Odd that "decades of research" have not produced a single demonstration that the method works statistically, nor a single identifiable researcher willing to give even a lukewarm endorsement.