Charles Dillon comments on Use Normal Predictions