> Unfortunately, more incidents with cyclist will > help with this measurement of uncertainty. Wouldn't different categories of uncertainty be subject to different methods of measurement? To me, there are different classes of uncertainty. E.g., I view the uncertainty alluded to immediately above as being of the variety which becomes more certain with experimentation. Whereas, where in the OP you talk of "fixed and variables", this type of uncertainty would become more certain with the application of probability and chance? No?