I'm going to share with you my understanding on this as I was told by a few people who are more in the know concerning x-ray fluorescent spectroscopy, I am not a scientist and honestly, I do not think anyone on this forum can truly say that they understand it's nuances. What I do know is what Tyler(wildcat123) correctly mentioned earlier, for the spectroscopy to work, It needs a certain amount of data(known originals) to use as comparisons. What that number is, is unknown to me but, I believe it requires hundreds of comparatives. The more comparatives, the finer the data. While 99.9% sounds great, I think you'll see that it's a an incorrect yet visionary number.
It's my understanding that this process worked very well on aged decals, such as those found on Third Reich helmets. All decals were manufactured in similar plants under similar conditions and applied to the helmets in the same fashion by the manufactures. This meant that there were more comparatives for these than there is for U.S. helmets. Now, the jury is still out concerning the TR decals and many folks who are much more in the know in the TR helmet world will tell you that this is not working well and has only limited merit.
The problem arose, again from my understanding, that spectroscopy does not work well when it came to paint. There just simply weren't enough comparatives. This was based on multiple formulations of paints made during WWII by both the Allies and Axis alike. Each company & each batch of paint had different levels of pigments, binders, bases, lead, solvent etc. Add to that, the enviroment in which the paint was made, stored, and/or used, played into the way the paint ultimately looked(chemically and visually) once it was applied and dried.
Example: A 506th A/B trooper paints a spade on his helmet using white paint supplied by a U.S. manufacturer. On the other side of the camp, another trooper is painting his spade on his helmet using white paint supplied by a British manufacturer.
Both helmets look exactly the same on June 6th, 1944, they have identical white spades. However, chemically, these two spades are completely different. They have: two different manufacturers, two different formulas, two different paint binders, two different pigment combinations, two different levels of lead contents, etc. BUT... they look exactly the same.
Fast forward to today: Which one of the two would you use for the legit comparative? Now multiply that one comparative by thousands of possibilities and the spectroscopy technique concerning paint... especially paint on paint, over steel helmets, done in the field, under various conditions, using multiple types of paint and formulas, could become highly controversial. Which is exactly what happened.
Again, let me state that I am limited in my knowledge concerning X-Ray Spectroscopy but, I do know paints, having worked in the automotive paint industry for a great many years. The debate concerning formulations is a very legitimate argument thus, I think spectroscopy, as a collectors tool, is open to error.