I must admit to never having considered the question that there might be a good reason to try this kind of manoeuvre. Backward departure through translational lift? Military skill to depart-from-dodge quickly?
Anyway, the backward manoeuvre can be even more interesting, if you are not careful. I have not tried to fly that profile myself, but someone did demo it to me recently (very windy day).
When you increase speed and pass through translational lift, as the disk flaps back the direction of travel means that it lifts the tail (it would have been the nose in fwd flight, and we simply push through with the cyclic). Going backwards, as the tail lifts the air-flow then gets under the horizontal stabilizer (if you have one) and lifts the tail further. That creates a positive-feedback loop as the angle of attack on the stabilizer increases. The ship will rapidly adopt a "very severe" nose-down attitude. If the pilot is not anticipating this event there is a natural tendency for him / her to pull-back on the cyclic which - particularly in a teetering head rotor machine - creates a severe risk of amputating the tail. The control input reactions are therefore not intuitive.
Glad to have been in the demo, and even more sure than before that I will not be trying to effect a departure backwards.