Uncertainties are inherent in all physical systems. There are uncertainties due to variations in material properties, manufacturing, assembly, and also in the way systems are deployed and operated by users. With the advent of modern wireless technologies and the Internet of Things we are confronted with systems which change their characteristics in a dynamic way according to user preferences. Traditional EMC models are deterministic reflecting static situations and assuming complete knowledge of all system parameters. If we assume that some system parameters are random, then it follows that responses will also be random which can only be described in a statistical sense. Typically, responses have a distribution which is characterized by its statistical moments (mean value, standard deviation, etc). In a previous issue of the Magazine (Vol. 6, Part 3, 2017) we have discussed some of these matters in connection with randomness in transmission lines and interconnects. The traditional way to deal with randomness is to operate a model (or do testing) many times with parameters spanning the range of the random variables and thus obtain the distribution of likely responses (Monte Carlo technique). This is very time consuming as it requires tens of thousands of tests (model implementations) to be performed.