Long live transparency – the data video controversy

14 Jul 2015 Posted by
The first rest day of the Tour de France produced, as usual, some off-the-bike stories, but in 6 years of covering the race, I can’t recall such a bizarre sequence of events as this one.

About five days ago, Antoine Vayer, he who provoked Brailsford into calling performance analysis “pseudoscience”, started to post graphs on Twitter showing Froome’s power output and heart rate on his climb of Ventoux in 2013. They include an attack, off a base of long periods of 400W, up to 1000W, and maintaining 600W during attacks. During this period, his heart rate barely changes – 157 pre-attack, 161 post.

Next, someone industrious managed to take a file of the raw data and synchronize it to video footage of the ride, so that you could see, second by second, how speed, HR, altitude and power changed. It was fascinating to watch – in fact, it’s a model for how the broadcast should look.

I’d love to provide you with the link to this video, but Sky’s lawyers have had it removed, and the person who did this no longer has a Twitter account (whose doing, not sure – initial reports said it was legal action and suspended), which I find an absolutely extraordinary response. The data by itself didn’t mean all that much, and if it was fabricated, just say so. Or heaven forbid, use the opportunity to explain and gain some points for the now extinct concept of transparency in the sport.

There are many possible explanations for the dislinkage between power output and HR, by the way. He may already have been very close to his maximal heart rate (but then you have to say, that’s a mighty fine effort to ride Ventoux at 95% of max in the third week of a Grand Tour, and also to attack with a 250W increase off a maximum heart rate, four times, from a base of 400W. Not bad). Perhaps the HR was faulty, maybe it’s irrelevant or typical for an elite rider attacking to peak power off threshold power. It’s probably not worth overplaying.

The power output data is really interesting though – back then, we estimated the power on that climb to be 389W (or 388W using Dr Ferrari’s method), and I can assure you, having seen the raw file with second-by-second data (it’s doing the rounds), that the estimate is well, very exact. I’m hesitant to post exact figures because Sky’s legal team had a person’s entire account removed, even though this data is all on the internet already.


Secrecy breeds skepticism

The response has been amazing. As in 2013, the data was first dismissed as fabricated. Then as hacked (which is a tacit acknowledgement of its validity). Or maybe irrelevant. The usual attacks that it proves nothing came, to which I’d respond by saying “welcome to a six-year long conversation, take a side-order of context with that indignation”. Fact is, we’ve been here before, and it’s the reaction more than the revelation that is so amazing.

While on the subject of power data, I cannot understand why teams believe they should keep it secret to the extent that Sky seem to. It is analogous to Justin Gatlin asking for the clock to be covered up for his 100m races – that way nobody will know that beating him will take a 9.74s or faster. But they know this already – all the pro teams know what the power output requirement for 10 min, 20 min, 45 min, 60 min is. Transparency would hurt nobody, except those who need performance not to be known (Hey Justin, that’s a free tip – ask them to hide the clock so nobody knows how fast you’re going).

Finally, back in 2013, it was a remarkable performance, and putting it to concrete numbers is probably unlikely to sway you from whatever you believed (or want to believe) to begin with. I wish the opaque curtain of PR & legal action could be lifted. Today was not a good day for winning the minds of a watching public, or for the believability of the sport.

Long live transparency.


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.