Off-Topic:
Gullwing Doors:
On-Topic:
I think it's a shame that people are saying what is the best, and what is not. To me, the best shape is one that productivly holds all of the pieces that your strategy wants to do. How your wheels are doesn't necessarily have to be what your body shape is.
In 2002, Team 247 wanted to grab all 3 goals. They decided on a triangular shape, with 3 wheels in "crab-style). When all was said and done, weight issues caused the third goal-grabber to be scrapped. How well did it work? Finalists, Semi-Finalists, and Quater-Finalists at 3 regionals, including a 7th seed at one of them. (Note: This was 247's best year)
In 2003, 247 chose against the Triangle design because they wanted to make a wedge that wasn't always out. It had 4 positions, 0º, 45º, 90º and 'Closed'. When at 0º, the robot became an effective wedge that could fit under the bar, and push around teams such as the TechnoCats with seemingly no effort. At 45º it could be used to side-swipe the bins in autonomous to get more stacks downed. At 90º it could block portions of opponents view when strategically needed (such as a cornered bot, driver doesn't know what direction to turn). When closed, the robot had it's small footprint.
By having this mult-purpose wedge, weight was saved giving 247 room to create a suction mechanism, which aslo worked very well.
Also, the 2003 bot, with it's rectangular deisgn, had 3 wheels. That's right, 4 corners, 3 wheels. The layout of the wheels was the same as in 2002, but a different speed shifting style and wheel type.
Sorry this was so long, but my opinion is that your strategy should define your building type. Don't jump to conclusions about things. For example, the hemi-sphere.. using equations it is the best bang-for-your-buck, but if you have no use for a hemi-spherical design to do the job you need, what's the point. Same with every other shape.
Find out what you want to do, without a body design in mind. Then find a body type that suits it well.