Calculations of at least two runs with different values of a parameter
are required in order to discover to what extent the results
depend on the , if at all. However, statistical errors can
significantly exceed a systematic trend caused by changing
. To avoid this effect, we use a so-called method of
differential experiment (Ermakov, 1985[50]), when the same
initial set of points (ZAMS-stars in our case) is used. In this case a
systematic error acts similarly in all experiments and vanishes after
the results have been subtracted to find the -trend.
In order to study how the final statistical distributions depend on
the initial functions, the method of Green functions is used.
Essentially, the interval of allowable values of the initial parameter
(say, mass ratio 
) is divided into several parts,
and we perform the experiment keeping the parameter uniformly
distributed within each small interval.
Then we can restore any distribution
on this parameter (with a certain accuracy) by summing
the results for each interval with the appropriate weight.
To study dependence on the star formation rate function, we can collect statistical data for different intervals of ages of the stars. Their subsequent convolution with appropriate weights allows us to obtain the results for an arbitrary dependence of the star formation rate on time.