During a presentation about the current risk measurement discussions in our field I realized (yeah, not enough to say "epiphany" :-)) the key issue on the current methods is the complete lack of calibration and feedback.