It seems that I need to calculate nfrVrH theoretically. I think the normalization is simply incorrect. This makes sense, because the n! within the nKr comes initially from the sample probability, but there is only one of those entering the entire function... Thus, I probably have one too many n!'s.

However, I don't think I can just divide by n!... So, I'd better calculate what
the normalization should actually be.

Changes to be committed:
modified: examples/nfrVrHDemo.py