In order to evaluate the results of the Biologic Passport, the UCI adopted the so-called "ABP-method". 
This method compares data of the athlete with values of universal variance, or established a priori: 

- for Hb = 28.22 

- for OFFs = 71.74 

These are nothing but AVERAGE VALUES of individual variances, or within-subject variances, obtained from several studies (Haematologica 2006;91:356-363. tab. 1 pag. 357) carried out on: 

- 80 French rowers, subjected to 409 tests: an average of 5.1 tests for each athlete 

- 288 Italian footballers, subjected to 841 tests: an average of 2.9 tests for each athlete 

- 124 German athletes of different sports, subjected to 603 tests: an average of 4.8 tests for each athlete 

- 630 sportsmen of different disciplines, subjected to 1731 tests: an average of 2.7 tests for each athlete. 

It's easy to notice that the number of tests being subjected to each athlete is very low; therefore, the use of such data with the purpose of establishing a universal variance value is at fault. 
Moreover, it's not even data on a homogeneous group of professional cyclists, but a mixture of athletes of different disciplines. 

The strongest objection though, is about the decision to compare individual values to the average value of a population, without considering the variability within that population. 

The same Michael Ashenden and Yorck Schumacher, among the "Founding Fathers" and "UCI Experts" of the Biologic Passport, state at pag. 357 of that same article (Haematologica 2006;91:356-363): "In theory it would be possible to estimate a different value of variance for each athlete, but this would make any application of the formula very complex and would require at least 6 samples for the baseline reading. Therefore we opted to use the same variance for all subjects". 

Calculating an individual variance is not really difficult; evidently, the Authors needed to have and use a variance value obtained with less than 6 tests.... 

Even WADA, at page 3 of the BIOLOGICAL PASSPORT OPERATING GUIDELINES, states: "The collection and monitoring of values... will constitute an individual and longitudinal profile... with a subject becoming his own reference. This contrasts the traditional approach of the athlete's variables being measured against norms in the athlete population at large". 

While WADA declares profiles with a 99.9% specificity as sanctionable, the UCI discretionally takes upon itself the right to sanction lower specificity values, on the basis of rather unspecified "corroborative evidence" (article 23 of UCI rules, version 3.12.09). 

Essentially, the UCI wants to be able to sanction whoever it feels like, outside of every definite code of rules, utilizing an approximate and faulty method.