New molecular beam experiments on the dissociation probability S-0 for N-2 on Ru(0001) are presented. These are in general agreement with prior measurements and exhibit very unusual behavior; a very slow increase of S-0 with incident kinetic energy E and the fact that S-0 is still only similar to 10(-3) at incident energies considerably above the barrier. A simple dynamical model is developed to describe this unusual sticking behavior. The key aspect is that there is considerable energy loss Delta from E upon initial impact with the surface (principally to the lattice) and only E-Delta is then available to surmount the activation barrier in the exit channel. Using experimentally measured values of Delta from scattering experiments gives good qualitative agreement of this model with the measured S-0. One implication of the strong energy loss is that there is an apparent violation of detailed balance when comparing only the reactive fluxes of activated adsorption and associative desorption.