Sachs-Wolfe effect So for photons at the bottom of the well at the time of last scatter, they will climb out of it losing energy. But, from there on, the potential wells (density fluctuations) do not just disappear, so those photons will keep going through the "uphill" and "downhill" motion while travelling towards us, so why wouldn't the effect even out by the time we receive the photons? Why is it guaranteed that these photons will still look cooler due to the initial climb?
 A: There are two reasons the "uphill" and "downhill" motion do not always cancel out.

*

*The photon originated at the top or bottom of a hill. This is the ordinary Sachs-Wolfe effect. It implies that (at large scales) hot spots on the observed cosmic microwave background (CMB) came from cold spots in the early universe, and vice versa. This is because hot spots have excess density, which gravitationally redshifts exiting photons, cooling them beyond their original temperature excess.


*The hill's height changed between the photon's descent and its ascent. This is the integrated Sachs-Wolfe effect. It generally does not happen during the matter epoch; density variations then grew both in amplitude and in size (the former due to gravity and the latter due to cosmic expansion) in such a way that the potential remained constant. However, that was not true during a brief time period shortly after last scattering, because there was still a significant radiation density at the time, which slowed the growth of structure. It is also not true today, because dark energy is slowing the growth of structure. Thus, both shortly after the CMB was emitted and at recent times, gravitational potentials decay, leading to integrated Sachs-Wolfe effects.
