I'm no scientist but to try and stop this conversation going bonkers, here is what i have come to understand about it. Say a cement based adhesive once cured, has an initial tensile bond strength of 2.6N per mm2. During complete saturation in water that bond strength reduces to 1.1N, but will increase again when it dries. (These figures taken from the datasheet of Mapei Keraflex S1 - don't ask me why this happens.... maybe mapei can answer that for you). A ceramic tile full of water will increase in weight (as
@Plan Tec Tiling proved) and keep the adhesive saturated. So dry you have a higher bond strength holding a light tile compared to wet where you have a lower bond strength holding a heavier tile.
With solid bed fixing you slow the rate at which water can pass through so the majority of the tile and adhesive will stay dry in intermitent wet areas (showers and wet rooms) with the higher bond strength, but also have full contact with the tile for even weight distributon and holding capabilities over a larger surface area. If solid bed fixing is not achieved you have free movement of water between ridges of adhesive that have not been collapsed or worse, between 5 small dots of adhesive. This causes both the tile and the adhesive to saturate quicker, reduce the bond strength and have sporadic surface coverage trying to hold a tile of increased weight as its full of water. All this in conjunction with ill prepared substrates such as plaster, plasterboard or plywood that turn to mush over time because of moisture ingress and you are looking at a catsrophic failure. Even with a tanked or waterproof background, with insufficient surface coverage of adhesive you are running the risk of failure due to a small amount of weakened adhesive no longer being capable of holding potentially an increased tile weight.
Does that make sense and does it answer your query?