It is well known in forensic science that any claim by the prosecution about the likelihood of the truth of a hypothesis lacks a sound logical foundation. The consistent treatment of anti-doping as a forensic science may be a first major step towards a much needed revision of the rather marginal role for the defense. For more details, see K. Faber and M. Sjerps, “Anti-doping researchers should conform to certain statistical standards from forensic science”, Science and Justice, 49 (2009) 214-215.

Let’s forget about the abstract logical problem for a moment. The UCI experts claim that doping is ‘highly likely’. Should that be convincing? Not at all! That verbal statistical statement translates to a numerical certainty in the range 95-99%, which has never been good enough in a doping case. The IOC decided many years ago that the certainty should be at least 99.99% (‘practically proven’). Why? With ~250,000 conventional tests per year, that could lead to an innocent athlete being convicted every two weeks – with any real defense essentially being ruled out. That was believed to be an acceptable risk of competing.

I just wonder: why didn’t the UCI experts do their homework on two essential issues? The UCI experts are invited to provide a rigorous calculation to support their verbal statistical claim (‘highly likely’), which should not even be convincing! I’m absolutely confident that it’s a mere guess, their personal belief that this rider is guilty. Clearly, such beliefs should never enter an expert opinion.

Klaas Faber

Independent anti-doping expert and former forensic scientist