Updating a model using the actual value of a new observation. Typically it refers to only an update in the level. You then obtain a sample of h-step-ahead forecasts based on originally estimated coefficients. Also called moving origin or rolling horizon. Armstrong (2001e) describes the use of successive updating.