the chi square gets lower the higher the error of observed data. You may have trouble if the error is underestimated (high chi) or background has been subtracted (low chi).
For the Rbragg, we need more information about your problem.
I think you should understand that the best fit of your refinement will be as good as the quality of your data. And needlessly to say your data are rather not good. The resolution, noise to signal ratio are very poor and you have just few peaks. Even worse the incorrectly fitted first strong peaks will greatly dominate the result. The misfit of first reflections may be caused by bad instrumental profiles. The shape suggest mainly Lorenzian contribution. Check your calibration.
It seems that the main problems are: the background (how much coefficients do you use for the fit? you can try to reduce the number of coefficients, it seems too curved) and the rutile phase (check the structure, the atomic positions mainly or thermal parameters), it seems that wrong intensities are calculated.