The definition of average velocity is:
![]()
According to this definition, the vector average velocity is
obtained by dividing a vector by a scalar.  Since the scalar (
) is always positive,
this definition indicates that the two vectors (
 and 
) have the same
direction.  We would like to apply this definition to find the velocity at the
turning point.  Then 
 and 
 in the definition
correspond to an interval that spans the turning point, i.e., with an initial
point just before the turning point and with a final point just after the
turning point.  The diagrams below illustrate this idea graphically for a ball
that is turning around.
Suppose
a ball rolls uphill past point 1, turns around at point 2 and then rolls
downhill past point 3.  Since the ball has turned around, points 1 and 3
correspond to the same location.  The position of the ball is a vector that
points from the origin to the location of the ball.  The location of the origin
in the diagram is arbitrary.  Its location will not affect the velocity we are
trying to compute.  Since points 1 and 3 represent the same point, the position
vectors at these two instants are also the same.  If this is true, what is 
?  What is 
?