If multiple items (product warranty allowance, etc.) are being measured (items with different probabilities) and if it is possible to calculate a rational expected value, is the liability estimated using a weighted average calculated for the probability of occurrence for all conceivable results? (Expected value method)