When the stress lies outside the element, then it ceases to be purely compressive. It becomes, to some extent, a pair of forces, one compressive, the other tensile.

All the forces on each pole must sum to a force acting along a line parallel to the pole's longitudinal axis and lying within the pole---preferably coincident with the longitudinal axis.

Imagine a pole that is long and flexible, with forces on it (other than gravity) that sum to a force acting along a line (We'll call it, "AB.") that lies just_within_the_outer_surface_ of the pole. Now, if that pole's position deviates from the vertical by more than just a very few degrees, and if line AB lies on the upper side of the pole, then it is possible that the force of gravity will deflect the pole, placing AB outside the outer surface of the pole.

The force acting along line AB then will behave like a bowstring. The pole will behave like a bow. Catastrophic failure will ensue.

My conclusion is that special attention must be paid not only to the flexibility of the poles but also to the manner in which forces are applied to the ends of the poles.

A ball-and-socket joint, designed so that it never exceeded its range of motion and installed at every pole-end, would be ideal for this application.