I appreciate the analysis and effort and look forward to the next post, but just as a note, the increase in r-squared when adding angle to the regression is basically meaningless. Mathematically, adding variables to a regression cannot reduce r-squared. (Unless you’re using Adjusted R-squared.) I’m not really a statistics expert, but here’s a brief blurb from Wikipedia, if you’re curious:
<a href="http://en.wikipedia.org/wiki/Coefficient_of_determination#Inflation_of_R2" title="Inflation of R2"

]]>