I would say go for the top row. This year is interesting in that the main game functions have been done and well documented. The gamepiece manipulators can be derived from 2007, and the lifting mechanisms from 2007/2008. There is a lot of value in looking at the successful robots from those years.
For lifting you generally have the choice between a multi-stage elevator or an arm. Both are rather simple mechanisms requiring 1 driven axel (usually 1-2 motors) and both can reach the top row without any excessive complications. There are pros and cons to both, but none are actually as significant as some would make them seem.
Remember, the inner tube has to be hanged, this does not mean that the lifting mechanism must reach the top rung. The bumpers must surround the entire frame perimeter. Your game piece manipulator must have some way of extending beyond the bumpers to pick up pieces (unlike 2007). If designed correctly, this mechanism can also get you the extra height boost to reach the top rung and if you're clever, you may not even have to power it.
Below are images of a quick sketch I drew earlier today illustrating an example of a basic robot design that can hang tubes on the top rungs of the scoring grid. (The gripper assembly could be powered by pneumatics or motors and the parallel bar linkage would be driven by motors with some sort of reduction.)
http://i.imgur.com/rDKI8.jpg (Dimensionless)
http://i.imgur.com/WjsFL.jpg (Dimensioned)