The adjective randomized implies that the order in which squared differences between the central value and all measured values in the set are added does not impact the numerical value of the variance. In contrast, the adjective ordered implies that the measured values in a set are in the same sequence as they were in a sampling unit or in a sample space. If both terms are properly defined, the terms variance of a set and variance termss of an ordered set would be sufficient.

Both equations apply to sets of measured values with equal weights such as test results for core samples of the same length and density, bulk samples of identical mass, or central values of ore sections in equidistant boreholes. Core samples of variable length and density, bulk samples of variable mass or volume, and central values of ore sections in boreholes at variable distances, require that proper weighting factors be applied to all terms in the above equations. The equations required for this calculation can be found in Sampling in Mineral Processing (see Reviewed papers).

After geostatistics was hailed as a new science, the variance of a set became a sort of "sill value". Similarly, the analytical variance became a "nugget effect". While it is true that geostatistical neologisms reveal paucity, the penchant for neologisms and the use of the "krige" eponym as an adjective and a verb explains why so many geospeak terms are at variance with those defined by ISO Technical Committee 69-Applications of Statistical Methods.