Not sure if this matters, but I'm using this for nonlinear equations, but it's still the same concept for the most part with the exception that I'm using newton raphson iterations to find the final solutions. I'm trying to figure out why weighting the errors doesn't wildly alter the solutions for the state variables? Obviously it gives greater weights to some "measurements" (now you know what I'm using this for) but how does it not just make the answers wrong??? I can't really visualize why - please help me out here. Thanks!!