We investigated the atomic fill site probability distributions across supercell structures of RT12-xTi (R=Nd, Sm, T=Fe, Co). We use a combined molecular dynamics and Boltzmann distribution approach to extrapolate the probability distributions for Ti substitution from lower to higher temperatures with an equilibrium condition to assess how temperature affects the predictability of the structures fill path. It was found that the Nd and Sm based Fe systems have the highest filling probability path at lower temperatures but the cohesive energy change due to Ti substitution in Sm and Nd based crystals indicates that a more stable system could be achieved with a combination Co and Fe in the transition metal site.