series data one typically considers one‐step (or k ‐step) ahead predictions, based on a time series model fitted on past and current data, and checks for the independence of the randomized PIT, in addition to checks for uniformity.…”