From what I have seen most modern strains will start to revert back to their "wild side" after a few generations of inbreeding or back-crossing. This is not always true, but IMO it definitely takes a good amount of selection once you get past that f1 or f2 stage with modern strains to prevent deterioration. Also, I think often the dominant traits tend to not be the traits that we humans always want either. Traits like high potency, and purple coloration, tend to be recessive from what I have noticed. Traits like extreme vigor and high yield often go hand in hand with lower potency and poor quality, so though those traits are good for the plant form a survival stand point, they are not usually what we human want (Unless maybe you are a cash cropper who cares nothing of quality, and there are plenty of growers out there like that.) Hence, the undesirable traits start to pop back up again when modern strains are haphazardly bred with no selection.
I'm no scientist though, so I can't say for sure this is how it works, and I can't cite any solid scientific info to back up that claim, it is just what I have observed. You have to remember that 99.99% of our strains today are hybridized, there are virtually no true parents left out there. When dealing with a true landrace I would think the genetics would be much more stable and less likely to deteriorate over just a few generations. Finding good, unrelated genetics, to create true F1's has become very hard IMO.
Oh and Sea you can select for the expression of hemp, I know someone that more or less did that. They were growing some old school Mexi strain, that at one point may have been good, but they inbred it for like 10-15 generations with the ONLY consideration being yield. The result: an early, "bullet proof" plant that was high yield and super vigorous, but that produced buds that were almost totally 100% absent of trichrome heads and smelled and tasted like wet hay.