I'm working on tRA again, and what I'd really like to do is change it from a strict linear weighting to something more dynamic. It probably won't be using crazily complex equations to fit all the data perfectly or anything.

I'm not much of a maths guy anymore so ... so far:

• Figure out xO/PA (trivial). This gives you an idea of how many men are on base.
  
• Need to derive a linear (maybe parabolic?) equation which gives the linear weights already derived for an average xO/PA but changes when you're off average

I have no idea how to do the last bit, so any suggestions are more than welcome.