We claimed that the critical fractile formula, Equation (12.8), is appropriate because the optimal Q should satisfy cunder(1 - F(Q)) = cover F(Q), that is, the cost of under stocking times the probability of under stocking should equal the cost of overstocking times the probability of overstocking. Assume that Q satisfies this equation [which is equivalent to Equation (12.8)]. Use a probability argument to show why Q - 1 and Q + 1 are both worse than Q in terms of expected cost.
