This doesn't really seem to pan out in reality. Often people with small dense LDL have it as a result of other risk markers. There's a good discussion here - https://www.youtube.com/watch?v=kplh30RmYo8 - the long and short of it is that when you perform mutual adjustment for number of particles, both large and small LDL particles are associated with a fairly similar risk. But if you have the same mass of cholesterol in two individuals (LDL-c) and one has SD LDL and the other has LB LDL, the latter has fewer LDL particles and therefore lower risk, so it can appear that LB LDL are less risky if you're only measuring cholesterol mass.

As for statins not working well in primary prevention, I'm not sure what you mean - the JUPITER trial showed significant risk reductions for primary prevention, for example.