The 2014 Tour: Performance implications & a reflection on the origins

25 Jul 2014 Posted by
Vayer book

Many of you have no doubt followed the performance analysis discussion of cycling, which has picked up momentum in the last three years.  Let me tell you where my part in this story originates…


David Walsh – “Use this”

In June 2009, I was privileged to spend a few days with journalist David Walsh (he of Armstrong-takedown fame, a man who pursued Armstrong despite being shunned, sued and slated from every direction), here in Cape Town. He gave me his copy of this book (see picture above), loosely translated “Can you win the Tour?“, written by @festinaboy and Frederic Portoleau, which documented the power output of the top riders during the mountain climbs of the Tour de France.

David’s words to me then were that this method of performance analysis can be used to help show that Lance Armstrong doped to win 7 Tours de France. The idea was that performances have physiological implications, and at some point these implications become implausible, which leaves doping as the only possible explanation.

Physiologically, this is entirely true – the size of the engine (VO2max), the efficiency of the engine, and the ability of the engine to run near maximal intensities are clearly important factors determining performance, and so without bending the rules too much, we can, based on known physiology, construct a reasonable set of assumptions for each and therefore predict performance and its limits.

It’s not pseudoscience, though it does require some assumptions, all of which I readily acknowledged from the very first day when I analyzed performance, that of Alberto Contador on the Verbier in 2009 (only a month after meeting and talking with David Walsh)


Five years on – a new situation, but kind of the same

Here we are 5 years later, doing the same analysis, but improved vastly by people like @ammattipyoraily and @veloclinic, who’ve joined with @festinaboy to improve the way we analyze performances, with the intention of finding confidence in the performances (at least this is my desire, not to catch cheats but to find ‘qualified hope’).  

We now have historical records, which have been used to produce benchmarks and expected times, corrected for altitude and gradient, applied to many more riders than just the winners – this is real progress, thanks to this group of ‘hobbyists’, collectively called the “pseudoscientists” (hat tip, Sir Dave).

Of course, the reality is (as always) far more complex than the theory, and we all recognize this.  Single performances lead to ‘performance pixelation’, the cure for which is a dose of common sense.  And we understand that we cannot judge and condemn a rider to the doping doldrums based on an estimate of W/kg alone. And while we may differ in our approach, the concept underscoring the method remains as true today as it was, in theory, in 1998, in 2005, and in 2009 when David Walsh introduced me to the concept.

Yet I read how the method is now ‘pseudoscience’, and dismissed as ineffective, presumably because riders other than Lance are winning, and I wonder, “what am I missing here?”  Why was it valid years before, when Armstrong was the target, but when it was Froome or Wiggins, too many assumptions had to be made?  What’s the difference, other than the flag beneath which a given rider competes?

The attack is strongest from within the UK.  Even Walsh, who quite literally put me onto the idea, has dismissed it.  What worked for Lance, and for Contador and Rasmussen, is not applicable to Sky or Froome.  That’s the hypocrisy of patriotism, I guess.  I guarantee you that when Spanish, Colombian or Italian riders win the Tour, with record performances and 6.1 W/kg climbs that challenge physiological limits, many in the UK will be first to cry dope.  Anyway…

I also read that the doping discussion has NOT happened in 2014 (again, from David Walsh), which is bemusing, because we’ve had the same discussion this year as last year.  In fact, I could have posted the exact same articles from last year and just changed “Froome” to “Nibali” and left it there.  The reasons I’ve been more silent now are a) work pressure and b) the realization that doing a stage-by-stage analysis invites people who don’t quite understand the method to make out-of-context comments about it.  I wish to actively avoid the pixelation of the past.


The 2014 Tour – a picture much like that of 2013

I hope, in the next day or so, to find the time and energy to put down more comprehensive thoughts on all the mountain stages of the 2014 Tour. They will show that it has, by and large, been quite similar to the Tour of 2013 in terms of both the pattern of the race in the mountains and the magnitude of performances and victory by the top rider.  If anything, it is cause for even more optimism, because 2014 has been slower – the power outputs produced by Vincenzo Nibali are consistent with those produced by Froome and Quintana in 2013, and possibly even lower (but again, direct comparisons invite all kinds of rebuttal, so why make them other than to inform a common sense discussion?”)

My expanded thoughts (and these brief ones) again confirm my belief, held since 2010, that the sport is likely cleaner, but not yet where it wants to be, clean.  The internal optimist in me takes heart, for example, from the fact that the performances of the 1990s would have dominated Nibali in the same way that he dominated the rest of this year’s race, and in the same way that Froome/Sky dominated last year’s (the similarity I mentioned).  Nibali, for instance, won today’s stage to Hautacam by 500m. In the height of doping in the 1990s, he’d have been close to 1km down on the fastest climbers.  There’s much cause for hope in that finding alone, as well as the fact that 26 men have climbed Hautacam faster than Nibali today.

However, the cynic remains (justifiably) alive and sadly viable, because the performances we’ve seen this year, and in 2013, are in good company with those of known dopers, perhaps not dating back to the sport’s worst era in the 1990s and early 2000s, but certainly the latter part of the 2000s.   While Nibali and Froome of 2014 and 2013 are relatively similar, they are also similar to the 2003 – 2005 era, which provides some reason for skepticism.

Nibali’s ride of Hautacam today ranks #27 of all-time, with those ahead and behind sharing doping as their common thread.  You can go down that list and cross off all the riders implicated in doping, either by testing or confession, and you’re left with a very thin list indeed.  And while a rider cannot and should not ever be guilty by association, it also takes a special kind of denialism to ignore this fact.

Further, the presence of some of those dopers in the management of the sport of cycling sadly pushes one to lean towards a cynical interpretation of data, however badly hope wants to peek through.  Given a single number – power output on a climb – there are many possible interpretations.  But when the dominant riders of today are managed and coached by unrepentant dopers, it’s difficult to ignore the cynical option.  Someone needs to tell me why I should expect a different outcome now?

So we leave ourselves with a conundrum, a series of unanswerable questions. I cannot answer them, but will attempt to shed some light on them as soon as I can, on the website, so bear with me, at least until early next week (I hope sooner). What we cannot be left with, however, is the idea that this method, so ‘valid’ in 2009 to prove the Armstrong lie, is suddenly inaccurate, useless and irrelevant, just because we want it to be.  That is bewildering to me – that David Walsh, who so convincingly promoted a tool (that’s all it is) for one rider can dismiss it so easily for another.

The data can always be interpreted differently (and yes, some of us differ on the extreme position we are prepared to adopt), but it remains the same data, and the picture emerging is no different now than it was a year ago – it’s ambiguous, but worthy of interrogation. The allegation that we are ‘silent’ on these issue now, particularly coming from a man who catalysed my own interest in performance tracking, is hardly what the sport needs.

It needs neither cynicism nor naivety, but rather informed discussion. I hope this can be encouraged.

I hope to make my more well thought out contribution to it tomorrow.  Until then, good night.


This post is part of the following threads: Tour de France Analysis, Tour de France timeline – ongoing stories on this site. View the thread timelines for more context on this post.