Revisiting The Mysteries Of Andrus Veerpalu
With somewhat cryptic reports appearing that Andrus Veerpalu failed a drug test back in January (he retired just before World Championships this year), it’s probably a good idea to revisit my old post on the subject. In that post I considered the conventional wisdom about Veerpalu that he had an uncanny ability to pop outstanding races in big championship events. When I looked at all of his results as measured simply by rank, my basic conclusion was that he sometimes seemed to do this, but the difference wasn’t huge, and I could find some examples of skiers that were more extreme instances of people over-performing at major championships. In the end, I thought his record was ambiguous, but I could see why people would say this about him.
In this post, I think I’m going to end up backing off of my skepticism a little bit, but not because of an unconfirmed drug positive. Rather, if we acknowledge that Veerpalu was an extreme classic specialist, so much so that over his entire career he’s done 98 classic races and only 16 freestyle once (and 19 pursuits in various formats), and use a somewhat more sophisticated measure of performance, things begin to look a little different. If we focus in on just his classic races, plotting them by standardized percent back from the median skier we get the following:
I’ve highlighted the OWG and WSC races in red. First, let’s acknowledge that the truly unusual race from 2006 was plausibly the result of infamously tricky weather conditions.
So based on this I’d probably grant you that in classic races, for the period of time between 1998 and perhaps 2003, Veerpalu did display an uncanny ability to ski his best at OWG or WSC races. But since then he’s been declining slowly and still occasionally pops a good race, but not so uniformly at major championships.
I’m skeptical that we’ll know anything for sure in the near future about these developments. Veerpalu is sure to deny everything and if FIS hasn’t said anything about it by now, I sort of wonder what the truth of the matter is, although the FasterSkier article I linked to suggests that the B samples hasn’t even been tested yet.
In any case, I thought this might be a timely clarification of my earlier post.