Data collected in real S-on-1 LIDT experiments performed with a nanosecond, 1064 nm automated station are used to calculate the damage probability with the ISO-recommended (conventional) method and the recently-suggested cumulative method. The damage probability points versus fluence for each type of calculation are fitted using both, linear and nonlinear curves. The resultant four data sets corresponding to each real experiment are used to compare important parameters as: statistical uncertainty of damage probability points, fitting errors, damage threshold fluences for actual number of pulses, and the extrapolated threshold fluences for very large number of pulses. We suggest and analyze also a limit case of the cumulative method, when the damage probability points are calculated for each interrogated site. Both, the recently-suggested cumulative method, and our limit case, look very promising.