On The Problem of The Suitability of Forecasting Formulas Found by Statistical Methods