ambitious thinking with this one but potentially making a mountain out of a mole hill. Id suggest that the wheel rotation is more trouble than its worth considring how small scateboard wheels are.
If your shots are short you could easily animate the rotation by hand to get better results, this is because the script your using is only evaluating the translate x which means it would only be accurate if the skateboard is travelling in the translate X of its parents space (translating something in local space with the move tool on the is not the same as translating in X).
To do this properly you would need a script that evaluates the skateboards position on the previous frame and grab the length of the vector between to two to drive the rotation.
This is probably a bit over the top so I'd animate by hand, which means you can make it look cool if one set of wheels is in the air (spin spin spin!) which is trickier to set up if you rotations are locked up by an expression.
Also on this note you probably wont see the wheels spinning unless the camera is REALLY low, close up and the wheels would have to be quite dirty to notice.
For the parenting issue it sounds like the geo is allready constrained to something or it has keyframes on it. If its constrained it will move back to its constrained position every frame. If it was keyframed before you tried to parent it to your rig it will make a new keyframe on that frame (if auto key is on) and look correct but the rest of the animation will stay the same, so if you change frame it will pop out of position.
If this is the case you can either delete the constraint / animation and then parent it (also unlock any translation / rotation channels). Or if you want to keep any animation on it ( i guess not) you would have to bake the skateboard into the animation space of your curve. If you want to know how to do this just ask and ill write another reply.
With the banking, your hierachy should be something like this :
root curve> main pos / rot curve > wheel ctrl curve for each set of wheels, the pivot would be at the point where the truck joins the board + main board ctr (animate the wheels seperately to the wheels... banking) > individual wheel controls)
Id Imagine for your main control it might be easier to have an IK system to get a more convincing control of the whole board. To do this just make a joint that goes from back to front of the board, make an IK chain and parent the joint to your main control and another curve under to be the parent of the IK handle to control the direction.
Hierachy is v important with rigging, as is an understanding of rotation orders and transformation in order to get zeroed out translates and rotates for each controler.
I know this is a bit much, its worth reading up on the topics to solve the problems.
With the deformations im not sure what your asking for, the main deformations would be in the bend of the board itself which could be achieved using simple deformers, clusters, blend shapes etc.
If your trying to make a toony board you will need to look at topics like volume retention, joint expressions etc.
Hope this kinda helps.
FX supervisor - double negative