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.

Tuesday, July 07, 2009

Tour de France 2009: Team TT

Cancellara hangs onto the Maillot Jaune...by a whisker

Fabian Cancellara has survived in yellow for at least another day. But only just. After yesterday's dramatic split that saw Lance Armstrong gain time on all the other GC contenders, today's team time-trial was always going to be about whether Astana had the firepower to overhaul the starting deficit of 40 seconds, which would see Armstrong, on his comeback to the sport, start stage 5 wearing the yellow jersey.

And Astana did possess the firepower - exactly. For by the time the 39.5km had been completed in Montpellier, Astana crossed the line as stage victors, EXACTLY 40 seconds ahead of Saxo Bank, with Garmin Slipstream separating the two, 18 seconds down on Astana.

Those 40 seconds meant that deciding who got yellow went down to fractions of a second, and it was Fabian Cancellara who hung on, denying Armstrong a return to yellow in the comeback Tour.

There was a moment, on the open road, when Armstrong was in "virtual yellow", when Astana having actually gained a lead of 41 seconds over Saxo Bank, and it was only in the final 9km that Saxo Bank and Cancellara were able to find that half-second that sees him in the maillot jaune tomorrow.

Analysing pacing of the team time-trial

It's always interesting (especially for me with my research background on pacing) to examine how time-trials are paced. Given the very tiny margins between victory and defeat (shown by the tiny gap that kept Cancellara in yellow today, for example), errors in pacing strategy can have huge implications. This is particularly true in a team time-trial, because pacing must take into account the weaker riders' abilities. If the early pace is carried by stronger riders, then weaker riders will experience a much more dramatic slow down in the second half. The general principle in pacing is that whatever you gain by starting too fast is generally paid back in duplicate (at least) in the second half of the trial!

For example, Caisse d'Epargne shot out of the start gates today and have the fastest time at Checkpoint 1, by 7 seconds. By Checkpoint 2, they're down to 5th, and by the finish, down in 7th place, having lost over a minute to teams over the final half of the course.

So below is a basic anaylsis of the pacing strategy adopted by Astana and Saxo Bank in the time-trial today. Remember the course profile, shown below - tough first half, with a lot of uphill sections, and then a tricky second half with descending and tight turns. The strong winds, narrow roads and terrain made this TTT one of the more attritional in recent years - just ask Boygues Telecom who at one stage had four men off the road in a grass field.

Below is a graph comparing the speeds achieved by Astana (in blue) to those of Saxo-Bank (in yellow since they had Cancellara) over the four intermediate splits. In an ideal world, we'd be able to plot power output, not speed, to take into account the variable terrain and wind, but this is mostly for comparison between the two teams anyway, so it's passable. I've indicated on the graph the time difference between Astana and Saxo Bank for each interval, and at the bottom, in black, is the difference between Armstrong and Cancellara on the road, given that Cancellara started the day with a 40 sec lead on Armstrong.

What probably jumps out right away is that the eventual difference between Astana and Saxo Bank was created in the first half of the race. To be precise, Astana beat Saxo Bank by exactly 40 seconds, and 38 of those came in the first half. There was, as I mentioned above, a point at Checkpoint 3, where Cancellara had lost the yellow jersey by one second.

However, as you'll see on the graph, Saxo Bank did the fourth interval 1 second faster than Astana, and it was this 1 second, in the final 9km of the trial, that sees the Swiss champion in yellow for tomorrow's stage. It's difficult to fault the pacing strategy of either team, however. Neither of them "blew it" to the same extent that Caisse d'Epargne did in the second half.

Astana's big move came between 10 and 20km - it was here that they did all the damage, and actually won the stage.
Garmin-Slipstream, who finished second on the stage, did this interval in 13 min 8 seconds, Astana did it in 12 min 45 seconds, 23 seconds faster.

So Astana were considerably faster than anyone else in this second inverval on the route, and their stage win was ultimately the result of this effort. Whether this was because the course was particularly challenging here, or the team had decided that this was the moment to turn the intensity up a gear or two, I don't know. If you refer back to the route profile, you'll see that the second interval from 10 to 20km was certainly the most difficult part of the route. The average speeds shown in the graph above confirm this - stage 2 was considerably slower.

Tour momentum builds for the mountains

In any event, the result means that Cancellara can sleep with the yellow jersey under his pillow for at least another night. Given that the next two days look good for sprinters, that Mark Cavendish and his team are hungry for more wins, and that other teams who haven't had a look in on a sprint finish will be keen to do so before the mountains, I don't see breaks being successful.

It therefore seems likely that Friday's first mountains will be ushered in with Cancellara in yellow, and then it will be a battle to see who from Astana emerges. It's becoming increasingly difficult to see anyone other than an Astana rider winning. Today, Cadel Evans lost 2:35 to Contador and Armstrong, and Carlos Sastre 1:37, making their lives very difficult. Perhaps the Schleck brothers have a shot, though I can't see them doing enough in the climbs to make up for what should be time lost in the individual time-trial.

Astana team leadership - Armstrong in yellow equals would have meant more problems for Contador

The real talking point, as many have thought it would be, is how Astana work out team leadership. One man who might be a little relieved that Astana didn't find another half a second today is Alberto Contador. The prospect of hitting the mountains with his supposed "domestique" Lance Armstrong wearing yellow would hardly have done him much good.

The fact that Armstrong and two team-mates were at the front driving the pace on in yesterday's split sent a pretty strong message that any promised support for Contador is entirely conditional, and only if Contador is winning the race clearly will he be able to rely on Armstrong's support.

Had Armstrong been in yellow, and if the race were able to reach the first big mountain stage on Friday with that situation, then the Astana team meetings would be very interesting indeed. Would they have allowed Contador to attack the yellow jersey in his own team? It's one thing to attack when you have a team-mate ahead of you in the overall race, quite another when he is wearing the race leader's jersey. Contador might therefore be a little relieved that this is not the case.

Certainly, quotes by Contador yesterday do not suggest a rider at ease within his own team. He was quoted after yesterday's stage as saying "I don't want to express an opinion on the tactics of the team. I'll let everyone draw their own conclusions".

Your conclusions are welcome!

Next few days

Nothing is predictable in the Tour, as we saw yesterday, but the 'script' for the next few days calls for early break-aways, chasing pelotons, and sprint finishes, which should not be too eventful in the grand scheme of the race. The next big rendezvous is in Andorra on Friday, the first mountain top finish of the race, and time-permitting, I'll do a post ahead of that looking at the power outputs of Tour riders on those climbs.

Enjoy the sprint finishes!


energetich20 said...

I think we really shouldn't be drawing any conclusions about the phantom Armstrong Contador feud. Doesn't really exist. They are both at the top of their game and are going to have one heck of a time climbing and maintaining their lead.

At this point, I bet Astana is aiming to try and take all three podium places and split the winnings!

Quite the piece of riding those guys did out there today though!

Alex Simmons said...

Just a heads up for you:

If doing a pacing (or any) analysis on power meter data, I suggest being (more than ordinarily) skeptical of any power data captured by a CPU that is not the same model as the power meter (e.g. a PT CPU for a PT power meter, or an SRM PC for a SRM power meter).

For example, any Garmin unit with "Smart recording" enabled, produces unreliable data which badly distorts simple calculations like Average Power, let alone what it does to more sophisticated analysis measures and methods.

I have been doing extensive pacing analysis from the ITT - in this case on Bradley Wiggins power data and this data anomaly is readily apparent with AP varying by over 6% depending on how the data is managed by the software or spreadsheet. It isn't possible to determine what values are actually real.

Garmin's "Smart recording" isn't.

Yet the riders with Garmin seem to have this unhelpful "feature" enabled.

Oliver said...

Makes for some good racing ahead.

On the relative strategies of teams. Garmin with about four good TT riders decided that 5 was enough and not to get 'slowed' down by those who couldn't keep up...in the end letting the fifth guy hang in for counting purposes.
Cadel's team imo lost more time waiting for those who fell or mechanical problems than they gained from their presence later...esp losing 1min in first half to leaders, but not making it up.

Contador vs Armstrong...could have been beneficial to Contador not to 'help' as much...that wasn't apparent and especially in run to finish it was Contador at front and Armstrong riding 4th or 5th.
So he certainly wasn't concerned that lance may end up in yellow...which he nearly did if he had himself bunched up a bit closer to Contador.


Anonymous said...

Armstrong's efforts have been amazing! If I had to think of a word to describe it, that would be hard. Supra-physiological? If the guy wins, Ashenden needs to go back to the drawing board... this is becoming frustrating to watch.

Ross Tucker and Jonathan Dugas said...

Hi guys

Thanks, as always for the comments!

To Energetich20

Perhaps. Contador's quotes certainly indicate that he's not 100% happy, but it may just be the media portrayal of the whole thing that is in play, and that it is a phantom problem. Once we're in the mountains, then we'll know what's going on, based on how they pilot to the last of the big climbs. I definitely agree that they'll aim to take top 3, though.

To Alex:

Moot point, since I don't have any power data at all! If you do have some that you'd like to share, I'd be happy to let you do a post and put the numbers up here! It would be really fascinating to see a guy like Brad's power, and then compare it to the power in the third week.


Good point about Garmin vs Lotto, especially regarding pacing. Garmin's decision really paid off for them. I should have used that as an example in the post, because it's a really good one!

as for the work being done, I'd love to see stats on how much time each rider spent in the lead. Every time they cut to Astana, it seemed that Armstrong or Kloden was on the front. I very rarely saw Contador out in front. That might be a misleading observation, perhaps the commentators have skewed my perception by talking about Kloden or Armstrong, or maybe it was just co-incidence that they showed the team only when Contador was not in front.

My perception was definitely that Contador didn't do much work. Until the very end, that is, and then he did certainly push, and you're right, it suggests that he didn't much mind the outcome! They must have known how close they were, so that is interesting.

To the last anonymous poster:

Touche! You're right, it is developing into just that...I'm trying hard to dissociate my race comments from the personal feelings around its "protagonists" for the sake of objective race coverage - trying to stick to the science as much as I can, hence the pacing graph and so forth!

What do you mean that Ashenden will go back to the drawing board though? Not sure I follow that...

The mountains should be very interesting, and hopefully not too frustrating. The only thing I will say is that the Tour has been frustrating to watch for about 8 years (for me, personally), unless one willfully ignores all the evidence and opinion from within the sport, because it's been so clearly 'corrupted' that performances are scarcely believable. Finding the balance between cynicism and portraying what the riders actually do is a dilemma for me, I must confess...

Anyway, let's enjoy the riding for now, and see what transpires!


Alex Simmons said...

I'd be happy to do a post of some kind. I sent you a paper I did on optimising pacing some time back but don't know if you received it or not.

Problem with Brad's data is, since Garmin used Dumb (Smart) recording, the data is suspect, so a comparison to a later TT would be a bit like assessing Armstrong's efficiency over the years ;)

Saris (makers of Powertap) realised the mistake and have now made sure all team Garmin rider's 705 units have disabled Dumb (Smart) recording.

Nevertheless, here is a discussion I started at wattage forum (need to be a google groups member) on the topic of the ITT and pacing strategy.


As I discovered later on analysing the data, the elevation profile I started with didn't have the resolution it should have and the climb especially was somewhat different.

I also made an estimate on what the technical nature of the descent was worth in lost time vs a non-technical run.

If I can get a decent elevation profile for the later ITT, I can do another model.

For reference, here is the link introducing the pacing optimisation model and where a copy of the paper can be found:

Anonymous said...

I think the main reason that Caisse d'Epargne did a good first half, was simply because is more hilly. Unlikely the second that had some windy flat parts good for powerfulls ITT specialists like Cancellara, Wiggins.