I'm going to commit all these changes, since they all look pretty reasonable.

TODO: Figure out normalizations for nfrxry.

• Thought experiment suggests that, even though I get the norm "correct" for n=2 for the moon distribution, that might just be symmetry. Thus, I should go back to the "rugbee ball" distro where I expect that samples with ordering like (1,1)(2,2) should be MORE probable than samples like (1,2)(2,1).
  
  Changes to be committed:

modified: debinningCore/dataGen.py
modified: examples/nfrxryDemo.py
modified: examples/quadSortingDemo.py
modified: ipython_notebook_slides/debinningWithNewStatistics.ipynb
modified: ipython_notebook_slides/orderStatistics.ipynb