Welcome to the Science of Sport, where we bring you the second, third, and fourth level of analysis you will not find anywhere else.

Be it doping in sport, hot topics like Caster Semenya or Oscar Pistorius, or the dehydration myth, we try to translate the science behind sports and sports performance.

Consider a donation if you like what you see here!

Did you know?
We published The Runner's Body in May 2009. With an average 4.4/5 stars on Amazon.com, it has been receiving positive reviews from runners and non-runners alike.

Available for the Kindle and also in the traditional paper back. It will make a great gift for the runners you know, and helps support our work here on The Science of Sport.

Monday, July 20, 2009

Tour 2009: Contador climb

The anatomy of a climb: Contador on the Verbier - its place in Tour climbing "history"

News Update: Based on Contador's amazing climb of the Verbier, a french scientist has estimated Contador's VO2 max to be 99.5 ml/kg/min. This was the catalyst for accusations of doping from numerous commentators, including Greg Lemond. However, is this estimation accurate? Does Contador really have a VO2max of 99.5 ml/kg/min, and can we say he might be doping? For discussion on these questions, read our latest post here.

Well, the response to the post yesterday describing Contador's climb has been overwhelming - being in SA, I slept through most of your emails and woke this morning to a deluge of comments and analysis of the climb. Thank you to everyone for your input. It would be wonderful to do a more detailed breakdown of the climb, but that would require accurate information regarding the profile and the specific gradients per kilometer.

Unfortunately, there is little chance of consensus on the actual figures for the climb. Many of you have commented in the last post and referred me to profiles, breakdowns, specific kilometer-by-kilometer gradients, and thank you very much for that. However, not a single person's input agrees with anyone else's!

For example, on the Tour website, the climb is listed as 8.7km at 7.5%. However, the profiles many of you provided vary quite a bit from this. The Tour profile indicates a 9km climb with a total rise of 638m (a gradient of 7.1%). Other websites were saying it was 8.8km at 7.5%, or 7.1%, or 8.3 km long. There was even a suggestion that it was only 7.2km long! Finding the real profile seems a mission impossible! This has obvious implications for the calculated wattage and VAM on the day.

I believe that the distance of the climb is 8.7km - I base this on the fact that at the bottom of climb, as the peloton turned left, the banner signalling the start of the King of the Mountains "competition" very clearly indicated 8.7km to the summit. I can believe that the Tour website might be incorrect, but I find it difficult to believe that they would also err on the actual measurement of the route. Also, one of you said that an SRM from a Tour rider showed the length of the climb to be 8.74km. So I think it's a safe bet to say that the climb is 8.7km long.

In terms of the vertical rise, it is more difficult. It seems that most are saying that it rises 638m. A few commenters said this, and it was also the assumption used Dr Michele Ferrari in his own analysis of the climb at 53x12.com. Finally, SRM data from yesterday suggests that the climb is 640m, which differs from the 653m that is inferred from the official Tour guide and even from the mapmyride.com route, which has a 653m rise over an 8.7km length.

As a result of all this "disagreement", doing a detailed analysis on this kind of data is fraught with assumptions and therefore possible errors. As it is, Contador's climbing rate is either 1858m/hour or it is 1900m/hour. A small error - 2%, but still, going into more detail than this makes that error even greater.

So unfortunately, a detailed breakdown of the climb is not possible until one knows the altitude and distances with greater certainty - at the moment, the figures just don't match up. The mere fact that there is so much "disagreement" from one profile to the next, where the climb begins and ends, and what the gradient was means it's pretty speculative to try to work out power output per kilometer. I was really hopeful this would be possible, because it would have been fascinating to track how Contador's power peaked and then stabilised as he first followed, then attacked and built his lead.

Comparison between Contador and the history of the Tour

Nevertheless, it is still interesting to compare Contador's climb to those of riders in the past. I said yesterday that Contador's rate of 1900m/hour (based on the 8.7km @ 7.5% assumption, recall) was a record in the Tour. I know a lot of people have said that VAMs alone cannot be used to compare one climb to another. If you do so 'blindly', I agree, but provided you acknowledge that length of the climb and its gradient influences VAMs, as you'll see below, then this analysis is still quite intriguing. It will reveal, for example, that Contador's high VAM was achieved on the least steep of all the top 10 climbs in the history of the Tour. Therefore, on a steeper climb, it would be even higher. Yes, it was also a short climb, but more on that below.

If we go with the more conservative climb of 640m (as per SRM data), then the climbing rate is lower, at 1864m/hour. That is still a record in the Tour, mind you, as the graph below shows. This is a graph of the twelve highest VAMs ever in the Tour de France (click on it to enlarge if it's too small):

So Contador surpasses Riis' climb of Hautacam 13 years ago, as well as those of Pantani, Armstrong, Leblanc (a surprise name on this list!) and Ullrich. Just as an aside, Pantani's form in 1997 was astonishing.

VAM and relative power output, and influence of gradient

VAM (or vertical ascent in meters per hour) is a measurement popularised by Dr Michele Ferrari (yes, that one) as a means of comparison between riders and climbs, because it can be used to calculate relative power output.

It is important to understand that VAM is affected by the gradient. According to Dr Ferrari himself, the higher the grade, the greater the VAM at a given power output. The relationship between VAM and relative power output is defined as follows:

Relative power (W/kg) = VAM (m/hour) / (Gradient factor x 100)

This gradient factor ranges between 2.6 for a gradient of 6% and 3.1 for a gradient of 11% (The gradient factor is equal to [2 plus (% grade/10)]. You can read more on this relationship here)

Implications for the Contador climb

This has implications for Contador's climb. For Contador, the relative power output can be calculated by taking the VAM (1864m/h) and dividing by 275 (the constant for 7.5%). This gives a relative power output of 6.78 W/kg. (Ferrari calculates 6.73W/kg, but he uses a VAM of 1852m/h, because his ascent is 638m, not 640m. This shows again the discrepancies in measuring this Verbier climb!)

The point I have to make is that on a steeper climb, the same power output would produce an EVEN HIGHER VAM. Therefore, if you look at that graph above, Contador's record VAM would have been higher on any of the other slopes - Alpe d'Huez at 8.1%, Hautacam at 7.7% and Joux Plane at 8.5% would produce higher VAMS. For example, had Contador been on an 8.5% slope, producing the same power output, his VAM would increase to 1932m/hour.

In other words, Contador's climb was a record IN SPITE of the more gradual climb.

Confounders and explainers

There are a number of things that have to be factored into all these performances, particularly that of Contador. I touched on some yesterday, but here are some other contributing factors:


It's been reported that there was quite a strong wind blowing up the valley on the climb. Alex very helpfully calculated what impact a wind would have on the required power output on the climb. It turns out that with NO WIND, the power output required on the climb is approximately 422W. A tail-wind speed of 3m/s (10km/hour) reduces the power output required to 387W, which is a pretty sizeable difference. Of course, the climb cannot have had a tailwind all the way up - it had hairpins and so there will have been headwinds and tailwinds. However, this is an average tailwind, and it seems reasonable. I tried to watch for signs of strong winds on the climb, but must confess it was not noticeable.

Also, in the graph above, there is no controlling for the wind. Perhaps LeBlanc had a mighty tail-wind on Hautacam in 1994? Perhaps Pantani faced a head-wind in 1997 and could even have been faster? It's impossible to factor that in, which is why it can be risky making judgements in isolation! That is why averages over longer time-periods provide more meaningful information than once off events. The average power output on climbs over the course of a Tour tells you more than single climbs (but more on that in other posts). However, it's safe to say that wind can have a substantial impact on climbing power calculated from ascent time.

Climb length

Many have been quick to point out that the climbing rate should be higher, given that Verbier is a shorter climb than most of those done at the end of Tours. This is certainly a factor, since most of the climbs in the above list are 35 minutes long (Hautacam) or even longer (Alpe d'Huez). Soler's climb in 2007 was short - 22 minutes, but the difference in length is certainly partly responsible. Therefore, Contador's record VAM is at least partly due to a shorter climb.

I say "partly", because I don't believe that the effect of length is as great as many seem to believe. It's certainly a factor, I don't wish to dismiss it, but not as large as one might first thing. For example, when the Tour did the time-trial on Alpe d'Huez in 2004, the climbing times of all the main riders was only just marginally faster than when the same climb was done at the end of a 200km stage (all the other times in that list above). Similarly, long climbs like the Tourmalet and Mont Ventoux are climbed only a few percent slower than the shorter climbs, and so while length plays a role, and would account for some of Contador's record ascent rate, it's not as simple as saying "shorter equals faster".

Race situation

I will say that the way this Tour had gone, the first big finish was always going to be spectacular. The Tour was effectively dormant for 8 days, and the Pyrenees were done with minimal attrition. Therefore, given the situation and the way that the race had developed, the climb was always going to be fast. This again illustrates how isolated climbs can't be taken out of context, and is the reason one should look at a collection of climbs to reduce the impact of these confounders as much as is possible.


Finally, I did mention in yesterday's post that given the change in pro-cycling over the last few years, one would expect a drop in climbing rates, not new records. People will wonder about what this record ascent means - it's only natural given cycling's history!

I've hopefully managed to explain some of the other factors that must be considered in the Contador climb, but this question remains, without a doubt. It would be naive to dismiss it out of hand. I will say that performance analysis of single performances does not constitute proof of anything. In fact, it's a weak method of inferring doping. That was never my intention in yesterday's post, by the way (in case it came across that way). The better approach is to look at all climbs and work on averages, as I did for Tour winners from 1989 to 2001 in a previous post.

Why? Because doping has an effect on the repeatability of the performance, just as much as it affects performance acutely. Many will think only of the acute doping effect, but in fact, most of the doping products exert an even bigger effect on recovery, and hence the ability to produce this level of performance over and over. Think testosterone, growth hormone, cortisone, insulin - all are used to reduce stress response or improve post-exercise recovery. Even EPO would have this effect. Therefore, one cannot infer too much from a once-off performance. Rather, you have to look at a collection of performances, which also partly addresses variability provided by wind speed, temperatures and race situation.

In time, however, this performance will be placed into context - one of perhaps 10 climbs in the 2009 Tour, just as there may have been 10 climbs in 2008, 2002 or 1996. At that point, one will get a better idea of what is happening, and hopefully the analysis we did of the Tour winners 1989 to 2001 will be comparable to what is happening now.


That's the short analysis of Contador's climb. It was spectacular, without a doubt - a record in the Tour, even factoring in wind and climb length. There are too many unanswered questions regarding wind, absolute power, gradients and distances, however, which is a pity. Hopefully the discussion and the great debate it produces is worth the absence of a definitive answer! Certainly it has generated a lot of discussion, for which we thank you once again!

Bring on day 2 in the Alps!


Donate and support The Science of Sport

Thank you for taking time to read and follow our coverage of the Tour de France here at The Science of Sport. We hope you've enjoyed the insights we've tried to provide! Your time and energies are greatly appreciated!

We run this site as a "labour of love" and will continue to do so, but of course, any support is greatly appreciated. So, if you would like to donate and support our continued efforts to bring you the insights and analysis, please consider donating to our site. You can do so in one of two ways:

  1. If you are reading this on our site, then simply scroll up to the top of the page, where you will find the DONATE button at the top right of the page.
  2. If you are reading this in our email newsletter, click here to be taken to our site, where you will see the DONATE button at the top right of our homepage.
Thank you once again, and keep on reading!

Ross & Jonathan


Anonymous said...

VAMs are strongly influenced by the average percentage of the climb (which makes it quiet difficult to compare it with other climbs), by the Altitude of the climb as itself (Doctor Michele Ferrari) assumes 3% by every 500m height gain as a reducing factor by increasing altitude. Mr. Ferrari calulates 6.7 W/kg for Conatdor in Verbier quit astonishing but a little bit apart of Pantani (>7 W/kg) or even Ullrich on the Madeleine in 1998 (around 6.9 W/kg).

Anonymous said...

Interesting in the TV picture the display of the km to do shows 8.3km when Spilak reached it at the point where the climb is announced to be 8.7km with a sign on the road, just around the corner.
Moreover in found that for the stage to Arcalis the 10k mark is situated at very different locations when comparing Tour 1997, Vuelta 2007 and Tour 2009. One really has to look up the lenght and height of the climb "in nature" to be really sure aboout the parameters.

Ross Tucker and Jonathan Dugas said...

Hi Anonymous 1

Yes, quite true, VAM is affected by the gradient, but in this particular case, it actually means Contador's climb is even more impressive. That is, the higher the grade, the higher the VAM will be at a given power output. Ferrari's equations describe the relationship as Rel Power = VAM/Gradient factor. Therefore, for a given relative power, the VAM will be higher as gradient rises.

Now, given that Verbier is the least steep of the climbs in that list, Contador climbing at 1850m/h corresponds to about 6.73W/kg at 7.5%. That same relative power output would produce an even higher VAM on a steeper slope. For example, 6.73W/kg would produce a VAM of 1952m/hour on a 9% slope.

So Contador's performance, to reach that kind of VAM on a "moderate" gradient, is even more exceptional.

But you're quite right, it's amazing power output. Regarding Pantani and Ullrich, yes, those are extra-ordinary too. Do you happen to have the times and gradients of the two climbs in question? Would be interesting to compare?


Ross Tucker and Jonathan Dugas said...

Hi Anonymous 2

Yes, very odd. There's definitely some funny business going on with the distance measures! Pity, because it deprives one of digging into the climbing performance a bit more! I don't know which to believe anymore! One day, I'll head off to Verbier and try to measure it myself!


Anonymous said...

"shorter equals faster".

Shorter most definitely means higher sustainable power for the duration though, which, in most cases, will equal faster.

Ross Tucker and Jonathan Dugas said...

HI anonymous

Yes, you're right. But I already said that in the post. What I said is that it's not a simple case of dismissing the higher power output as the sole reason for it.

So if you read what I said, it's that shorter climbs ARE a factor (read that section again, I acknowledge exactly what you're saying), but that I believe people overstate this, and so it's NOT AS SIMPLE as just saying shorter = faster.

Also, remember that the Verbier is not the shortest climb ever done - they do a lot of 8km climbs. None of them are on the list. So you can't put Contador's record down to "a shorter" climb. Part of it, not all of it.


Anonymous said...

Regarding Pantani and Ullrich 1998 on the Madeleine, I once had an internet article on it bei Dr. Stefano di Santo calculating it, currently the link


doesn't work anymore, but I do have the video so I could try to figure out the parameters. And by the way I just realize once again it's incredible to ride at 25km/h on a 7% slope for 20min+.

Dr. di Santo:

Pantani 409 Watt/56 kg = 7.3 Watt/kg
Ullrich 495 Watt/72 kg = 6.9 Watt/kg

By searching the internet I found these paramaters obviously from this article:

1998 erbrachte Jan Ullrich
> (72 kg, inkl. Velo und Bekleidung etwa 80,5 kg)
> am Col de la Madeleine während
> rund 38 min eine mittlere Leistung von 495 Watt.

38min - I will see if thats corresponds with my video. But interesting W/kg seems somehow to depent on riders weight.

Ron George said...

Great analysis. Always like a piece on Contador's performance. Seriously, the guy is in a media blackhole, considering how another certain teammate has basically hijacked it.

Not certain I should say this, but I simply cannot accept Riis's and Pantani's performance paramters, no matter how good they were. One guy confessed to doping in front of the press, the other died, and drugs had a big part to blame.

I strongly believe that this year's Tour has done their job solidly in trying to be strict against doping. Thankfully, cycling is taking the lead on this and I'm very happy. No more will fans be cheated to wrongfully acquired performances.

Anonymous said...

"Thankfully, cycling is taking the lead on this and I'm very happy. No more will fans be cheated to wrongfully acquired performances."

Interesting how to change that with just the same people being in the cycling business for the last 30 years. The ones who drove their cyclists to the specialists for blood doping now represent the clean cycling, in my opinion can't be working.
Even if they are/were doped these performances can't be achieved just because they are doped, theres much more to it. Only a handfull of people is capable of winning a Tour de France each year. Maybe Contador is now in the good situation to do it without doping, but I don't blame the riders like Riis, Pantani etc., they just did their job.

Ross Tucker and Jonathan Dugas said...

Hi Ron

Thanks for the comment - yes, I think most will look at all those climbs with a fair amount of suspicion not only Riis and Pantani's!

Just to comment on the last point - I wish it were cycling that were cleaning up its act. My opinion on the UCI is that they are doing anything possible NOT to bring this more to the surface. Their hand has been forced by the likes of the French Anti-Doping agency, WADA and of course the pull-out of many sponsors and media. Up to about 2007, the UCI president still denied that cycling had a doping problem!

So I agree that cycling is being cleaned up (not clean, but under control), but I wouldn't say cycling is taking the lead!

Let's hope it does result in a clean Tour in the future!

Then to the anonymous poster, point taken.

I don't think Ron is necessarily blaming the riders personally, as much as saying that it's a shame that we're being "deceived" about true human performance capacity. I think anyone who knows cycling in the 1990s and early 2000s will appreciate how difficult a situation those cyclists were in.

I've often thought to myself that they faced an impossible dilemma, and I certainly don't blame people for giving in.

What bother me, though, and I think Ron is alluding to this, is the DECEPTION that surrounds it. It does cheat the viewer out of honest performance - nothing personal, but there is so much deception and lying about it, this veneer of being clean when it was in fact rotten to the core. That is a problem for me, not necessarily what the cyclists did. I've often heard cyclists describe themselves as "victims". I would not go that far - they still have choices between right and wrong, but I can appreciate the difficulty. Not the outright lying that accompanies it though. Then again, what else can one do...?


Alexander said...

I have the whole article from Dr. Stefano di Santo as txt-file.
He stopped the last 14.4 km of the 20km climb:

"14.4 km;
1095 m di dislivello;
una pendenza media del 7.6%;
dallo scatto al passo sono trascorsi 38'02" alla media di 22.75 km/h..."

You can find the txt file here:

Anonymous said...


Good analysis! Is the problem with comparing different hills, though, is that Verbier was a short, relatively steep climb. Contador could therefore sustain a high level of power (one would expect it to start dropping over a 10+ kilometre climb; AdH, for example, is 13.8 kms), equating to a high VAM.

Comparing different climbs is surely a fraught process for VAMs.


Anonymous said...

I don't know if this has been posted before but what about altitude?
From my memory the summit of the climb was <1500m of altitude, significantly lower than the comparable climbs (alpe d'huez).

Lincoln said...

What was the VAM of Andy Schleck, btw? Am curious as to what the VAMs were like of the guys who came 2nd, 3rd, etc on that stage. That is, were the conditions 'right' for high VAMs?

Great article

marathonman said...

I'm with Lincoln. How would Scheck's ride compare with the Top Ten?

Anonymous said...

If I have done my calcs right a rider 1:30 behind AC (like Evans or Armstrong) would be around the 1780 mark so down near Ullrich on the table.

Unknown said...

I came across this on velonews and may be of some use in finding the route details. May even be good to take Chris Ankers details and compare them to what Contador did.

Qstel said...

I am curious what Riccardo Riccò's VAM was during his climbs to Super-Besse and the Aspin in the 2008 Tour de France.

Ross Tucker and Jonathan Dugas said...

HI all

Thanks for the comments. Here's a response, one by one:

To Alexander:

Thanks so much - I'll check that article out and as soon as there is a gap in the action, will post on it!


Yes, true on the length issue - I dealt with the length of the climb and certainly, it is PART of the reason that the Verbier would have a higher VAM. I do just have to repeat the point I made in the post, that the length is not as important as some people suggest. If you look at Mont Ventoux or the Tourmalet, which are 21 or 17km long, they are climbed at about 3% less power than the shorter climbs like Alpe d'Huez. So it has an effect, but the point I was making in the post is that the effect is not as large as people make out, and one can't simply say that Contador is faster because it's shorter.

Second point - Verbier is NOT a steep climb. The average grade of 7.5% (or 7.1% depending who you believe) is less than any other climb. We know that VAMS are higher on steep climbs, so for Contador to set this record on the least steep climb is even more remarkable. And if you look in the post - you can calculate that at the same power output, his VAM on Hautacam would be 1930m/hour!

So you are right, you have to have common sense when comparing, but using VAMs still allows climb comparisons, provided you know the context. Also, you can use the VAM to calculate the relative power, which I did in the post, so I think the method has merit, though I take your point - a straight out comparison is not ideal.

To the next anonymous poster

The altitude at the top of Verbier was 1468m, Alpe d'Huez is 1860m - that's not really enough difference in that short a time frame. Verbier also starts slightly higher - about 200m higher. So I don't think that the altitude has too much influence. A percent, perhaps, but not significant.

To Lincoln and MarathonMan:

I did work out Andy Schleck's VAM - it was 1801m/hour, if you use the height at 640m. His time was 21:19. Armstrong's was 1731m/hour.

To Biking Badger:

Thanks for the link - I check that page regularly, it's got great information. I actually used the data from Sorensen to finally decide which parameters to use - the 8.74km and 640m rise. When the Tour is less "busy" and I have more time, I plan to spend some time on that data and see what it shows!


Finally, to Qstel:

I don't have the data for Super-Besse, but on the Aspin it was 1624m/hour. Not that quick, but not the final climb of the day, of course.

Thanks all!

Anonymous said...

The chart is impressive because the base value is not 0. Maybe you should change it to show that Contador wasn't so phenomal according the lenght of that climb.

Thanks for you good posts.

Anonymous said...

One way that you could account for other variables (such as tailwind) would be to run the numbers for the other top 5 or 10 finishers. How much did Contadors VAM outstretch Schleck's? How did Armstrong's performance compare to his finishes on Alpe D'Huez. If Armstrong in particular had a higher VAM than he did in 2004 (when he was notably strong) it would be safe to say that conditions were favorable for a record to be set.

Sigmund1 said...

First a big thank you for a great blog and some great postings. I am new to your site but will follow it intently in the future.

I have a couple of comments regarding this post and your earlier post on historic climbing performances and whether the recent performances relatively to historic performances indicate doping.

Equipment has been mentioned as a possible explaining factor. I would add improvement in both training methods and nutrition as possible contributors.

For example, the Norwegian commentator for TV2 has repeatedly made comments about french team's opposition to sports drinks during races up till quite recently (we're talking into the 21st century here). Apparently they preferred water.

Also, a former Norwegian pro has made comments about the lack of sophistication of continental training methods. I have heard other cyclists comment on the conservative approach to training in continental Europe.

I would imagine better training methods would allow quite vast improvements on even relatively recent performances.

One of your readers commented on pain threshold as a possible contributor to Armstrongs improvements and I must admit I don't understand why this isn't a relevant factor. Personal experience tells me that the ability to dig deep is very much an important factor in what separates the true champions from those finnishing in second place.

Anonymous said...

"Contador achieves exceptional performance with 490 watts of average "power standard" for 20min55s. It develops 445 watts in real power with 62 kg of body weight is a power of 7.2 W / kg. Undoubtedly, his finest achievement in the mountains. For relatively short climb between 20 and 30 minutes on the Tour de France, we have not found more than 460 watts (Armstrong and Basso to La Mongie in 2004) but this type of end is not so common on the Tour . We usually climbing between 30 and 45 minutes. The "world record" belongs to Bjarne Riis with 480 watts for 34 minutes to Hautacam in 1996. Contador's performance is lower."


These guys do a great analysis.


Anonymous said...

To me you make a mountain out of a mole hill (and a significant error in analysis).

Contador's vertical ascent up Verbier (8.7km) was 4% faster than Armstrong on Alp d'Huez in 2004 (15.5km).

Compare the world record for the 5k to the 10k or the half-marathon to the marathon and the "normalized" difference is about 4% and 6% respectively.

So your claim - "I don't believe that the effect of length is as great as many seem to believe." - is clearly wrong. The human engine producing the power cannot sustain the same output over twice the distance. This effect is as significant as your "result" that Contador is faster.

Hence all the other longer climbs are slower. (Not to mention Pantani was only 2% slower on Alp d'Huez...)

Ross Tucker and Jonathan Dugas said...

Hi anonymous

And I feel that you are making a significant error in comprehension. I haven't dismissed the role of the length of the climb at all. In fact, if you read the post, you'll see that I have clearly stated that we expect a faster climb and higher power output on a shorter climb. That's in the post, in black and white. So I agree with you, length affects the climb.

But what I am saying, and tried hard to clarify in the post (apparently without success!), is that in the sport of cycling, the impact of the length of the bout on performance is not as great as it would seem. It is certainly not the same as for running - we know from research that pacing is quite different in cycling compared to running over the same distance (and even time). Similarly, power output does not decay with length at the same rate for running and cycling.

You throw out the running example, which misses the point entirely. Why should cycling and running behave the same way as length increases? So I contend that it's not "entirely wrong" to suggest that length in cycling has the same impact as we tend to think from running.

The example of Alpe d'Huez is interesting - a 14km time trial produces a time only 25 to 35 seconds faster than riding the climb at the end of 200km stages. If shorter length had the same impact, the 14km time-trial would be much faster. It's not. Imagine a marathon runner doing the final 1km of a marathon at almost the same speed as they could run in a straight out 1km? Haile Geb would run the final 1km of a marathon in 2:28! That's the implication.

And then you're also missing the obvious observation that short climbs have been done before - Verbier at 8.7km is not the shortest climb in the race, there are a few in that range. Not one features on that list. And yes, there are reasons for this, but if your explanation of shorter = faster held, that would not be the case.

And finally, you look only at one climb parameter - its length. The point remains that a steeper climb (seeing as how you want to compare it to Alp d'Huez) has a higher VAM, so Contador on the Verbier (7.1%) is an underestimate compared to Alp d'Huez (8.1).

What I am saying, just as I said in the blog, is that it's not as simple as saying Contador was faster because it was shorter. There's more to it than that.

And lastly, this is all intended as point of discussion, and I presented all the mitigating factors in this post - wind, length, grade, race situation. You've lost sight of the wood for the trees, though. Your insinuation is that you are reading a journal article published as a study.


Ben said...

Ross, your comment that long climbs such as the Ventoux are climbed only 3% slower than shorter climbs like l'Alpe d'Huez makes it sound like 3% is a minor difference. But 3% is a huge difference between elite racers.

You can see an example of this in your bar chart of the fastest Tour climbs, near the top of this post. Rather than beginning the x axis at zero, which is the proper way to present such data, you have it beginning at 1700. This exaggerates the differences in the length of the bars. In fact, a 3% difference between climbing rates at VAM=1800 is a difference of 54 in VAM. This is about the difference between Contador/Verbier/2009 and Leblanc/Hautacam/1994, or Pantani/Joux-Plane/1997.

So Contador's performance is not as off the charts as it sounds, and this is obscured by the unfortunate way that you presented the data. Axes that do not begin at zero are a common problem, especially in newspaper graphics, but while doing a PhD you should have learned about this problem and know to avoid it.

Anonymous said...

Oh oh, the minutiae police are out in force.

Guys, I agree that you're right, the graph should be plotted with an axis of zero. But this is not a scientific journal, and as I read this (correct me if I'm wrong, Ross), you're trying to create a debate and analyse a current affair (the Tour) and you're doing a great job of it.

So you guys with the specifics - well done, you're quite right, but seriously, guys, you're nitpicking and fighting battles in the wrong place here. Ross publishes articles in journals - perhaps you should spend your scientific energy critiquing those papers, since they should be held to the standards you seem so willing to beat people over the head with.

Near as I can tell (and again, correct me if I'm wrong), this site does the best job I've seen anywhere of taking sports events and casting a scientific eye of them.

Keep up the great work, and you guys with the minutiae, well done, pat on the back. Now seriously, get with the objective of what you're reading - your contextual interpretation is completely incorrect.


Mister Suss said...

This is unrelated, but props on the write-up in the NYT magazine last Sunday!

Ross Tucker and Jonathan Dugas said...

Hello Mister Suss

Thanks a lot! It was an honour to have the topic covered! Pity it had to be so short - I read the comments at the bottom of the piece and it's clear that a lot of people didn't quite get the idea, and it seemed a little oversimplified! But nevertheless, thanks very much, good spotting and thanks for reading here!


Grand National said...

Altitude could play a part in the performance in this race. You can't under estimate it!

robert merkel said...

Mr Anonymous, you seem to be missing something here - that there have clearly been other climbs of comparable length through the history of the Tour, including during the 1990s. But it seems that nobody else climbed them fast enough to make the top ten list.

To Ross and Johnathan - you guys said in the post that you've got data for other climbs - can you post the fastest times for climbs of roughly similar profiles for comparison purposes?

And frankly,the people complaining about the graph are quibbling about a minor side issue that has nothing to do with the validity or otherwise of Ross and Johnathan's analysis.

Anonymous said...

I think most commenters are grateful for this analysis and simply want to debate the conclusions - without any malice. At least that's the case with me.

I'm interested, though, as to how to factor in the 3% difference in power output on longer climbs for comparison between Verbier and Hautacam and AdH (your comparative VAM graph). If Contador's VAM on Verbier was adjusted down by 3% as an example of how he might've climbed a longer climb, he'd be closer to 1800 (Pantani, Joux-Plane 97).

Cyclismag.com calculates power a bit differently, and concludes that Riis still has the record. Given that, at the very minimum, Riis' ride was EPO boosted, a conclusion that Contador was faster again is chilling.

My (very basic) analysis of AdH in 2008 is here: http://le-grimpeur.net/blog/archives/52

Anonymous said...


Thanks for your long response.

I like your website and your scientific approach and the discussion it generates. Sometimes you appear to throw in a bit of hyperbole though and conflate issues. To me good science simplifies and gets to the heart of the matter. It does not try to create mystery by looking at minor effects counterbalancing in detail (unless necessary - which is not the case here, in my opinion).

The reason I chose to compare Contadors climb on Verbier to Armstrong on Alp d'Huez was to simply make a reasonable comparison between two maximal efforts by the best in the world at the time. The majority of the 4% difference is easily explained by the difference in length. It is the first order effect, if you will.

Perhaps you would like the answer to not be so simple. As you say, that's good for discussion :)

-Paul (formerly anon)

ps> Your comparison of a marathon runner's last 1km to a "straight out" 1km time is not at all analogous to comparing a climb at the end of a 200 km stage and the same climb in a time trial. Try climbing after a 200km threshold-level solo effort and you may see something similar. I assume you know that!

Anonymous said...

Can you please tell me how much I could have increased my potential to produce increased VAM scores if I had done a hard interval on the trainer rather than reading all these posts?

Anonymous said...


Great analysis as always. really enjoy these posts, whether they're on running or cycling or whatever

Just to comment on the above criticisms:

I come from a scientific background, now working in patents, and so I appreciate the concerns that people raise about your "sensationalizing" the stories. But I think those people are missing the whole purpose here, and I therefore agree with George up above.

Guys, this is not a peer-reviewed journal. That's not to imply it lacks scientific quality, because I'd challenge you to find another resource on sport that matches it. And all these criticisms about the short length inflating the VAM - what about the more gradual gradient? Had this climb been as steep as Alp d'Huez, the VAM would have risen by 2%. So it seems that if one is going to nitpick about the length, you should at the very least consider the gradient. But then I read this post, and Ross has done all this, factored it all in.

This is a news website that is trying to provide insight into the sports events that we all enjoy. So by all means, raise concerns, but the kind of insinuations I've been reading (not just here but on some forums) are completely off base - if you want the "dry straight out factual" articles, then stick to the Journals that scientists publish in.

And I must just say - go to Pubmed and search for "Tucker R exercise" and you'll see a host of articles - so you're not reading some armchair enthusiast who doesn't cut it in science and so writes lame duck stories on the internet. So I'm sure Ross appreciates the need for "dry" unbiased and objective analysis, but this site is not the place to do it.

I have no doubt that Ross and Jon could both adopt the approach you're all 'requesting', but this site would be much poorer for it.

So keep it up guys. Love the work.

James Klein

Anonymous said...

Re: distance and vertical height

Can someone not use Google map to determine the distance and altitude (e.g., http://www.daftlogic.com/sandbox-google-maps-find-altitude.htm). I would do it for you but I'm not so interested in cycling (but generally speaking, I do enjoy reading your blog)... :-P

robert merkel said...

Google Maps/Earth is very much less than perfect when it comes to altitude readings for roads cut into the side of mountains.

Its distance measures on twisty roads also have some fuzz factor (remember, cyclists have a fairly wide choice of lines around a corner).

Anonymous said...

We can argue about VAMs, how it should be calculated, etc. but that does not change my feelings.While watching today's TT, it finally became crystal clear to me, that Alberto Contador is on some kind of performance enhancing drugs.

Anonymous said...

Why is Andy Schelck's 1801 not on the graph? The graph was supposed to be the top 12 ever?

Marco said...

Has anyone precise numbers for Contador's climb of the Angliru in 2008?


The VAM was also estimated to be more than 1930 m/h by Cozy Beehive at http://tinyurl.com/mmy7zv

On that day this final climb was steeper (up to 20%+) and at the end of a longer stage (199km) than Verbier.

Great blog!

Anonymous said...

Comparing the climb to Verbier to longer climbs like Alpe d'Huez is like comparing a 5k run (Verbier) to a 10k run (Alp d'Huez). If you compare the WR on the 5k and 10k you see that the 10k WR was run in a 4% slower average pace than the 5k, so this has quite a significant impact on the equation if you want to compare a 14k climb to one that is almost half its length. Also, if you compare the time differences with the competition on the same climb you see that it is quite similar to the strong climbs of Pantani, Riis and Armstrong. So, although Contador's climb to Verbier was quite a demonstration of strength, I don't think you can say that it's better than the races of the old champions.

Ross Tucker and Jonathan Dugas said...

Hi to tne 3rd last anonymous poster:

No, the graph was a comparison between Contador's performance and the top climbs in history. In the future, when I analyse the top 12 climbs, I'll put Schleck's on, but this post was to put Contador's into context. It will be included in that list moving forward, of course, but my intention was to look at Contador's because it is topical and was being discussed.

But you raise an important point - one should look at all riders, not just one.

To the last anonymous poster, yes, that point has been made numerous times on the thread, and also in the original post, so I don't want to repeat an argument for 8th time. That is, after all, why the whole body of this post was dedicated to explaining the context of the climb. Nobody is "blindly" saying Contador is the best ever, statistically it is, but there are a number of factors (not just length - again for the 8th time) that account for this.


Anonymous said...

So Andy Schleck's 1801 at Verbier surpassed all climbs that Armstrong's done is his Tour career?

The description of the graph is labelled "This is a graph of the twelve highest VAMs ever in the Tour de France"

Pop said...

>"I say "partly", because I don't believe that the effect of length is as great as many seem to believe."

Believing is NOT science!

Ross Tucker and Jonathan Dugas said...

To anonymous

And if you look at the graph, you'll see 12 bars, plus Contador's at the top - the post is a comparison of Contador and the climbers of the Tour. Therefore, what I have done is to look at Contador and compare him to the riders in that graph - there were 12, so the heading is appropriate.

But yes, you're right, Schleck's climb is faster than Armstrong. Again, I'd say there are a number of factors that account for this, but yes, it's surprising and anyone who didn't think "right" would be deluding themselves. It's not proof, of course, but it's interesting.

Then to Pop:

Of course it is. What do you understand by science? Of course it's "believing". Do you believe that runners are fitter than sedentary people? Do you believe that high carbohydrate diets make you fat? Do you believe that dehydration causes heat stroke?

Regardless of what your answer is, your "belief" is based on the collection of evidence that you have at your disposal. So science is very much "belief". Perhaps its a semantic issue. It's not the belief of faith, but a belief of evidence. Please read the Mission and Vision of this site before you respond with the typical "science" accusation.


Anonymous said...

Instead of being astonished that "Contador's high VAM was achieved on the least steep of all the top 10 climbs in the history of the Tour," you might just consider that an indication that Ferrari's "certain numer between 2.6 and 3.1" might be wrong, especially for the lower gradients.

For example, just moving from 2.6 to 2.8 is a difference of 50 watts! This is a huge unexplained factor.

Ron George said...

Congratulations Ross. You received a mention on Wall Street Journal abotu your analysis. See : http://online.wsj.com/article/SB10001424052970204886304574306860619951786.html

Ross Tucker and Jonathan Dugas said...

Hi Ron

Thanks a lot - yes, the journalist contacted me the other day saying they wanted a piece, could they use some of the figures and numbers. They chose the angle of "could the Mont Ventoux record be broken?", which as you'd know is probably not a priority for the top guys today! Be an interesting comparison though... Nice to be mentioned!


Anonymous said...

Interesting comments. Wonder what it would take to outfit a couple of the team cars with a setup to measure distance, gradient, wind speed, altitude, and amount of oxygen available (must vary with both altitude and pressure)? How much did it improve the rides kph to fence off the last 2 or 3km of a ride? Robert

djconnel said...

Nice article! The effect of length can be estimated with the critical power model. If AWC/CP = 60 seconds, then going from a 40 minute to a 20 minute climb increases VAM by 2.5%. This reduces Contador's VAM down to around 1815. Still fantastic, but in line with Riis 1996. Riis was perhaps on a heavier bike, but had no helmet. With Al you can get within a helmet mass of 6.8 kg, but did Riis? I'm not sure. A 1% increase in total mass would reduce VAM by around 0.57% (after considering wind resistance) if wind resistance was 20% of total, then Contador's down to 1798 on Hauticam, not counting difference in grade. Coefficient of rolling resistance is also a large factor: easily 1% difference in VAM. I don't know the relative road conditions of the two climbs, although Verbier was by accounts excellent. I suspect tires were similar. Drivetrain efficiencies are another question. Again I don't see why there would be a difference.

On time trials, riders have the disadvantage of no draft. Of course if you attack early enough you don't get a draft, either. But the draft you get early on can be of benefit. There's not much of a data set available in hillclimb time trials in the Tour. Vaughters set the Ventoux record in a time trial. The L'Alpe time trial is too isolated a point, given the importance of wind, to draw too many conclusions.

djconnel said...

I put an analysis of the effect of road grade on VAM on my blog.

Ron George said...

Dear Ross,

If you'd like fully understand vertical speed (or VAM), please check out my treatise on climbing rate here.. I hope as scientists, my perspective and take on the topic will appeal to you. I'm looking for critiquing at the same time from intelligent people,so it would be a privilege to have you guys read my article.


Note : Dan's perspective (above) through derivatives was excellent and I have linked to his blog on mine.

Ron George said...


One more point that I'd like to add in your analysis. You did not show an error estimate. Neither did Ferrari for some of his values. Please read my article. I'll show you calculation of Tom Danielson's record VAM up Mount Washington and also present how to do an error analysis with an example. This will give you an idea of how important stating the error in anyone's VAM estimation is. Contador's VAM of 1852 m/hour is not to be taken as an estimation. Thanks.

Ron George said...

Sorry, I meant Contador's VAM of 1852 m/hour is not to be taken as an absolute. Thanks.

Anonymous said...

Dear Ross and Jonathan

Congratulation on your millionth visitor.

I Denmark a blood researcher has just stated that Armstrongs Haematocrit values rise from the 11th july to the 14th of july from around 40,5 to 43 and that it was the same at the first and last day of the tour. This indicates in his opinion that blooddoping could have occurred as one would expect it to drop as it did for armstrong during the Giro d'italia.
a) how do one get hold of the riders haematocrit values?
b) If you have access to them wouldn't this approach be interesting for several riders?
AS the riders get IV-fluids dehydration is unlikely as cause of this increase right?

Biletul Zilei said...

Well worth the read. thank you very much for taking the time to share with those who are starting on the subject. Greetings

Anonymous said...

If anyone's maths are as bad as mine they will probably want to take refuge in this page as well!
Online VAM calculator

Martin Christensen said...

I believe Michael Rasmussen (kylling) also beat Riis' record. It could be interesting to know how he compares to contadors performance.

Anonymous said...

Interesting comments. Wonder what it would take to outfit a couple of the team cars with a setup to measure distance, gradient, wind speed, altitude, and amount of oxygen available (must vary with both altitude and pressure)? How much did it improve the rides kph to fence off the last 2 or 3km of a ride? Robert

Anonymous said...

Why is Andy Schelck's 1801 not on the graph? The graph was supposed to be the top 12 ever?