Some video I took in Professor Rachel Pepper's lab to see the separation in a Newton's cradle after the initial collision. The party line when it comes to the dynamics of a Newton's cradle is that conservation of energy and momentum require that the speed of the outgoing ball matches that of the incoming ball, but this is an oversimplification. With more than two balls involved, there are an infinite number of solutions which satisfy these conditions. You can see in the initial collisions (videos 1,2, and 3) that the balls scatter, each with its own speed. However, because of this scattering, every subsequent collision is a two-body collision, and so is fully constrained by conservation of energy and momentum. Video 4 is particularly illustrative in this regard.
I was inspired to take these videos by reading Herrmann and Schmälze's paper "Simple explanation of a well-known collision experiment" as well as some followup articles.