AS has the Resolution in a digital form that can be run through Google Translate:
http://www.as.com/misc/resolucion_caso_contador.pdf
Thanks for the link. My doc translator only got through the first half, IOW, missing all the important stuff. So I had to cut and paste small sections after that. I eventually got tired of doing that, but I think I have gleaned most of the major points. The argument that Bert had a clean passport is there, and as someone else noted, the argument that CB contamination is not compatible with blood transfusion, but no supporting data or arguments are provided in this particular document.
What I found most interesting, though, is Bert’s argument for contaminated meat. The prosecution stated that 286,000 cattle were tested last year or in some time period. This is about 0.30% of the 87 million figure (total cattle population in the EU) that I posted in another thread, which agrees well with the 0.25% figure Bert’s lawyers claimed. The prosecution said only one CB positive resulted from all these tests.
Part of Bert’s reply is this:
In order to obtain optimal results from a statistical point of view (that is to say, level of confidence of 95% and error of prediction of 1%) 8,586 head of cattle in 2007 would be due to have analyzed, which contrasts with the 97 head of cattle that indeed analyzed themselves in this year. 2) With the sample of head of cattle analyzed, the probability that the controls carried out in the Independent Community of the Basque Country, detect head of cattle contaminated with Clembuterol is extremely low of 0.001221 in 2007. This level of sampling does not allow to assure from a statistical point of view that does not exist meat contaminated by Clembuterol in the CA of the Basque Country. 3) These data, it is come off that the conducted controls are very far from being optimal the statistical ones to affirm on a base that the analyses could be able to detect, although is with a minimum rigor, cattle fattened illicitly with Clembuterol.
If I understand this correctly, they are saying only 97 cattle were tested in the region from which Bert’s steak came, which is not enough for a statistically significant conclusion. The actual amount required is about 8,500. This agrees with the analysis I provided on another thread EXCEPT that it ignores the point that 8,500 cattle (actually quite a bit more, I believe: on the order of 20-30,000, based on the 286,000 in all of the EU) were tested in all of Spain. The number tested in the entire country was plenty large enough for statistical significance. But since this number was taken from the entire country, of course only a small fraction of it would come from the area where Bert’s steak is thought to have originated.
This is a very clever argument, just the kind I would expect from lawyers. To illuminate it further, consider the situation with polling voters. Let’s say you poll one thousand voters in the state of New York, and on this basis find that candidate A is leading candidate B by a certain margin. But of that sample of one thousand voters, only 50, let’s say, live in Albany. Does this mean that we can’t say anything at all about voting patterns in Albany?
Well, yes and no. It could be that people living in Albany have a very different voting pattern from people in the rest of New York. To determine that, yes, you would need a much larger sample from that city. But you would not expect them to vote differently unless you had some independent reason for that. Voters do of course exhibit regional differences in their preferences, sometimes enormously so, and there are well known reasons for this: urban vs. rural, white vs. minority, economic status, etc.
But are there regional differences in CB use in cattle? Is there any reason to believe the region where Bert’s steak came from is any different from the rest of Spain? In the absence of any direct knowledge, no. We get into Bayesian probability here. In the absence of any such direct knowledge, the odds that any small region in Spain is vastly different from the country as a whole are very low. This is what the large-scale sampling tells you. So based on all the available data, there is no reason to believe that the situation in the Basque area is so different that eating meat from this area would greatly increase the chances of contamination.
IOW, the problem with this argument is that it could be used to invalidate any statistical sampling procedure. One can always claim that one is concerned with some very small specific region which by the rules of the game provides only a handful of the total number of samples. Remember, that even if Bert could not have traced his steak to any particular area in Spain, of course it did come from some particular area. So one can always claim that the particular area involved differs statistically from the larger environment.
But to repeat, the rebuttal is: prove it. Provide some reason for thinking this should be a special case. Without a good reason (and maybe this brother of some dealer who supposedly was busted for CB a few years ago could turn out to be a reason), the odds that this region is very different are very low.
Further, there are two known facts here the lawyers seem to avoid. First, only one example of contaminated meat was found in all 286,000 EU cattle tested. This is so ridiculously low that there really isn’t a basis for carrying out more intensive tests on a regional basis. If there was only one positive in all of the EU, what are the odds that there is one more, let alone a bunch more, in some small area that wasn't heavily sampled?
Second, while the 97 cattle tested in the Basque area is not a large enough number for a highly significant result, it has some significance. You could certainly conclude from it that any meat from that area is far more likely to be clean than not. I don’t know the exact significance, but it surely suggests that the odds of eating contaminated meat are on the order of 10% or less. This is not beyond reasonable doubt, but then again, doping sanctions do not have to be based on reasonable doubt. The WADA code says somewhere between preponderance of evidence, which is 50%, and reasonable doubt, which might be considered 99%. Where you draw the line here seems to be up to the judges on the panel. I would say 80-90% certainty is sanction able--and remember, it takes a somewhat distorted argument even to come up with this figure.
In fact, the lawyer's argument here, if I understand it correctly, is somewhat misleading. Suppose out of the 97 cattle tested, ten were positive, about 10%. The argument is that because of the small sample size, this 10% figure is highly uncertain. The actual prevalence of doped cattle in the entire Basque population could be quite a bit higher, say, 15%, or lower, say 5%. This is correct.
But in fact, 0 cattle in the 97 tested positive, or 0% So even if the small sample size makes this 0% figure uncertain, the real figure is still going to be very low, at most a few percent, which means the odds of eating contaminated meat are still very low. IOW, the lawyers are asking the panel to focus on the high degree of uncertainty, when in fact this is mostly irrelevant, because the possible upper range is still very low. It doesn't really matter much how uncertain the 0% figure is, if the worst alternative is only a few %.
Not surprisingly, given the problems with this argument, Bert’s lawyers also question whether the detection standard is good enough to spot meat that may have enough CB in it, not to cause public harm, but to trigger a doping positive. The prosecution claims that for Bert to test at the level he did, the meat he ate would have to contain at least three times, and possibly as much as 100 times, more CB than the minimum level current tests can detect. Bert’s team tries to argue that this may not be the case, that levels set to protect the public may not be sufficient to prevent levels that trigger a doping positive.
I think this is a much better point of attack than trying to argue against the cattle testing statistics. Bert's team points out that if cattle are doped, there is a substantial period following the end of doping before they are slaughtered, to allow CB to clear the system. In this way the cattle can pass the urine test, but might still have significant levels in their muscles/meat. This gets into some tricky pharmacokinetics, so there is plenty of leeway to support various theories. Though the prosecution's best case (from Bert's point of view) is that levels in meat would still be three times higher than what would be needed to trigger a positive test, I can well imagine Bert could find some experts to debate that. However, in this document, there were no specifics to back up this claim, as far as I could see.
Finally, another very interesting observation, as reported earlier in the news, was that Bert actually tested positive for CB for four days in a row. What makes this interesting is that on the last day, his level was higher (17 pg) than the day before (7 pg). I don't think his worst critics are going to argue that he had a second transfusion during that period, or that he microdosed CB. And as far as I know, he hasn't claimed to have eaten meat again during that period.
So why is the value higher? It almost has to be a background effect, IOW, the 10 pg difference reflects either imprecision in the test, or fluctuations in ongoing levels of CB that are always present in his body. Regardless of the explanation, this is perhaps the best argument this cynic has seen for a threshold value. If levels below 20 pg/ml. aren't reliable, then you can't sanction riders for having them. Bert's maximum level of 50 pg/ml is still well above this, but a little lower and you have a case.
Facebook
Twitter
Instagram