Rhino, I need to disagree with your blanket statement... there are a lot of differences between the standards other than just flange depth.
In the NMRA spec, it's pretty clear when you look at the tolerances for the measurements, while there is a target value, the "toy train" standards have such broad tolerances there might as well be no standard (yes that is over the top).
Interestingly enough, when the NMRA overhauled the "G" standards, they pretty much copied G1MRA for ONE of the NMRA standards, but actually mucked that up a bit. (the "standard" standard, not the finescale or the toy train one)
I follow G1MRA pretty much myself, but keep my flange depth to 2mm or sometimes a tad more. Also the flange thickness not too thin.
On the LGB, they are designed to be a flange-bearing frog, i.e. the flange edge is supported in the bottom of the flangeway in the frog, not by the rail tops themselves.
This is a very interesting way of doing things, rare in the prototype, and if you run all your rolling stock with identical flanges, it can work well, all these dimensions interlock and interplay in a complex way.
This means if you want really good running, you need to follow a standard for the wheels and track and switches/points/turnouts. I define really good is running a mainline train of 40 cars up and down slopes with many switches and leaving it unattended for hours.
I won't go into the explanation here, but there is a very tight relationship between back to back tolerances, gauge, and the frog geometry, you cannot radically change one without changing the others. I learned this when I thought all I had to do was set the back to back gauge at 1.575"...
My site actually explains this... it was a real revelation when I started and had a mix of equipment and tried to make things better. One of my best friends was a track inspector for many years for the real railroads, what an education.
Greg