You are assuming that the equilibrium assuming complete fairness is 50% women in every profession. Which will simply not be the case (doctors is a good example, the female ratio is more like 60% today). And if it is off in any direction from 50% then the gap cannot be measured ignoring the profession.