Our practice as an OEM is to supply the "minimum" ATG cases as
part of the data package. We will also flight match the aerodynamic model to the limits of the available aircraft certification data, recognising that the ATG cases are only a sample of the aircraft. We will also provide additional proof of match data as part of the package to validate any unusual characteristics we think might prove contentious during qualification of the sim.
(I'm mainly talking about the flight model here; I'm not so familiar with our systems simulation approach, and that is certainly more fraught with problems of configuration control; generally once the aircraft is in test the likelihood of a significant change to the basic aerodynamics is small, and we can use most if not all of the developemt data to build our models. Systems may still be evolving up to (or even beyond) the certification date, so there's never really a firm, fixed basis for the systems side of the work.)
To some extent the sim manufacturer has to use "seat of the pants" to fill out the inevitable holes in our data; they know (or should) what sort of thing works for training purposes for cases where we will never have the data. To some extent it's an emperor's new clothes situation, in that past practice ends up validating what you do next, because it 'works'. But that's pretty much a definition of engineering
.
While there are undoubtly things the authorities could do better - and not always in the direction of requiring more validation, there are things they could do to make everyone's job a lot easier while sacrificing nothing in terms of fidelity - there is, of course, a qualitative eval by the authorities, it's not just a matter of matching a few manoeuvres and ticking the box. Of course, if the eval pilot has a different opinion as to how things should be than anyone else.....