I think Tony is closer to the mark than you. That's because the concept of increasing L requires more coil turns, proportional to sqrt L, which means increased R proportional to L, if the wire size is reduced to keep the same size coil, so one gets no net gain in Q.
In the end the available Q is determined by the overall coil size, independent of L. This makes sense, in a Nature's revenge sort of way. Murphy shows up here too: Rac/Rdc is much less than one, due to proximity effect, and litz wire is required, even at modest frequencies, to avoid a total meltdown of Q.