The statistical properties of a cellular automaton can be predicted by drawing a ``mean field curve'' based on randomly distributing the cells throughout a configuration. Using the rules for compound probabilities in the evolutionary function predicts the distribution of the second generation, from which properties of the automaton, such as fixed points, can be inferred.
This approach must be corrected for the correlations which build up from generation to generation, creating a need for calculations which reveal the periodic configurations exactly; but then it is just as easy to include shifts along with the periodicity.