This phenomenon is not new in the slightest. The board's leading technical experts and pioneers,
rohins and myself,
n00b, have already investigated the anomalies of the charge shot trajectories.
Our results were inconclusive for the longest, until
rohins discovered the specific input to trigger a "loop"ing path in the shot.
Right as the shot leaves Samus's cannon, U HAF 2 UP B. The up B input combined with the airborne reverse momentum (think Newton's third law) will cause the spin attack's torque around its initial hitbox to transfer to the charge shot, since Samus's animation is stuck in recoil from the gun. This results in an upward looping path of the charge shot. The reason it has a large, sweeping curve for its trajectory is because a fully charged charge shot has a larger radius and higher mass than Samus herself. This is explained by the tightness of her Up B spinning as a result of her tucking her body in, reducing her corporeal radius, which consequently increases her tangential velocity of the screw attack. You can see the difference in mass between Samus's charge shot and herself, because Samus's dash attack (she hits with her whole body) does not have anywhere near the knockback of a full charge shot. (F = ma, where m is mass and A = acceleration. Samus's dash attack and charge shot move almost at equal speeds, meaning the difference in force is due to the increased mass of the charge shot)
The phenomenon is further detailed in the following thread:
http://www.smashboards.com/showthread.php?p=6612170