Nice. Oscar Bonilla's visualization still gives me a better grasp, I feel. Maybe you could add a graph in Oscar's style - especially his last graph?
Euler-style diagrams may give a better conceptual representation. but unfortunately they are fairly useless for intuiting about actual quantities - people are simply aren't very good at comparing relative sizes of areas, and are much more accurate when comparing lengths.
I agree that the method I've chosen could be more enlightening, I'm just not sure how to do it while retaining the ability to make comparisons about the actual values.
For more on this, see http://lilt.ilstu.edu/gmklass/pos138/datadisplay/badchart.htm
A while ago when Bret Victor's amazing article Up and Down the Ladder of Abstraction was being discussed, someone mentioned that they'd like to see one made for Bayes' Theorem. I've just completed version 1.0 of my "Bayes' Theorem Ladder of Abstraction", and it can be found here: http://www.coarsegra.in/?p=111
(It uses the Canvas html5 element, so won't work with older versions of IE).
There's a few bugs in it, and it leaves out many things that I'd like to (eventually) include, but I'm reasonably satisfied with it as a first attempt. Any feedback for what works and what doesn't work, or what you think should be added, would be greatly appreciated.