AbstractComputer simulation of cooperative, many particle systems (eg. real gas, Ising lattice) encounters a difficulty inherent to the ensemble theory. Under given conditions, the typical configurations constitute a vanishingly small fraction of all possible particle arrangements. Moreover, near to certain phase transitions, the typical configurations abruptly acquire pronounced long range order. Hence a computer method (like Metropolis Monte‐Carlo), which is not explicitly oriented towards discovery of such extremely rare arrangements, will almost certainly become inadequate. An attempt is made to formulate quantitatively these elementary observations, with the view of deriving a unified search for methods, that will enable to simulate the emergence of long range order in random system