How many models were used in the scoring process? Were there non-model, empirical scoring rules?