Hey. Looking at EuStockMarkets[1:1860] -- the DAX. ar() estimates an AR(1) model with \alpha_1 ~= 1 and high \sigma^2. But forecasting from a given X_t_0 according to an AR(1) model with \alpha_1 = 1 should be X_t_0 for all X_t, t > t_0 ..... ? How do you do this forecasting in R? predict(ar(),n.ahead)$pred gives something different .. Thanks! Lars [[alternative HTML version deleted]]