I think these calculations are precise, but the input data are not. I can get as many decimal points as one would like, but it makes no sense to go below a millimeter. Besides, for better or worse I made an assumption about the likely saddle height.