Here, I propose an enhanced statistical approach for simulating electronic circuits using LTspice, complemented by statistical evaluation methods implemented in Excel.

Advanced statistical simulation of electronic circuits
Anyone who has ever simulated an electronic circuit with Spice may be familiar with the functions that can be used to investigate the effect of the component tolerance.
On the one hand, there is the Monte Carlo simulation, where you can assign a tolerance to component values and with each simulation run the random component values are changed within the specified tolerance. The result is a set of curves in which the width of the tolerance can be determined using the cursor.
Some Spice simulation programs also offer the option of parameter analysis, whereby one component is changed in turn with the maximum specified tolerance and all other components have the nominal value. The result shows the influence of each individual component, whereby there is no statement on the effect if other values also have a deviation from the nominal value.
In some programs there is a "real" worst case analysis that changes all component tolerances on the basis of the parameter analysis so that the output value under consideration is either minimum or maximum, "fake" worst case analyses simulate all MIN MAX variations.

you will find the complete tutorial in the pdf