Well, theoretically, acceleration at the point of impact does not matter, if it is in a straight line. However, after the point of impact it may matter if the method of producing the acceleration (ie leg drive) can continue to be used by the player. But really, straight line accelerating and deceleration do not matter, only velocity at the point of impact in a theoretical model that Groucho is talking about. The exceptional problem that Groucho / Scarfman referred to is either modelling the impact in an ideal scenario in a minute quantum of time, or a more real world model where the impact takes place over a second.
Angular acceleration / deceleration, though, may be different (ie changing direction). I think this is what Groucho said above, where small guys changing their direction with no loss in speed have an advantage. By changing the point of impact, you can change how much force is required to stop you (that is where, no doubt, the biomechanics that Bruce is so interested come into play). Especially where the force is being applied (hips (close to the centre of gravity), chest, thighs, etc), or from what direction - eg if the force is normal to the guy running, or in the same direction, or coming from some weird angle, or maybe even what part of the stride/gait the player is in (eg driving off one foot during the initial collision might provide more stable force throughout the collision for the runner).
But we're now at the Eddie Jones level of analysis, I think.