We reconsider the treatment of Lise and Jensen (Phys. Rev. Lett. 76, 2326 (1996)) on the random neighbor Olami-Feder-Christensen stik-slip model, and examine the strong dependence of the results on the approximations used for the distribution of states p(E).