Text this: Prediction and policy: Do empirical gross calorific value prediction help reduce coal testing overload?