Are Simple Explanations Always The Best?

For every complex problem there is a solution that is simple, neat and wrong.

H.L. Mencken, quoted at Trust But Verify

For those who have been following the Floyd Landis story, the temptation to accept the simplest explanation is powerful. And the idea that when given two equally likely explanations, the simplest explanation is the best explanation is an old philosophical principle known as Occam’s Razor.

The initial stories about the allegations against Landis painted a picture of solid evidence against him. Testing of his A and B samples after Stage 17 of the 2006 Tour de France indicated that he had traces of synthetic testosterone in his system. Or so it seemed. For casual observers, this was enough to be convinced of his guilt.

Over time, however, it’s turned out that the story is much more complicated. And for a good period of time, the mainstream media seemed to lose interest in the story. Just another doper caught, tried in the media, and sent off to ignominy where he belongs. At least, that’s how it appeared.

With the publication of Michael Hiltzik’s investigative series in the Los Angeles Times a week ago (part 1, part 2), interest in the mainstream media seems to be picking up.

Wayne Coffey, writing for the New York Daily News, produced an article that attempts to strike a balance. Unfortunately, there’s a few clinkers in his story.

First, there’s this:

A source with in-depth knowledge of WADA’s modus operandi and lab protocols disagrees completely with Baker’s metabolite interpretation, and doubts the authenticity of his t/e figures. For one thing, Baker’s interpretation is based on outdated guidlines, the source says.

“If the numbers are what they say they are, the case wouldn’t be going forward. The case is going forward,” the source says.

The only problem with Coffey’s source’s explanation is that Baker’s figures come straight from the LNDD lab documentation package given to both Landis and the anti-doping authorities. The T/E ratios that Baker quotes are the lab’s calculations. So if Coffey’s source is saying that those ratios are wrong, and he or she has the proof to back that up, then what’s really being claimed is that LNDD is holding back the real data.

That seems a bit far-fetched to me. LNDD may be guilty of many things, but I don’t think they would purposely conceal the true evidence from an athlete accused of doping. What purpose would it serve? The bad PR they would get once the truth came out would be huge. The damage to their reputation, beyond repair.

Would the anti-doping authorities purposely withold the real data? Again, if that ever came out in public, the damage to their credibility would be enormous. So Coffey’s source on that assertion seems to be just plain wrong.

Perhaps it’s the anti-doping authorities who are misinterpreting the numbers? No less a person than Dick Pound does so, right in the midst of Coffey’s article, when he’s quoted as saying this:

[Landis] has to find some way to overcome the fact that there is an A and B sample that is up to its eyeballs in testosterone.

That’s an overly simplistic explanation. By going through the lab documentation package, what you’ll find is that Landis’ measured testosterone level was less than one-quarter of the level considered to be high by WADA’s own standards. Landis came in at 45.4 ng/ml versus the 200 ng/ml testosterone concentration that WADA considers to be excessive. What caused Landis’ excessively high T/E ratio was a very low epitestosterone level, not excess testosterone in his system, as Mr. Pound suggests.

And that’s if the lab’s measurements are correct. There are some very qualified scientists that have looked over the lab documentation package and reached the conclusion that all is not as it seems in relation to the data reported by the lab (here and here and here).

Coffey also offers this tidbit:

Two years ago, Spaniard Jesus Manzano, one of the top cyclists in the world, went public with an array of revelations about the pervasiveness of drug use in cycling, saying that “a cyclist couldn’t do the Tour (de France) or Vuelta (Tour of Spain)” without doping.

I suppose anyone who’s raced at the Tour or the Vuelta or the Giro might be considered one of the top cyclists in the world, but Manzano, as best I can tell, is not someone who’s won a large number of major races. To me, the phrase “one of the top cyclists in the world” suggests that he’s at the level of an Armstrong, Basso, Ullrich or Riis. What Manzano is best known for is being the whistleblower whose testimony led to the ban of the Kelme team from the 2004 Tour de France over suspicions of doping by the team’s doctor, Eufemiano Fuentes. Fuentes is also the doctor at the center of this year’s Operacion Puerto scandal.

Coffey uses Manzano to back up the idea that all the top cyclists dope, and to make Floyd Landis appear naive when he says he has no reason to believe any of the top 25 riders in the world dope. Manzano’s allegations may have been true for the Kelme team, but I wonder just how much he knows about cyclists on other teams.

It’s always tempting to look for explanations as to how another athlete could be doing better. And the idea they’re doping is a very easy explanation to latch on to, even if you have no direct proof. Sometimes — very likely most of the time — athletes do better just because. No other reason.

I used to train with one guy who, if certain riders beat him in a race or even in sprints during training rides would head down to the bike shop afterwards and buy the latest, greatest gadgets to make his bike lighter and faster. It never helped. The same riders would still manage to beat him the next time.

Some people, tired of the same kind of situation, might resort to chemically enhanced performance. It might work. Then again, it might not. Perhaps Manzano felt that way, and then justified his own doping by rationalizing that everyone’s doing it. Hard to say.

Coffey makes a valiant attempt at putting out a balanced article, but he needed to do a bit more fact-checking to balance out the assertions made by the anti-doping side. In any event, the bottom line for the Landis case is this: The simple explanation is tempting, but very likely to be wrong. The truth of the matter is much more complicated. Not in terms of what Landis did or didn’t do, but in terms of showing definitively that he did or did not dope. In this case, the simplest expanation may not be the best.