Letter to the Editor-Some Observations on the Use of Lanchester's Theory for Prediction