With the ball in the only-slightly-larger tube, the pressure is relatively constant until the ball reaches the top, and the leaking gets significant.
Correct, you should essentially have two levels - at the bottom and at the top. Drilling small holes at constant increments would help achieve some (albeit semi-graduated/non-linear) level control. You could also make a very shallow angle cone out of clear plastic sheet rolled up, so at the bottom it is only slightly larger than the diameter of the ball, and at the top it is maybe 1.5 or 2 times the diameter of the ball. This will give you "infinite" graduations. That is how some industrial flow meters are made. If you knew the drag characteristics of the ball, you could use your device as a flow meter! That'd be pretty neat.