Abstract Using profile functions to predict subsidence induced by mining activities has two major advantages: quick calculations and easy writing of computer programs capable of undertaking such calculations. The possibility of quickly obtaining many subsidence profiles, by slightly changing the initial data, allows for a statistical approach, which does not consider the value of the strain at a given point but the likelihood of finding a certain strain at such point. The study of subsidence effects is more realistic with this approach and also eliminates the errors that could arise either from the estimation of the initial data or from the calculation of the strains starting from a given subsidence profile. In this paper, the authors present a work methodology applied to the case of a profile function developed for steep coal seams.