What Are The Odds?

It’s been a busy day for the Rantster. Haven’t had a lot of time to read everything on TBV yet, but a couple of things caught my eye. Duckstrap has an excellent post on the Daily Peloton Forums with a very thoughtful analysis of why he believes the evidence in the lab pack can’t support a guilty finding. And TBV quotes Floyd Landis, posting in an online forum:

… What percent error would be reasonable to inflict upon someone the penalty which I have already paid. To be more clear, what is an acceptable margin of error for a test with such stakes?

This has me thinking: Just how accurate does a test have to be to be definitive?

What standard do we set in order to guarantee that a medical test gives us the proper result, and neither a false positive nor a false negative? Think, for a moment, about a test that may have life-and-death implications: the AIDS test. Just what would a “tolerable” false positive or false negative rate be?

One in one hundred? That’s pretty easy to answer. That’s way too many false positives. Imagine how many lives could be adversely affected if people thought they were infected, but in fact weren’t. And flip the coin to the other side, false negatives. One in a hundred tests yielding a false negative is as bad, or even worse, because that person could engage in risky behavior that might result in the infection of other people.

One in a thousand tests false negative or false positive? Still too many.

How about one in a million? Getting closer to an acceptable rate, but could that still be too many? It’s a tricky question to answer, isn’t it?

Now think about the accuracy of a test. How many times is a test correct, regardless of negative or positive result. Is only one wrong result in a million an acceptable rate for an AIDS test? How about a DNA test used in a murder trial?

Think about the OJ Simpson case for a minute. During his trial, the DNA evidence presented by the prosecution was said to have an accuracy of such that only one other person out of a million billion (edited per a comment to the original) would match Simpson’s DNA. Pretty accurate, eh?

Sure, but then think how many people live in the metro Los Angeles area inhabit this world. Several million Five or six billion, right? So, even with a test that accurate, there could have been a few other people in Los Angeles walking around whose DNA samples matched the evidence from the crime scene. Is that enough to show reasonable doubt? I suspect it was for the jurors.

If the DNA test would only match one other person out of ten million billion or a hundred million billion, it would have been more convincing, wouldn’t it? Simpson managed to find the magic “Get Out of Jail Free” card. Or almost free. He’s got some pretty hefty legal judgments against him that I doubt will ever be paid in full.

And as the outcry over his book and TV interview shows, there are many who scorn the man. But he’s still out there on the golf courses, looking for his wife’s “real” killer. (Perhaps he should look in the mirror.)

Anyway, let’s get back to the main character in this sad saga: Floyd Landis. He’s paid a hefty price already. He’s lost his job. He’s lost the respect of his peers (and many of the general public, as well). He’s lost his best friend. His life-long dream has turned into a nightmare. But is what’s happened to him justified? Are the tests accurate enough to guarantee that there’s no other conclusion but that he doped his way to victory on Stage 17?

No, I don’t think so. But I’m not sure that it’s a slam dunk for the Landis side, either. There’s a lot of variables at play here, and some missing information that LNDD hasn’t provided. Some of that may bolster the anti-doping lab’s case and some of that information may bolster Landis’ case. After reading Duckstrap’s analysis, I’m of the opinion that odds are the reported results can’t support a guilty finding.

There’s a lot we don’t know about the case, even though Floyd has posted the lab documentation package for all to see. First and foremost is what the defense strategy really is. I’m wondering if the slide shows have been test runs to see what presentations give the best traction to their arguments, and which arguments work better.

Clearly, it’s a PR effort aimed at shoring up his battered image. I have a sneaking suspicion that there are things in the lab pack that Landis’ team knows about, but haven’t pointed out yet, so as not to show all their cards. If that’s part of the strategy, I sure hope it works.

Landis, himself, notes that this is a war where the winner gets nothing, but the loser loses everything. Which brings me back to a question I’ve asked a number of times: Assuming the ultimate outcome is not a guilty finding, who can or will give back to Floyd Landis everything he’s lost in the last four months? What are the odds that anyone can?