A new statistical method for evaluation of hot channel and hot spot factors is presented. A new definition of “hot spot” is proposed with which the probability of exceeding critical temperatures can be correlated to the size of the zone in which they occur. In contrast to previous methods, the hot channel factors are demonstrated to be independent of the assumed spot size, provided that the uncertainties are correctly specified. Therefore, a new criterion is proposed for specification of the uncertainties which are random variables along the fuel pin axis, and the concept of a “specific standard deviation” is introduced. The different effects of the uncertainties, whether they act on single elements of the core, on groups of elements or on the whole core, are taken into account by an appropriate procedure. The statistical analysis takes into account the whole core with its particular axial and radial nominal temperature profiles. The principal results obtained by the SHØSPA code for the sodium-cooled fast reactor Na-2 are discussed.