A theoretical sensitivity analysis of total lost time and saturated flow rate is conducted based on the method proposed in the Highway Capacity Manual (HCM). In addition, the accuracy of the timing calculation algorithm suggested in the HCM is verified using field data from three intersections. It is demonstrated that there is a positive correlation between the estimation error rates of the signal cycle length and the phase lost time. Also, the estimated value of saturated flow rate must meet the specific requirements under different saturated conditions to guarantee the accuracy of the signal cycle length. However, through analysis of field data collected on the discharge headway in three intersections, it is also found that, if the 4th vehicle is set as the initial spot for the stable discharge headway, as is recommended in the HCM, the error of the phase lost time will be over 40% when the line length is over 10 vehicles. Moreover, the calculation error for signal cycle length is not guaranteed to fall within the 15% range when the length of line is over 15 vehicles. It is suggested that, to improve the applicability of the HCM method, a more accurate description of the distributed regularity of the discharge headway is necessary when calibrating key parameters.