Too many variables. Too many unknowns
Some ideas based on MIT and US Navy studies on understanding the dynamics of three-dimensional objects freely falling through water.
As a crude rule of thumb, objects with a centre of mass offset from the centre of volume will tend to behave as follows:
Small offset: flutter, see-saw oscillations, tend to horizontally aligned during descent.
Medium offset: plane away from the vertical dependant on degree of offset.
Large offset: near vertical descent in direction of heavy end of object - will usually flip once on first entry if CM is above CV at entry.
Aspect ratio will influence drag and stability, eg an elongated shape will have a different cd falling on the long dimension than end on. Tumbling motion is also observed for objects with small aspect ratios.
Heavy compact objects, for example rocks, when falling freely in water show a variety of characteristic motion patterns determined by their shape. An object might rotate around a single axis, which continuously shifts its spatial position, or develop a single-axis rotation where it rotates around the longest axis, which usually remains perpendicular to the direction of flow.
For a rotating object, the trajectory will be influenced by the direction of rotation and the angular momentum.
Given the above, is there anything we can say about the heavy bits.
Engines: Heavy and compact, and likely to fall near to the vertical.
Unknowns: Nature of attitude at impact - was there any momentum in the shaft which could possibly result in the whole structure rotating once free in the event that the shaft seized and as a consequence influencing the trajectory - possibility of tumble if initial condition CM>CV.
APU: Heavy and compact, so likely to fall near to the vertical.
Unknowns: At what point did it become detached from associated structure and enter free fall.
What else?
FDR: Not heavy, but compact cylindrical shape (missing ULB) and likely to have CV close to CM, so good candidate for vertical descent.
Unknowns: At what point did the FDR cylinder separate from the empennage before free fall.
Conclusion: Too many unknowns....