The native globular proteins do not possess random coil structure. The random coil structure is obtained in the laboratory, by solubilizing the protein in solutions of chaotropic agents (eg 6 M Guanidinium hydrochloride or Urea 12 M). The structure that thus is obtained (random coil) is used as reference state, for example, for measurements of thermodynamic stability. The structure of which you speak about is what is topologically called aperiodic structure not fixed in time, i.e., a structure that does not have a recognizable topological organization and that continuously changes over time. Loops or fluctuating and very flexible segments of any size generally go under this name. When you get the famous sigmoidal curves of the two state denaturation equilibrium of a protein, for calculations of stability, ΔG, you must be able to recognize with certainty one of the two states and this is the one that should correspond (often but not always ) to 100% denatured, i.e., the random coil. If the denaturation of protein is a perfect two states, having a 100% denatured state as random coil is, in general, a condition for an efficient process of refolding.