edmajor wrote:
Cyclingfan wrote:Working from an incorrect axiom of the upper limit being constant. So regardless of whatever computation was done, it's a case of rubbish in rubbish out... unfortunately.
In practice anti-doping agencies are using variable thresholds, but my calculations were exact tracings of numbers quoted by Sky News and there mentioned academic paper. So in that context, 1 in 70 likelihood that she's innocent given she scored 109.86 in the test is very much correct.
One thing I'm not sure is how the paper's testing variance compares to testing errors in practice, but one would assume it's comparable...
Yes under those limits of course, however they don't have a basis in reality. I can give variables on how fast a block will slide down a slope of a certain angle. Do the calculation and get the correct answer. However, inertia isn't taken into account, friction , the difference in materials, drag, any minor changes in gradient , the block will slide at different speeds in different locations because of localised differences in the gravitational field etc etc.
I'm not doubting the calculations precision but it's accuracy. Due to the limit changing as a function of time, which isn't accounted for in the calculation.