Powder diffraction crystallography pioneer Peter Stephens once joked that a Rietveld refinement is never perfected, but merely is abandon. This is because once one has found a good fit, there are a seemingly infinite number of things that can be attempted to make the fit better. Most will not. Nonetheless, it is worthwhile checking the sample profile models to see if a better fit can be obtained by adding a small number of parameters. One can test if LGmix refinement improves the fit. Revert to the previous model if not. Likewise, I have sometimes been fooled by good fits where I thought I was done, but I was surprised to see them improve further when I included anisotropic microstrain (irony here:) using Peter Stephen’s microstrain model. Again, check to confirm that microstrain and crystallite size broadening can or cannot be refined together. Sometimes this will not be stable in the early stages of the fit, but will work well once the structural and background parameters have reached good minima, sometimes things seem to start well, but end up with values that don’t make sense.