I suspect it's the same with the dog. The fallacy in the writer's theory is that calculus came first and the behavior follows. In truth, calculus is simply a tool used to describe the behavior that already existed.
Why do you even need to assume their might be ANY intellegence? How does a cannonball know exactly what path to fly based on complicated physics of ballistics?
If I throw two balls in close succession, the second ball lofted while the first is in the air, my dog will always lose interest in the first ball and more often than not, fail to catch the second ball.
If I throw the second one after he catches the second and begins his return, he will immediately drop the first ball and go after the second ball.