They plan (or at leats planned in 2017) a direct return to Earth without entering Mars orbit. See the slide headed "Mars Transportation Architecture" on this presentation.
I originally wrote:
Given that, the only reason not to simply boost straight up is the
possible small advantage from Mars' rotation.
It has been pointed out to me that this is wrong. It took me some time to see why. The key thing is that you want gravity to be acting, as nearly as possible, perpendicular to your velocity, not (anti-)parallel to it. That way it changes your direction but doesn't reduce your kinetic energy.
At the equator the rotation is about 0.3 km/s, so, absent atmosphere, launching horizontally due East at the right time of day would save that much. At higher latitudes, the gain is less. More or less the same gain would be available by launching upwards until a few hundred meters clear of the ground and then turning sideways, which avoids the need for different landing gear.
On the other hand, although low density, Mars' atmosphere extends up quite high, and exerts significant drag at high velocities, so spending extra time in it will have a performance cost. I don't have figures for that to be certain, but I suspect that as on Earth it will be necessary to pick a trajectory that balances the desire to exploit rotation and avoid gravity drag with the desire to get clear of much of the atmosphere before building up too much speed.