This thread highlights one of the problems with understanding what certification is about.
There are no guarantees, only measures of probability (read "very educated finger in the wind guesses") and repeatability (read "it's not much use if the data used is too airy-fairy").
Vmcg is a very artificial concept intended to provide the low speed line-in-the-sand for takeoff sums. Whether you can or cannot maintain control a bit below (or above, for that matter) (the certification) Vmcg on the ASI depends on a bunch of things which relate to how close you are to replicating the certification conditions .. thrust (flex, elevation, OAT), CG and crosswind, especially.
NWS, for instance, will be influenced greatly by CG .. the certification approach ignores it on the basis that, for the aft CG case, it's probably not going to be of much use in any case.
For the line operation situation .. if you are faced with a speed schedule on the Vmcg (or Vmca) limit, then consider the factors which go into the certification numbers and how your actual situation/configuration might be positioned .. but be prepared for the need for rapid decision making .. in the case of failure on the ground, you might well be in the situation of being near-instantly out of control ... of course, you snap the throttles closed and hope that the thrust follows .. otherwise it is time for dodgem car antics. If you have a failure in the Vmca limiting case, say, early in the rotation .. you may need very aggressive aileron control, according to Type characteristics.
And, if we sit back in the smoking chair and contemplate the matter over a wine .. why are we launching at minimum speeds ? If the runway is longer than needed and it is just a case of low weight .. why not adopt an appropriate higher speed schedule and avoid most of the worry ?
If everything is loaded up against you .. and you have to go (ie can't delay or do something else to reduce the risk) .. that's fine .. but one would feel a bit of a goose crawling out of the wreckage knowing that one really could have waited half an hour for the wind to die down a bit before launching ...