I agree, I am not sure why he keeps saying the same thing over and over...and this time he is talking about ensembles which are NOT forecasts vs. means. Does not "mean" they cannot be way off as to individual members but the mean maps are never forecasts. Using a mean map to assume results is like counting weights of 100 people walking into a building then "forecasting" that the next person will be a mean of them all. In many cases these are also members purposely fed false info to see how it would affect the end result. Their entire goal in life is to cross check the op's. If the mean is way off from the op's (upper air stuff, not snow maps) then that places the op's depict into question.