Given discussion with Micky (2018-05-08), add a note about the convergence to the true parameter in the limit of large samples in the Error-on-means.ipynb. It converges to your true parameter, but how quickly is relative to the size of your uncertainties (and hence weights).