I can't see how friction is factoring into the directions the balls initially scatter, as it's not taking the rotation of the balls into account. With that bein

I can't see how friction is factoring into the directions the balls initially scatter, as it's not taking the rotation of the balls into account. With that being said, wouldn't the balls all go in the same initial direction as they do in the linked example? Aka, the ones that are moving would move in that direction. Theoretically they could eventually bounce off walls and collide into each other in such a way that they eventually stop (though then there's the possibility of other balls that are still moving colliding into them later).