In this note we explore the following surprising fact: In regression with trend and seasonality, the prediction risk is constant for all seasons of a new cycle, despite the fact that it increases with time when the seasons ...