Login to VeloClub|Not a member?  Sign up now.
September 24, 2017
September 23, 2017
September 22, 2017
September 21, 2017
  • An interesting addendum here from Dr Stephen Lane:

    One interesting observation from the GSK lab testing was that the combination of Froome’s Osymetric chainrings and Stages power meter reported an almost identical output to that of the ergometer used to control power during the ramp test.

    You might remember that Team Sky adjusted Froome’s stage 10 data from the 2015 Tour de France before release, to account for the often observed ‘over-reporting’ of power when using Osymetric chainrings. The Stages website supports this move, suggesting a discrepancy in the range of 4 to 5%. Furthermore it has been suggested that non-round chainrings can reduce oxygen consumption during submaximal cycling. This could potentially explain Froome’s observed superior cycling efficiency.

    • jules

      I’ve never really bought into Sky’s argument about adjusting power. due to sampling rates, and by definition the lack of sampling in between the sampling points, there is always a degree of interpolation of power measurements. but it’s not immediately apparent to me why the chainrings would throw that out of whack. the ‘source of truth’ is the force delivered through the crank arm and rotational velocity of the cranks. I don’t see the oval chainrings interfering with those measurements much, or at all.

      also Froome’s lactate threshold of 430w is insane.

      • I guess this is digressing from the main impetus of the article (which is well written – thanks Dr Lane), but I thought the issue with measuring power on o-sym chainrings is that the velocity of the crankarms changes throughout one stroke. This has been known for many years (at least 10). See this thread by Coggan and others https://groups.google.com/forum/#!msg/wattage/5prBUKY20s0/npZID_tb-5AJ

        • jules

          members only Fab!

      • Alex Simmons

        Non-circular chainrings violate the assumption made by crank and spider based power meters of constant rotational crank velocity during a pedal stroke.

        The more eccentric the shape, the more the crank velocity varies during the pedal stroke. While torque is sampled at high frequency and then averaged per crank revolution, crank rotational velocity is not measured with such frequency and instead only as an average for the entire pedal stoke. As a result of the incorrect assumption of constant rotational crank velocity, the power phases of the pedal stroke are over weighted in the calculation of power, and the weaker phases of the stroke are underweighted.

        Essentially the error comes from the assumption made by power meters that for each completed pedal stroke:
        Integral of (? × ?) = Average of (?) × Average of (?)

        That holds true for the most part with circular chain rings (although not always) but is always violated with eccentric chain rings.

        i go into much more detail in this post:

        and DJ Connell does a good job on these rings here:

        • jules

          hmm OK. I’m with you I think!

        • Luke Bartlett

          power is watts, which is joules/second, which is force times distance per second. distance would be circumferential distance the pedal has rotated during that force sample. this is deduced from cadence. cadence is approx constant when cycling circles. cadence goes up and down when pedaling non-round chainrings. to properly measure you would need to measure the rotational speed of the cranks at each force sample rather than time each revolution through a cadence sensor type measurement.

          • Alex Simmons

            I already said that with this statement that non circular chain rings violate the following assumption made by power meters:
            Integral of (? × ?) = Average of (?) × Average of (?)
            ? is torque
            ? is rotational velocity

          • Paul Jakma

            The problem is two-fold:

            1. Force is not constant over the pedal stroke – it varies greatly. Therefore, you must sample it at sufficient resolution to be able to fairly approximate the force. (So at twice the rate of the highest frequency component in whatever function adequately describes the force against pedal strokes).

            2. The force is sampled (easily enough) in the time-domain. However, this time-domain signal must be converted to the distance (or angular velocity domain). And _here_ is where Alex says the problem lies, in that he says powermeters are not sampling crank rotation fast enough to determine this.

            Certainly, most (all?) used to just /assume/ that cadence was pretty much uniform, and hence that the time-domain to distance-domain conversion was a trivial scalar function. However, with oval rings this is not true.

            Further, it used to be the case that powermeters couldn’t even measure cadence in anything but units of whole revolutions, as they used simple magnetic field sensors passing over a fixed magnet each revolution.

            However, solid-state accelerometer chips have been cheap as chips for over a decade now, and powermeters started incorporating them long ago. The intrinsic “only in whole revolutions” resolution that the traditional magnet cadence sensors were limited to is not there with accelerometer measurement. They can be noisy though. However, solid-state accelerometers and (perhaps more importantly) the signal-processing software to condition their measurements have improved greatly over the years. A good number of powermeters have even ditched the magnetic sensors and have been calculating cadence solely from accelerometer data. You can also buy magnet-free ANT+ cadence sensors that strap to crank arms.

            What is unknown is exactly how accurate accelerometer cadence can be to below the sub-revolution level.

            Certainly, there are powermeter companies out there – large and leading ones in nations with class-action laws – who say the software on their powermeters can self-calibrate in the field for non-round chainrings and maintain the advertised level of accuracy power. (Note: that could be a calibration step to work out the eccentricity of the rings and calculate correction factors for later use in the time->distance domain (or analogue) transform – doesn’t have to be sampling cadence fast in normal use).

            That’s the issue. It’d be nice to be able to settle it for sure for the various brands out there.

        • Paul Jakma

          I spoke to a power-meter engineer from a very well-known US manufacturer about this the other year. He indicated it had been an issue in their older power-meters, but had been accounted for, for a number of years. He didn’t say how exactly, but I believe it is a software calibration feature.

          Googling, it seems that feature was released in about 2012 to 2013.

          • Alex Simmons

            Yet when we test them the claim that accuracy isn’t affected guess what happens?

            In a nutshell, I don’t/won’t believe them until they provide the data to demonstrate otherwise. One other (euro) PM company that uses the same sensor methods made the same accuracy not affected claim to me but then refused to provide the crank velocity data to prove it. They couldn’t back up their claim. Stick the eccentric rings on their meter and test them and guess what happens? Yup, the data is affected. Stages had trouble getting reliable cadence data out of their accelerometer signal, let alone reliable multiple velocity readings per rev. But it’s a moot point in any case with a unilateral power meter.

            If meters really are measuring rotational velocity more than once per revolution then they should be providing the data. But they don’t. I’m certainly open to seeing the data and having them prove otherwise, and I’ll shout it from the rooftops. What I don’t get is why would they not provide such data? It would be leap above others in power meter measurement performance to provide it.

            Also, it’s not just these rings which introduce such CV variation, there are other use cases with regular circular rings where it can happen as well (e.g. high power but low inertia scenarios). Not quite to the same extent as with eccentric rings but it’s measurable.

            If you want to stress test any power meter in this regard, then simply perform some peak force-velocity testing. IOW do some maximal effort standing starts and plot the average effective pedal force v crank pedal velocity for the initial 5-6 seconds of the effort. If the data can’t reliably match the well established inverse linear relationship that exists before neuromuscular fatigue sets in (which happens after ~6 seconds give or take), then it tells you the meter doesn’t handle variable crank velocity scenarios as well as they might have you believe. Quite a few meters fail this simple sniff test.

            • Paul Jakma

              Is your position based on scepticism, that because in the past essentially powermeter vendor had accounted for oval/eccentric chain-rings, and that since /one/ company you spoke to who claimed to have fixed it apparently still had issue, that therefore you now will not believe /any/ now who say they have?

              That is fair enough. I just want to work out at what level there is data to contradict the claims of vendors for a number of modern powermeters that they handle non-round – versus what is (justifiable) scepticism of vendor claims though without specific contradicting data.

              • Alex Simmons

                Well I’m not the one making a claim of improved power meter performance (e.g. wrt non-circular chainrings) and *not* backing it up with the data/evidence. That’s up to power meter companies to do, especially those making such a claim.

                When two of them were put to the test, their claim was found wanting and they could not provide the data to back up their claim. When another was put to the test were not even getting cadence data right, let alone crank velocity measurement at higher sample rates. What’s worse in the latter case is they publicly attacked the scientist that demonstrated they had a problem. Eventually they acknowledged the problem and worked to improve their cadence data.

                So if I have a level of skepticism, I consider it justified based on experience and principles of how such meters work and the lack of data to back up such claims. Like I said, if I am wrong, or more to the point, if a manufacturer has really cracked this nut, then why don’t they simply publish the crank velocity data to prove it? They must already have it if their meters are not being affected and it would be a most trivial thing to do and it would be valuable data and demonstrate a clear lead over rival brands. And if they cannot or will not, then I will maintain my skepticism. Trust but verify.

    • Alex Simmons

      I actually asked Jeroen Swart about the power measurement being the same (since it shouldn’t be if both measurement devices are accurate) but didn’t hear back (it’s no big deal). I would take the crank data with a grain or three of salt given it’s not a tested production power meter and it was not an item the lab owns nor calibrates. If it was a publicly available Stages unit then the unilateral crank measurement renders it pretty useless for the sort of accuracy you’d need for such an assessment. IOW either way I wouldn’t place too much weight on the crank power meter data in this instance.

      About all you can say about the comparison of the Stages data with the Computrainer they used is that it demonstrated there was no motor used during the test ;)

      One of the problems with using CT data is the measured gross efficiency value includes a component of equipment drivetrain losses. You really should be measuring the power output at the crank since the energy out / energy in equation is then measuring the athlete only and not inclusive of drivetrain losses (which will be ~2-3%).

    • J Evans

      If the osymetric chainrings are better, why doesn’t everyone use them?
      And why did Wiggins change back to round?

      • J Evans

        This is a completely genuine question.

        Also, if there is a difference, they should be banned. In my opinion, all bikes should be the same so that it’s all about the riders.

        I suspect there’s little or no difference or everyone would soon ‘get used to them’ and be using them.

      • MD75

        Simply put: Are they really better, no 1 answer has been seen that is 100% tested. However, much like a lucky item someone feels helps them. Be it sport or anything in life. If YOU feel like it is helping you. Your more prone to get better results with it than without. In a sport that is. US football for example, a player feels they play better with a lucky pair of cleats or gloves. And maybe they do play at a slightly higher level when they are wearing them vs. without. The fact is, shoes and gloves aren’t making them better. The mind is what is doing it. Are they really 1% better…if they really believe it helps, yes. The mind is very powerful to suggestion.

        As for Wiggins, maybe he only used them because Sky requested it. Or he was getting some money on the side for using them. Froome is sponsored by Shimano (all team Sky is). Yet they cover the logo/markings of the rings. So, if he really thinks that they are giving him an additional amount of power, Than they might be, the exact numbers of how much is the question. 1%, 5%, maybe more. Bottom line is that if you think it enough and believe in it. You may just get a slight increase with them. I know I got a 10min drop in my best climb time with them over round rinds. I swapped back and still have a 3min increase over my best with round rings. Indicating that they must have helped me some.

        • J Evans

          Surely some tests have been done on these: I’m surprised that I’ve never read any empirical evidence anywhere.

  • Stephen J Schilling

    The numbers are, indeed, “pas normal.”

    • Warwick

      Any performance by any person at the top of their sport is very far from “normal”, it doesn’t prove or disprove that they are on drugs, “normal” people don’t win the Tour de France / London marathon etc etc etc

  • Darren Yearsley

    Of course the numbers are what are required to win the tour, because, well, he won the tour. The real question is how did he get to those numbers where previously he could not.

    • jules

      in my lay view, the study is useless for precisely that reason. if he produced numbers that were below those quoted here, people would be even more suspicious.

      • JBS

        Not completely useless. It does give a baseline to measure other winners against in the future.

        But in terms of trying to prove/disprove doping, yep useless.

  • My little brother holds the DEXA scan record at RMIT. 4.0%.

    • #lean

      • VERY strong too.

    • Esthermpollack1

      <<fb. ?????????????????????????????????????????????????????::::::!il836r:….,….

  • RayG

    No way does Chris Froome have “a more efficient pedalling technique”

    • JasonM

      It is not really about technique. It is a calculation of oxygen consumption and translating that to how much of that oxygen usage results in mechanical output (watts). It is far more about biochemical process than how they ‘look’ while pedaling.

      • Alex Simmons

        Yes, it’s predominantly a biochemical issue. Trained cyclists typically fall in the range of ~19-24% efficiency. Of course there are a few difference types of efficiency measure so it can be a little confusing.

        • However biomechanics will influence biochemistry. Can’t not.

      • RayG

        1. It was a joke. 2. The story clearly differentiates between biomechanical (pedalling technique) and biochemical issues and says both can contribute to overall efficiency.

  • jules

    this study is a good start, but what they need to do is longitudinal studies of riders’ parameters – i.e. under the bio-passport.

    it’s difficult to assess an athlete as legit or not from a set of numbers taken at one point in time. what is more telling is how their parameters – e.g. haematocrit, reticulocytes, track over a grand tour. some riders bounce back fabulously after the rest day.

    of course that info is too revealing so we’ll never get it. which speaks a lot on its own.

    • Nomad

      Good point. IMO, the situation with Froome would be transparent if we had complete and unbiased analysis of his ABP profile from the last 2 years. But how likely is that?

      Interestingly, Horner was under strong suspicion when he became the oldest rider to ever win a GT when he won the 2013 Vuelta. In response to the skepticism, he released bio passport profile from 2009 through the 13 Vuelta.

  • Callum Dwyer

    Is the second table correct? 4 W/kg from 1 year of three 1hr training sessions a weeks.

    • Mark Blackwell

      The table is a bit misleading because on its own it suggests that the difference between world class cyclists and the rest of us “trained cyclists” is simply training frequency, volume and background. Sad fact is, no matter how hard I train, I’m never going to push those numbers… and Chris Froome could probably spend his whole year lying on the couch (drinking coke, eating twinkies) and still beat me.

    • jules

      4 w/kg isn’t that high as a maximum reading from a graded exercise test. if you’re 80 kg, that’s bowing out at only 320w.

      it’s not the same as w/kg held all the way up a 40 minute climb.

      • Callum Dwyer

        OK, so it’s not your typical 20 minute FTP.

        • rosscado

          The table refers to power at aerobic capacity so it’s closer to your 5 min power (but technically the test protocol is a ramp test).

  • Gerard Grant

    You are intentionally disingenuous with your “At 6.5 W/kg, Froome’s threshold power is off the chart.” Stick to the facts. The two weights you have for sure are 70kg and 67kg which place him firmly on the chart. In addition, it seems you have one test which he performed in the lab with a set of figures. All you have for the TDF weight is the weight, not any performance stats for that 3 week period, for all you know he could have been a “?%” off (or indeed better). Or are the figures against 67kg and 66kg from Sky? What we know for sure is 6.1w/kg. We can take the same results and overlay them on his self reported weight and get 6.4w/kg assuming he wasn’t being flippant and really meant 67.45kg! After that we have a void where the other TDF Top 10 weights/power and superhuman efforts reside awaiting the light of day.

    • jules

      it’s clear from the study that they only have reported weights, apart from the 70kg when he did the test. it’s logical to believe his w/kg was higher while he was in peak, race condition.

      from memory, Sky have long said these tests are inherently flawed, I believe for the same sorts of reasons you quote above. but the study author can only use what data they’ve got.

      • Gerard Grant

        No it’s completely unclear where the ‘reported’ comes from. Cycling News has “At his 2015 Tour weight, Froome’s watts-per-kilogram would be 6.25w/kg”. People keep using the word “reportedly” but not, as in the test, dating or referencing that report in some way. You can say what his weight was and I can report that. Not much use though! lol

        • jules

          I appreciate your point. I’m just saying Sky aren’t falling over themselves to supply this data. Which is unsurprising – no one else is either. The study and this article authors’ hands are tied.

        • Steve

          I was under the assumption Froome gave them the weight from what he had put down as his weight at the beginning of the tour? so its a “reported” weight as it probably fluctuated through out

    • To be clear, I added the caption on that chart, not Dr Lane.

      • Gerard Grant

        Lol, That’s a bit naughty. Keeps the sensationalism high for no real reason. With his lab figures only – he’s going to podium easily, in a race with long hills/endurance and TT’s, every time he’s fit and doesn’t fall over, apart from that there’s no story.

      • Gerard Grant

        Lol, That’s a bit naughty. Keeps the sensationalism high for no real reason. With his lab figures only – he’s going to podium easily, in a race with long hills/endurance and TT’s, every time he’s fit and doesn’t fall over, apart from that there’s no story.

  • Eat More Lard

    This is all very insightful but what we really want to know is how many watts he is putting out running up Ventoux and are their (not so) marginal gains to be had in discarding the shoes and doing a Zola Budd?

    • ZigaK

      Abebe Bikila would be more appropriate?

  • Roger That

    I am still none the wiser to why he looks at his stem all the time.

    • Stewie Griffin

      it’s to help open the lungs, if you look down instead of up, it’s easier to breathe. So a more efficient way to oxygene consumption. This is the general assumption on the why he does that

      • J Evans

        That’s the general assumption? We don’t think he’s staring at his powermeter then? Odd, I’m sure I once heard it said that he claimed that looking forwards hurt his neck.

        • Stewie Griffin

          well, the assumption that it opens up the lungs was made by several belgian physio experts. Seen that they’re all about marginal gains, perhaps there might be more to it. It could also be just his style, just like Escartin was crooked on his back hence his nickname, “the crab” :)

      • I reckon its postural fatigue management.

    • Mark Blackwell

      I realise you’re making an amusing point, rather than really asking a question… but I can’t help myself: he says that it helps him breath better, and that it’s more comfortable for hard efforts.


      • Roger That

        Thanks, good to know! Ok, but how about the elbows ; )

        • Michael K

          Glued on little rear view mirrors to see his rivals behind him.

        • velocite

          I tried riding with my elbows out one day and it seemed to open up my breathing dramatically. I don’t imagine looking at your stem would do any such thing!

        • Michael

          His elbow position is simply a result of the way he chooses to grip the hoods.

  • Alex Simmons

    One other comment I’d make is the choice of 4mmol/l blood lactate concentration for the sustainable aerobic power measurement. 4mmol/l is a pretty arbitrary value, and in reality maximal lactate steady state (an intensity that equates well with Functional Threshold Power and is typically sustainable for ~40-70 minutes) can be lower or higher than that in any individual. A range from 3mmol/l to 8mmol/l for MLSS has been shown in various studies. We don’t actually know what Froome’s MLSS is.

    So why didn’t they test for MLSS instead of the 4mmol/l power value? Well that would have been a time constraint as MLSS testing requires multiple long duration tests and one can’t do two (very hard) incremental tests to exhaustion and MLSS x 2 f(or different environmental conditions) in the one sitting.

    For interest I did the numbers on the Gross Efficiency, VO2max and power at 4mmol/l equates to a fraction utilisation of VO2max of ~86-87%. That’s a pretty “ordinary” value for a well trained cyclist, and closer to 90% is certainly quite feasible for elite cyclists. It suggest to me that Froome’s actual threshold is quite possibly a bit high than the values shown in this paper (I’ve read the full published paper).

    • Valiant Abello

      You should comment more often. Thanks for the insight and intelligent opinion.

      • Vickijsherman

        <<fb. ?????????????????????????????????????????????????????::::::!il213r:….,…..

    • jakub

      He claims in his autobiography that his maximum 30-minute power, measured on Col de la Madone shortly before 2013 TdF was 459 W. It is usually assumed that your 20-minute maximum equates to 95% of your FTP – so something around 440 W is definitely possible.

  • Marc

    He weighed in at 70 kg and had 9.5% body fat. Then I stopped reading. Try fool someone else.

  • Adam Aldous

    Man…. remember when cyclists had integrity (or we thought they did) and didn’t have to go through this bullshit.

    • J Evans

      No, speaking personally, I never thought that they did.

  • Andrew Coggan

    Just an FYI: the chart in our book is version 4, whereas the most recent is version 8:


    Froome’s estimated sustainable power would therefore best be described as “close to the top of the chart.”

    (Also note that the power profiling tables have been superceded by power-duration profiling found in WKO4.)

  • Berne Shaw

    Add to these findings the fact he was eventually diagnosed with a major chronic debilitating illness and as a result of that was spared having his body used up by a larger early career of numerous tours and races. Once cured he has that advantage ie a younger less used body. Taken together we see his true abilities and potential.

    It follows logically he was not capable of these results whilst being ill. Once he was diagnosed he showed a normal progressive pattern of healing and a normal step wise pattern of improved capacity towards his current health and training.

  • brucegr

    I think we need to be mindful of power analysis. As an example, it is elevated watts/kg that set off alarm bells.
    If a cyclist wanted to mask watts/kg, intentionally make it lower, the easiest way is always report higher body weights. Froome’s weight was never recorded by an independent source during TdF. And duing the GSK tests, bodyweight is easily increased by taking more than usual water/food. THis will increase body weight and lean tissue readings, and thereby lower bodyfat %. When projecting back to the race, it will be presumed the rider had lower bodyfat % than actual. And ‘the gold standard’ DXA xray analysis, is not understood by most. It still relies on a lot of linear regression equations of population norms re height, weight, etc. This introduces a lot of inaccuracy.

  • jakub

    What I also find very interesting is Froome’s heart rate response to exercise. He is known to have a relatively low maximum HR, but even more interesting is that his HR at 250 W was around 120 bpm, while at 420 W it was 140 bpm, a difference of mere 20 bpm. This suggests a very efficient heart and massive stroke volume. He is also quoted of having resting heart rate as low as 29 bpm.

    • Berne Shaw

      So true. Look at two others Greg Lemond and Miguel Indurain. Very efficient hearts. Passive lung function. And prodigious VO2 max 93 to 95 for Greg and Indurain 89. These guys all above 149 lbs as well so not pure climbers. Few understand the physics of climbing in a tour. They weigh more so they do markedly more work than pure small climbers. Thus their efficiency recovery and capacity is must be significant especially as they reduce long flat TT and add more mountain stages.

      With this in mind Froome is a super elite athlete.

    • Nomad

      It sounds like his cardiovascular system was very similar to that of “Big Mig,” who had a resting heart rate at 28 and a VO2max of 88 (though I believe Indurain had a top end MHR in the upper 190s):


      Of course, on the exterior these two looked completely opposite with Big Mig going 6-2 176+ (the kind of size you see with some D1 college wide receivers).


Pin It on Pinterest

September 24, 2017
September 23, 2017
September 22, 2017
September 21, 2017