Revisiting the Battle of Midway: A Counterfactual Analysis

摘要

Lanchester-type combat models have been widely used to study the outcome of battles and to carry on counterfactual experiments to analyze the consequence of commanders' decisions, tactics, concentration/dispersion of forces, quality-versus quantity, and so on. This paper uses a stochastic salvo combat model to study the Battle of Midway. The model is calibrated according to the historical outcome. The model is then simulated and used to study four alternative counterfactual scenarios: (i) all of the launched American attack aircraft reach the Japanese carriers, (ii) the Japanese have one additional carrier, (iii) the Japanese do not wait to launch their attack aircraft, and (iv) the American carriers are spotted earlier. Contrary to the common wisdom that the result of the battle was an "incredible" American victory, Monte Carlo simulations show that the probability of the Japanese winning was very low and actually close to zero. Even in the most favorable scenario for the Japanese, the Battle of Midway remains an American victory.