Wejo,
There should be a stated uncertainty in the instrumentation measurement. It is the problem we get in science all the time with digitization- just because an instrument can report 10+ digits doesn't mean that that is the actual measurement uncertainty.
in my book, there should be a recorded measurement uncertainty to what the sensor can report. (A machine that applies pressure and release to the system and then look at what the variability in the reported values would be would easily give you a confidence interval for the measuring, and that assumes that the clock timing of the processors is accurate to whatever nanoseconds they seem to be thinking they can report (so there is the physical response of the sensor plus the loop time of the code and the processing speed of the equipment to account for).
While the paper says that 0.1000000 or whatever number of zeros is the theoretical limit, that would then have an applied uncertainty of the measurement equipment (scale to the real world) to go with that theoretical value.
In my mind, the USATF should absolutely be filing a protest since if they really are taking a reaction time.of 0.099999 or whatever and the uncertainty of the measurement is even +/- 0.00001, this rounds to 0.01000 +/- 0.00001 with the last digit being the first uncertain digit (where the confidence interval is applied). For instance, depending on the chips used, the resolution of the timing circuit could be 1ms (it's probably better than this for professional timing systems, but at the same time, the code for doing this will.run in some sort of loop and the sensors will have a given response time as well. The error will be convolution of all of those factors.)
At the very least this as a scientific basis form those numbers need to be applied to how that rule is written and applied. I for one doubt that the pressure release on those sensors and all of the equipment in the timing can be relied on to that degree of accuracy.
Just my opinion. The way they are reporting those numbers in my mind indicates that they are commiting the common sin of carrying too many digits into the final answer and not applying the error propogation (sorry, the inner professor in me is screaming). :)