Summary: The basis of the statistical method of generation of branched and crosslinked structures as a Markov process is analyzed with special emphasis on the effect of differences in reactivities of functional groups. For irreversible reactions, it is important that the transition probabilities are calculated kinetically and that bonds once formed are not allowed to be reformed. The statistical methods are often a good approximation of real situation even if not all approximations used in the model are fulfilled. Cyclization is still a serious problem in modeling; For practical application it is recommended to rescale calculated ring-free dependences against the experimental gel point.