A randomised design in which two event-types were randomly intermixed would have a first-order (M=1) transition matrix: A B A 0.5 0.5 B 0.5 0.5 In other words, there is a 50% probability of A or B, regardless of the previous event-type. (This transition matrix is rather redundant, but shown this way in order to compare with subsequent examples.) For this randomised design, the efficiency for bot