v2.2
Various cleanup
Replaced a "correcting" distributions contribution to normalizing the final dataset. Outputted data follows more natural curve.

v2.1
Simple cleanup of preferences. Will report blank vs 0.

v2.0

Added virtual preferences
Fixed final output drawing with newmin & newmax's to ensure when min = median, that the drawing doesn't break with /0 errors. Now any/all global weights can be disabled, and the output will correctly draw values around 50%.

v1.0

I setup roles to use a Weighted Geometric mean of 3 Monte Carlo'd input categories.

I'm hoping some people familiar with geometric means can review it, see what they think

Opens with libre office

easier demonstration
http://imgur.com/IuLdwjQ