2. Understanding the GUM S1 method

It was probably to answer the question of the "form" of the probability law that the WG1 considered it useful to propose another alternative to the law of propagation of uncertainties. GUM S1 (supplement 1 to the GUM, referred to by WG1 as "JCGM 101 : 2008", which quickly became the ISO/IEC Guide 98-3 S1 standard) explains how to go about determining this law. It is based on numerical simulation, sometimes referred to as "Monte-Carlo simulation".

The idea is very simple (at first glance). It involves simulating possible values of the input data, knowing their probability laws, and then calculating the corresponding Y value, depending on the mathematical model or algorithm. Each simulation therefore requires :

• simulate a vector of values ( $x_1,x_2,\dots ,x_n$...