Pejsa makes two assumptions to derive his formulas:

1) at a given mach interval the drag function can be approached by a power of the mach number g(m)=a*m^b a and b are real numbers. With the custom drag funtions BfX determines for each interval a and b. Then applies Pejsas method. Numerical integration assumes a constant drag value for a given interval.

2) a flat fire approximation

ad 1) if the mach interval is small enough Pejsa method, and the software has to sum the effects over many intervals, Pejsa method becomes is equivalent to a numerical integration, probably equalling the accuracy of a runge kutta integration. BfX has the ability to use a measured drag function, e.g. the ones of lapua. BfX results matches here the ones of a 3dof model, see the workbooks mman posted.

ad 2) a 3dof is able to go beyond the flat fire approximation

BfX is all about generating tables. However, Pejsa publishes many simple formulas which one can use with an electronic calculator are sufficiently accurate for most purposes. In that case it is much faster than 3dof.