Froome’s scenarios: The data, dilemmas and dark side of the moon

03 Dec 2015 Posted by
Tomorrow sees the release of the first part (or maybe all) of the data on Chris Froome.  I’ve already written a bit on this for Cycling News, which you can read here.

And Shane Stokes also interviewed Dr Mike Puchowicz and me and then wrote this piece, so these summarize much of my thoughts around the concept.  Because they’re mostly in those articles, I’m not going to repeat myself here, but I did want to use this post to add a few things, technical and otherwise, about this whole testing process and what Froome’s data may show.

So below, you’ll find an explanation, as simply as I could state it, of what they will have done, and you’ll find my thoughts on the possible “Froome scenarios”, both for his current results and those historical ones from 2007, which I believe will also be released or discussed in Esquire.




The concept

Cycling performance is the result of three measurable “inputs”. They are:

VO2max – the engine size, or the maximal capacity of the rider to use oxygen. This is determined by the brain, which activates the muscle, which in turns creates an oxygen and blood demand that is met by the heart and lungs, and so all those systems contribute to this measure. Scientists debate the semantics of whether the limit is in the heart or the brain, but it doesn’t really matter for the purposes of describing performance. In theory, the bigger the better.

Efficiency – however, it’s not quite that simple, because efficiency also matters, and it tends to have an inverse relationship with VO2max. The best cyclists are the most efficient, which means that they produce more effective work (power output) for a given input of work (metabolic energy). So you can be an elite cyclist with a ‘relatively’ low VO2max provided you have exceptional efficiency. Ideally, you want a combination of both high VO2max and high efficiency, but that is very rare indeed

Threshold power – this is the ability of the cyclist to sustain a high relative intensity (as a % of their VO2max, usually). The higher the better, and one of the distinguishing factors of elite endurance runners is the ability to stay at very high % of VO2max for a long time. Elite marathon runners, for instance, can sustain 80% of VO2max for the entire race


Behaviour driving inputs creating outputs – the model from A to Z

A lot of people have, in the past, asked about things like willpower, the desire to win, diet and so on. This modeling of cycling performance does not, in any way, discount these factors. The model is as follows:

  1. Behaviors determine the inputs.
  2. The inputs determine the outcome

Behaviours are things like training, diet, recovery, emotional and psychological factors, willpower and so forth. All important, but not discounted by this process in any way, because the sum of these behaviors produce the physiology, which is the set of inputs.

These inputs then go on to determine performance capacity, the outcome. So when a cyclist rides at 6W/kg for say 40 minutes, they’re doing so because their physiology permits them, and because their behaviors created that physiology. To repeat then – behavior determines inputs, inputs produce outcomes.

So, let’s move on.


From the outcome, working backwards – the Froome Scenarios

If we know the outcome, in W/kg, then it’s easy to work backwards and ask what combination of inputs (physiology) would be required. Again, you can insert your paradigm here – some riders have more willpower, greater pain tolerance, belief and so forth, and those factors influence the physiology. So it’s all there, folks, but it’s still measurable as a set of inputs and outcomes.

Below is a graph that shows a range of possibilities. What I’ve done is work off five different scenarios for relative intensity, and then shown you the combination of VO2max and efficiency that would be required to produce 6.1 W/kg (this is the upper limit of performance that we have seen win the Tour de France in the last three years).

I’ll try to break this down as simply as I can.

Let’s start with the first question:

What intensity, as a % of maximum, can we expect an elite cyclist to sustain for 40 minutes in the Tour at the end of a 5 hour stage in the Alps or Pyrenees?

I’ve given you the following scenarios:

75% – the blue line

80% – the red line

85% – the green line

90%% – the purple line

95% – the orange line

froome scenario

In my (informed) opinion, I would say that 85% is the likeliest scenario, given that the rider has done 4 to 5 hours prior to the climb. In fact, I suspect even that is a little high. In a straight up one hour effort, I’d expect about 90% of max, so I actually think something between 80% and 85% (the red and green lines, in other words) would be a realistic assessment of what is possible on a mountain top finish at the end of a Tour de France or any Grand Tour stage.


The “Froome scenarios”: from impossible to likely

Now, for each intensity, there is a combination of the other two inputs, VO2max and efficiency, that would make the performance of 6.1 W/kg possible.

Of course, these scenarios and combinations are themselves sometimes implausible, but for different reasons. Take a look at the graph below, where I’ve provided some shading to highlight the areas that are:

  • Impossible VO2max – grey shading
  • Unlikely efficiency – light blue shading
  • Plausible and likely combinations – orange and red shading, respectively

Screen Shot 2015-12-03 at 6.46.45 PM

So, in grey shading are the areas where the VO2max required is either too high (alien/ridiculous zone) or too low (good recreational cyclist zone). To me, anything above 95ml/kg/min is just not on – there is precedent for it in cross-country skiers and the odd cyclist, but I’m dubious a GT rider will be this high.

And anything below 75 ml/kg/min is too low – a Tour champion cannot be below this, they would require a crazily high efficiency for it to be possible – you can see this possibility as the orange line. To me, these extreme scenarios are not possible and would be untrustworthy.

Next, we have areas where the efficiency is either too high or too low. These are shown by light blue shading, to the left and right extremes of the graph. We know that Lance Armstrong for example had an efficiency of 23.1%, and that’s about the upper limit measured in elites with the exception of one contentious study that I’ve discussed on the site before. I’d be surprised if anyone at the top of the sport produces an efficiency above 24.5%, though I will grant that this is possible, just not likely.

Similarly, anything below 22% would require a very high VO2max and I really don’t think the best cyclists will be this low. Again, it’s possible, I’m just offering my subjective opinion that it’s not all that likely, based on what we know of former Tour winners and current contenders.

Then finally, a plausible area is shown by the orange shading – it spans an efficiency from 22% to 24.5%, and a VO2max from 75 ml/kg/min to 95 ml/kg/min. Those are pretty wide ranges, but if you want to be “conservative”, that’s what you expect.

However, to be a little more ‘adventurous’, I’ve further offered an area that I think is “likely”, and show this as the red shading in the center of the plausible zone. It ranges from 22.5% to 23.8% for efficiency, and from 80 ml/kg/min to 90 ml/kg/min for VO2max.


My prediction – a bit of ‘fun’

My prediction? The star shows it – 23.4% and 87 ml/kg/min on the assumption that Froome can cycle at 85% for 40 min at the end of a mountain TDF stage. To repeat, I think that’s aggressive, but it’s done in favour of the athlete so that you don’t label scenarios as impossible when they might actually happen. Let’s see what the results show, and whether Froome’s data is released in the detail that will allow us to add the real star to my little prediction game (and I’ll hope like heck it’s in my shaded area!)

In fact, let’s do this:  If I am incorrect by more than 3ml/kg/min on the VO2max, I’ll send a good quality bottle of South African wine to the first person who calls me out on Twitter tomorrow.  And, I’ll also send one to David Brailsford, and another to Michelle Froome, signed “Dear DB/MF, yours, the pseudoscientist”.


Playing with the scenarios

If you want to play around with this graph, you could for instance say that he can only be expected to ride at 80% of max, and then you’re on the red line in the graph. That would mean his VO2max would have to be quite a bit higher for the same efficiency (my 23.4% estimate, for instance) – in thise case, it would need to be 92.5 ml/kg/min.

On the other hand, if you think that riding at 90% of VO2max is possible, then you’re on the purple line, and for the efficiency of 23.4%, the required VO2max is 82.2 ml/kg/min.


The question of threshold

On this note, the VO2max and the efficiency are the easier physiological “inputs” to measure. Less simple is the ability to sustain intensity. Broadly, there are two ways to do it:

  • Indirectly, which involves measuring the lactate threshold or other thresholds and then inferring or projecting what might be possible for a 30 min, 40 min, one hour effort based on known/historical comparisons and benchmarks.
  • Directly, which means measuring the physiology of the athlete at 6.1W/kg and making comparisons between these variables (heart rate, lactate, RPE) and what was measured in the max test.

My personal preference would be the latter – I would have had Chris Froome do the max test to assess VO2max and a range of other maximal physiological variables like HR and RPE and lactate, and then done a series of sub-max tests to measure efficiency. Then finally, on separate occasions, I’d have had him ride at a range of power outputs from about 5.8W/kg to 6.2W/kg for at about 20 minutes, and obtained a time-series of how his physiology responds or “copes” at the required performance levels of the Tour. That would allow comparisons backwards and forwards, to see what kind of relative intensity he would be capable of.

I’ll be interested to see how they tackle this one.


Comparative data – the 2007 test and the dilemma it creates

There is a chance that Froome will also release the results of testing that was done on him in 2007, by the UCI. If that were to be linked to the 2015 data, then we’d be in a better position to evaluate the trajectory (or lack of one, depending what it shows), and that will be important. It would be like having two chapters of a book, which provides some context, and possibly alignment, between then and now.

If the data is ‘aligned’ in the sense that his physiology back then was similar to what it is now, then the upside is that his transformation will be revealed as a transformation in performance only, and not performance plus physiology (which inherently looks more suspicious). The downside is that it will ignite more debate about why the performances back then were so poor, not necessarily compared to what they are now, but compared to what his likely exceptional physiology would predict.

The bilharzia explanation will come up, but one might reasonably wonder how it is that the bilharzia didn’t affect that test? They’d then argue that it produces phases of symptoms, and he happened to be tested when he was healthy, but then people would ask why someone with that physiology (assuming it’s at the exceptional end of human physiology as I suspect it will be) was not performing as such at least some of the time.

On the other hand, if the 2007 data is inferior to what we see in 2015, then it helps in terms of explaining how his performances have improved – they’d have tracked the improvements in his physiology. It would however do damage to the perceptions of how the transformation in performance would have occurred, which is where most cynics are focused, because they’ll have a form of causality to point to: “Your performances improved because your physiology was transformed by doping”.

What would help in that regard is the pairing of this data with biological data, as well as more data over time (which, apparently doesn’t exist, as Froome reportedly never did a VO2max test with Sky). However, biological data does exist, both from the passport and what would be regular medical checks by the cyclist. That biological data, especially to interrogate the bilharzia claim, may remain the missing piece in this puzzle, because we’re basically about to get a look under the hood, and we’re either going to see:

  • No change in physiology, but a significant transformation in performance between 2007 and 2015, which brings bilharzia into play as a viable explanation that cannot be substantiated, or’
  • A significant change in physiology and performance, which would invite two theories – one for bilharzia affecting both, and another for doping affecting both.

As is often the case, more questions than answers!
The key words in this whole debate are “longitudinal” and “comparative”. That is, comparisons over time and through space. How has Froome changed over time, and how does he compare to other cyclists? He is one of first to provide this kind of data (Thibault Pinot has provided very detailed data already), and should be commended for an act that others will hopefully follow. However, his biological data is what would provide the context to his physiology and his performance, and unless all three strands are viewable, in parallel, the story will remain incomplete.


A word on the testers

Just something important that I want to say about the testers, or at least one of them. I know Dr Jeroen Swart personally, and I believe him to be one of the best physiologists in the world in cycling, and also someone who is sincere and who has integrity. I remember saying that to our South African Olympic Committee some years ago, trying to get those blind men and women make use of what we have, and I believe it as much, if not more, today.

I don’t know what process was followed, or whether there were conditions that had to be met, but I believe that Jeroen is a trustworthy scientist, to the point that I think the data produced will be beyond reproach. Similarly, I know Ken van Someren who heads up the laboratory, and I’ve no reason to doubt his sincerity either. There is, it must be said, an issue with the EIS link, something Dave Brailsford alluded to earlier this week. That’s the kind of conflict that people will pick up on, because it does mean, however you skin it, that there is potential for a conflict of interests. I’m not a fan of ad hominem arguments, so I’ll wait to see on the data, but it’s something to be borne in mind as a “possible” sideline to the story.


The platform – science vs PR?

And of course, interested to see the testing results in general. Or at least, how they approach them in Esquire. I know this medium has been criticized, and I understand why. I would have no problem, however, if the science is still front and center of Esquire. And, it absolutely has to be followed by a peer-reviewed paper, which should ideally have preceded this, or at least co-incided with it. I’m all for the translation of science, and so provided the piece is true to the scientific process, then I’ve no problem with its dissemination to a wider audience.

However, if we see a “puff” piece, written as a human interest story with the science as an aside, then I’d be less inclined to trust it, and will feel more likely I’m reading (yet another) Sky-PR stunt (there are already aspects of that in this testing procedure).

However, if the science is credible and fully disclosed, with as much detail as possible to leave no questions unanswered, then it’s great.

One final issue is around the timing – we’re not talking advanced testing here, so the delay is hard to explain. There’s so much data already on what he would have been tested for, that really, it could have been written up in a day, with comparisons to other riders. I realize people are busy, but I don’t get this length of delay. This article took the length of a one hour train trip to write, with the graphing and analysis, and while I realize it’s not of the scientific depth or volume of what would be required for a peer-reviewed paper, we’re not talking about the cracking of the human genome here.


Conclusion – get to the dark side of the moon, then decide

To use an illustration I’ve used before, you’re only going to be offered one side of the moon. It’s easy to see the moon on a clear night. What you need to do (not only in this instance, but in any debate) is work really hard, force yourself to get around to the back, to the dark side of the moon. And if that still looks the same, then you’ve seen the whole picture. But as with many things, it’s what you DON’T see that should inform your opinion, and so when this data release hits us tomorrow and then again on Monday, ask what’s not being shown. I hope it is as little as possible.


Finally, if you made it to here, this may not offend your sensibilities as much, but I thought I’d take the opportunity of saying that I appreciate each and every view, and will continue to run this site and try to write articles when my time and energy levels allow it, no matter what.  However, I’m also fully committed to keeping the site ‘pure’ in the sense that I don’t want supplement companies or the like advertising.  The only source of income is thus your support in the form of donations, and so if you feel it worthy, please consider donating at the button on the top right of the page, or click here to donate through Paypal.

Any amount is more than welcome.  I’m sure some of you would like to donate rotten tomatoes and bricks through my window – unfortunately PayPal does not offer this just yet, but thanks anyway!