Eight quick thoughts on the Froome data

04 Dec 2015 Posted by
You must have seen it by now – Froome’s data, for want of a better description out of Esquire, has been published, and you can read the article here.
You can cut to the science and leave out the Vayer-focused hang-ups and the vague human interest stuff by reading Shane Stokes’ summary here.
Here are some of my quick thoughts on the actual data, what’s there and what’s not.
 

Top

1.  “The engine was there all along – the strongest part

There are certainly important pieces in there. Perhaps the strongest is the revelation of the 2007 VO2max of 80.2 ml/kg/min. The jump to 84.6 ml/kg/min is therefore not outrageous, either as a relative or absolute change. So when Dr Jeroen Swart says it shows that “the engine was there all along”, the data supports that, at least in so far as being an elite cyclist goes.
That said, I’m not sure anyone actually thought he’d gone from a 65 ml/kg/min to an 85 ml/kg/min, but the similarity is the strength of that aspect of the data.
The actual VO2 in those tests would be 6.06 L/min and 5.92 L/min, produced at a relatively similar peak power output (540W vs 525W).
That figure of 80.2 ml/kg/min would be very high up in the pro peloton. It would suggest a performance potential well above what Froome then got “stuck on” for 4 years, unless he was massively inefficient, or unable to sustain high % of max during that period.
We can rule out the latter, because his sustainable power in 2007 was 420W, or 77.8% of peak power, basically the same as in 2015, when it was 419W/525W = 79.8%. The ability to produce that after 4 hours might be a key change, which lab tests would find difficult to measure. But overall, whatever improvements have happened have come largely from a relative change, attributed to a loss in mass, because the combination of those numbers adds up to the estimated performances in 2015.
 
Top

2.  A peek under the hood, confirming and asking questions

They just don’t account for the intervening 8 years, or the “how” these changes happened. As I wrote before, we’ve looked under the hood of a high performance car, and sure enough, we’ve confirmed it has a big engine, as it did 8 years before. We don’t know much else about it, and nor, most critically for those with doubts (it’s cycling, that should be everyone), we don’t know how that engine went from good to great, just enough to go from a mid-level rider to a dominant Tour champion (5.5W/kg to 6.2W/kg), or why on earth that engine was spluttering so badly for four years after the 2007 test. Apparently its driver had bilharzia, but there’s no measurement, no test, that might support that contention even though the disease affects the blood and a pro cyclist’s blood is as scrutinzed as it gets.
 

Top

3.  Sustainable power

Similarly, the sustainable power hasn’t changed – 420W vs 419W in 2007 and 2015, respectively. The weight loss of 6.6kg between the tests (it would be almost 9kg using the Tour weight adjustments in the article) accounts for the increase in relative power output at these numbers – 5.49 W/kg vs 5.99 W/kg (or 6.25W/kg at TDF mass).
 
The argument is made that at his Tour racing weight (that nebulous number that even Sky didn’t seem to measure – go figure, given how critical it is to this whole argument of his ‘transformation’), that would be 6.25 W/kg, which is pretty much what is done at the front of the race for 20 to 25 min efforts.
 
I’m not 100% sure the conversion works that way – it’s close, but the ability to sustain power outputs while losing significant mass (and 3kg from a very lean 69kg is a lot – remember that Froome drew a lot of attention for looking so skeletal?) is one of the things that has been pointed out as most suspicious.
 
The same argument applies from 2007 to the present – loss of 6.6kg with no change in power.  That’s big, but not inconceivable if a whole lot of things (training adaptations, weight loss, strength gain etc) are happening simultaneously.  Less clear would be the 2011 transformation, but there again, we have this lack of data the makes a picture look fuzzy. At least it’s a picture. Sort of…
 

Top

4.  The missing efficiency

What is missing? The main one, in my opinion, is efficiency. Presumably the VO2 corresponding to those power outputs would be relatively similar (unless something goes ‘wrong’ beyond his threshold power), and it’s a pity that wasn’t disclosed because it would help complete a picture of whether he’s improved his efficiency in addition to his relative VO2 and power output by virtue of losing weight.
The efficiency would also help in and of itself, without being part of this longitudinal comparison. It would’ve been interesting to see that data. Maybe in the journal article to come.
The other missing part is the threshold power output.  It’s mentioned as 419W, but as I alluded to in my preview article yesterday, the way this was approached was always going to be one of the most interesting aspects of the methods, and the resultant interpretation.  This is part of what will emerge when the peer-reviewed paper is published, no doubt, but it’s a pity that in a sport where performance is the outcome of A + B + C (in some combination), the focus was strongly on A, with no mention of B, and very vague discussion around C (the 20 to 40 min issue, see point 5 below)
Just as an illustration of why that all matters, here’s the ‘model’ for how VO2max and efficiency are related at a range of given % of VO2max, also from yesterday’s post.
Screen Shot 2015-12-03 at 6.46.45 PM
 
This does, however, highlight the problem with this order & delay of the reports. It’s not dissimilar to how the Oscar Pistorius case went down at the Court of Arbitration for Sport (I’m not referring to the now infamous murder conviction). There, the data was presented long before peer-review, and by the time that academic process was completed, the picture looked, well, completely different. In that case, the delay was a legal/commercial one, with pressure brought on a sub-group researchers to remain silent, which I hope wasn’t the case here.
 

Top

5.  20 to 40 minutes, because those are the same…

In the absence of the completion of the data with those, and a few other missing puzzle pieces, the data component of the Esquire peak doesn’t contribute all that much. I realize it’s not written for that purpose, but it does, as a result of its pre-occupation with Antoine Vayer and the very circular argument about what is sustainable for 20-40 minutes, create more questions than it answers, engine and body mass aside.
 
That pre-occupation with “20 to 40 minutes” is ridiculous. That range is so large as to be meaningless when discussing one aspect of data. It’s like discussing a 5000m and a 10000m running event and assuming they run at the same pace. They’re close, yes – a 5,000m is run at 61-62s/lap and a 10,000 at 64-65s/lap, but that’s a 5% difference.  And 5%, as Froome himself shows between 2007 and 2015 is enough to transform a journeyman into a legend in professional sport.  So that kind of vagueness doesn’t help.
Against this “20 to 40 minutes” backdrop, the article refers me as saying that anything over 90% of max would put a rider in the suspicious category.  Certainly, 40 minutes at the end of a 4-hour stage?  Yes.  20 minutes?  Not necessarily, it’s very different from 40 minutes, so I don’t get when they talk in such general terms, and it happens often in the Esquire piece. That kind of ‘vague’, ballpark talk can sometimes be an indication of uncertainty, which you do when the data doesn’t quite fit the hole it was intended for.
 

Top

6.  The major tactical oversight

Speaking of Vayer, it’s ironic that he is mentioned so often in it, dismissively as usual, because he is part of the great tactical oversight in this testing. That is, he and Paul Kimmage should have been there if Froome and Sky wanted to shift the perception of this being a PR exercise. That’s just basic tactics, has nothing to do with the science, but it’s kind of obvious and an opportunity missed (the picture below comes from an article I wrote for cyclingweekly on Thursday, which goes into more detail on this)
Screen Shot 2015-12-04 at 3.43.39 AM
Top

7.  Joining the dots

All along, in fact since July, the concept I’ve tried to write about is comparative longitudinal physiology. Space and time. This data release provides some of it, at least in terms of A to B, 2007 to 2015. So that’s a good thing. Unfortunately, it doesn’t come paired with biological data other than a few Hb, Hct and Ret% values, but like the physiology, they’re too far apart and too thin to draw any meaningful questions. What they have done is to ask a few questions about reticulocyte suppression and a potential Hb increase during a Tour de France? However, you can’t draw a picture with three dots – it could be a triangle, a square, a pentagon or an octagon, and herein lies the problem, as it does for the physiology.
 

Top

8.  Biology, physiology and performance

In conclusion, the performance transformation, by virtue of this data, has some physiological basis – maintenance of power output and VO2, co-incident with a drop in body mass (most of it fat). Fair enough. This is not a donkey to a racehorse situation, in terms of physiology, only performance, and that’s no partly explainable, but as seems to always be the case with Sky, only partly.
 
Trouble is the performance transformation came right in the middle of our two “snapshots”, and that leaves a big ‘hole’. I don’t think anyone can deny this. That hole has been filled by bilharzia (and a little bit of asthma), as far as explanations go.
 
The pairing of biological data during this period is what would close many circles. There is no good reason for such data NOT to exist – it’s there, it just hasn’t been provided. If the failure of Froome to turn this massive engine into good performances right up to 2011 (when he was being hammered at the mighty Tour of Poland) was down to that, then the biology would support it.
 
So it’s a real shame there’s nothing from that period, either body weight, or these basic physiological measures, because the self-professed most scientifically advanced cycling team in the world has zero measures of any kind on their most valuable asset. Go figure.
80.2 ml/kg/min is high, but it’s not eye-popping. There are plenty of riders who get there, or even beyond, and never come close to achieving the dominance of Froome in recent years. That says to me, again, that something is missing from this ‘big reveal’. We already know that, though, because VO2max is a poor predictor of performance in an elite sub-set of endurance athletes, runners or cyclists. Knowing this, it would have been nice to explore those. And yeah, I know, it’s only Esquire, but that brings us full circle. The point was to offer data in support, and what’s happened is that some data has been offered, and good for Froome.
 
But overall, it does come across as another point along a line of fuzzy data, half disclosure and semi-transparency. That’s what we’ve come to expect from cycling, and now athletics, and maybe all sport. Perhaps it’s time to embrace the fuzzy.
 
So, interesting data. Some excellent stuff, very fascinating, and as I’ve said before, I have no reason to doubt the actual numbers, or the quality of the testing, because I know the tester, and he is the best in the business.
 
The doubt was never going to be assuaged by numbers though – data whispers, it’s the stuff off the bike that shouts so loudly that it’s hard to hear the data speak sometimes. And Sky does a lot of shouting, as does Michelle, occasionally Froome, and this Esquire piece. But the data, if you listen carefully enough, tells part of the story, so good for him for doing it. It’s just the “how”, as it often is.
 
A nice exercise. Now, I have three days of Talent ID conferences to speak at.
 
Ross