For very large microstate energy errors, these low-order formulas disagree with estimates from Monte Carlo simulations of error propagation.