The whole trick to a scene like this is to make it full of energy that looks spontaneous and real. What appears to be a simple grab shot is a staged scene that was repeated over and over until we were sure we had it. The attitude of the models had to be right and the camera had to freeze the action and deliver a pin-point focus. Making this situation even tougher I placed a 1000 watt tungsten lamp behind the models and had it shining directly into the camera lens to create a flare. A strong back light, such as this, is going to substantially reduce the contrast in the foreground scene making it even more difficult for the camera lens to grab focus.
Shooting situations like this and the one from my last post illustrates the importance of matching the right camera to the job, and why for a job like this I choose a Nikon D4 and Nikon lenses. Very few cameras can deliver such a consistently spot-on performance.