I'm sure this has been studied to death, but my guess is that the extra risk during flight/separation, and the extra weight of designing the rocket to handle th

I'm sure this has been studied to death, but my guess is that the extra risk during flight/separation, and the extra weight of designing the rocket to handle the different forces encountered when attached to the mothership are not compelling enough to make it worth the extra few % in delta-V. (SpaceShipOne went to Mach 3, not 25, so the boost was greater by percentage)