Considering the steady state...

If we accept the P(t) is the product of instantaneous voltage and
current, then there will be some points on any mismatched line where P(t)
is always positive. In between those points, P(t) will have positive and
negative excursions.

I think that it is a reasonable interpretation that at those points where
P(t) is always positive, then there is never at any instant, a flow of
energy away from the load, energy is never exchanged during a cycle
across those points, it always flows from source to load.

It may be that energy is exhanged during a cycle at the load end of the
line, and it may be that energy is exchanged during a cycle at the source
end of the line, but if the line is sufficiently long, there will exist
points where instantanous power is always positive, and therefore, energy
always flows in the load to source direction at those points.

The notion that a reflected wave in general conveys power over the entire
path from load to source is not consistent with the above. This notion is
emboddied in common language when talking about 'reflected power', but
the language belies the actual phenomena.

Owen