Acceptable from the predictor's point of view. We need to know what he/she
believes the windows on the prediction to be for the prediction to be of any use
and for it to be evaluated. The only standards we can place on a prediction
from this end must be statistical in nature, and we can't do that with a
individual prediction.