The thermal cracking of heavy oils has been modelled on the basis of combinations of first-order reactions in order to describe the behaviour of pseudocomponents with time and temperature of reaction. An alternative formulation uses the concept of a time-varying rate constant of the form[Formula: see text]where β is the rate constant and γ is an exponent that has values of 0 < γ ≤ 1.0. This was used to derive a transform of temperature and time that we have called the reaction ordinate Rω, which enables the correlation of extensive kinetic data sets by means of three constants: β, γ, and the mean activation energy Ea. We demonstrate this correlation for two sets of published data. The values of γ were less than unity and are interpreted in terms of the Kohlrausch function, leading to the hypothesis of a continuous distribution of activation energies in the system. This form of correlation is similar to that of the continuous-time random-walk relaxation phenomena that has been explained in many ways including fractal time, and thus we propose that this is indirect evidence for the nonhomogenous nature of heavy-oil-cracking kinetics.