I have a workfile with a range of 1999-2024 , inside it I have series with different starting and ending dates.I am running an ls regression equation for a smaller sample smpl “2015 2015+1000” days , while this estimates everything in that specific sample after that I run a forecast for the out of sample part,with smpl “2015+1000 2015+2000” but I see that the predicted Ys are forecast from the start of the observation of one of Xs which is 2008 and it does so for 1000 days, so now I have predicted Ys that are mismatched in dates. Why doesnt forecast only forecast what I tell it to ? In the books it even uses a smpl command before each forecast to do it for a specific date range.