Just my thoughts about focal length and beam angle: I don't quite think they're related in the way that klem suggested.
Consider a perfect point source against a non-reflective wall (light emitted at 180 degrees in any 2D plane containing the point source emitter). Bring in the lens and you'll find a focal length.
Now put the point source behind an aperture that blocks all light except the central 90 degrees, or 45, or 17, or 3.445 etc. In each case the focal length will be the same (though the total flux will be different).
The only thing that matters to focal length, is where the point source is, not the range of angles that light can leave the emitter. Of course we don't have a point source and thanks to the silicone / glass lens that comes with your favourite LED it's not trivial to calculate where the 'virtual' emitter actually is. So we put down the calculator and try it out. Even knowing the exact curvature of the aspheric surface, I don't bother trying to calculate how far it should be from an LED.
As for the reflector, you get some kind of image of the reflective surface and you don't want to know how ugly that is! For most of the reflectors you get in modern flashlights, you'll find them far too deep anyway (for the aspheric to be at its focal length). So you need to cut them down and then they're too narrow to allow much light to pass through the lens.
For the XR-E it was never worth it, but for the XM-L you may benefit from a very, very wide reflector that sends light around the edge of the lens. However, for the space it takes up you could just put in 2 more XM-Ls with their own lenses.