I have given this question (and my paper) to a PhD who works heavily with probability and has a track-record of accomplishments. Also, I have another idea on empirical verification. So hang on...