Cleaner Peloton = Same Speeds?

[MartinGT]

I saw Festina girl retweet something on Twitter that said something along the lines of "With doping came increased speeds, since doping has declined, why havent avg Tour speeds?"

Something along those lines.

I have looked into the Tour Speeds from 1998 and its interesting,

1998 - Avg Speed 39.983 - Length (Km) 3875
1999 - 40.273 - 3686
2000 - 39.556 - 3662
2001 - 40.020 - 3446
2002 - 39.930 - 3282
2003 - 40.940 - 3427
2004 - 40.553 - 3391
2005 - 41.654 - 3608
2006 - 40.781 - 3657
2007 - 38.980 - 3547
2008 - 40.500 - 3560
2009 - 40.310 - 3560
2010 - 39.585 - 3642
2011 - 39.788 - 3430
2012 - 39.830 - 3488

Avg info from 1999 - 2005 (US Postal era) Avg Speed - 40.418 Avg Length - 3500.286

From 2006 - 20012 Avg Speed 39.968 Avg Length - 3554.857

Surly its not all down to technological advances is it? When did the minimum weight come in?

/edit - Tour Data taken from here - http://www.bikeraceinfo.com/tdf/tdfstats.html

 

[Ferminal]

I find overall speeds to be terribly misleading.

They would have increased without EPO as the sport professionalised. They will continue to fluctuate in the future and we will have a lot of difficulty attributing those changes to doping.

I remember reading an article a couple of years ago which commented on the entire history of speed changes, we discussed it hear but I can't seem to find it.

 

[DirtyWorks]

Sampling the entire TdF in particular is a good example of the curse of averages. It tells you nothing about athletic performance.

Another example of the curse of averages is put Bill Gates in a room with 10 homeless people, and suddely everyone in the room is wealthy.

 

[esafosfina]

I'd say that apart from massive advantages in bikes/equipment there's been a quantuum leap in training regimes and nutrition, let alone rider care.

 

[Aleajactaest]

Sampling the entire TdF in particular is a good example of the curse of averages. It tells you nothing about athletic performance.

Another example of the curse of averages is put Bill Gates in a room with 10 homeless people, and suddely everyone in the room is wealthy.

Poor example. The average speeds are for the winning times, not all times.

In any case, times and performance in all sports go up/down based on so many factors. In this case, better training, nutrition etc will all help balance out the effects of drugs. Go back 20 years and see how un-scientific the sport was. No aeromods, no carbon fiber, no lactic acid threshold testing etc.

It's natural for things to get better.

 

[MartinGT]

Marginal gains, damn, forgot about this

 

[Cloxxki]

I'd say that apart from massive advantages in bikes/equipment there's been a quantuum leap in training regimes and nutrition, let alone rider care.

Yet, how much have distances for the hour record changed since Merckx, for the regular race bike?

 

[ScienceIsCool]

I find it more interesting to look only at the flat stages. The speeds should be dictated mostly by the riders and not by any technological advances (weight, aerodynamics, frame stiffness, etc).

As you would expect, I have found that the longer a stage is, the slower the speed. This is true for the data I've looked at from 1980 to 2012 and the trend lines have the same slope. Also interesting is that the average length of a flat stage hasn't changed at all in 30 years. It's roughly 205 km and can span from 160 to 250 km.

Looking at 1980 (pre-oxygen vector doping), 2000 (height of blood boosting), and 2012 (are we cleaner?), there are some surprising results.

In 1980 the average speed of a flat stage was 36.8 km/hr. In 2000, the average speed of a flat stage was 42.9 km/hr. And this year, it was 42.6 km/hr.

My conclusion is that the likeliest explanation for the differences in speed is that oxygen vector doping (and other modern methods) allows for better recovery after hard efforts. I also conclude that things aren't remarkably cleaner in 2012. In fact, the average speed of a flat stage actually increased as the race went on. In the first half of the Tour this year a flat stage was raced at 41.1 km/hr (avg 203 km length). In the second half it was 43.8 km/hr (avg 200 km length).

John Swanson

 

[MrRoboto]

Can you look at some mid 90's tours too?

 

[maltiv]

Using average speed as an indication for anything is ridiculous. There are so many different variables that are significant which have nothing to do with the riders' endurance: weather, road surface, equipment, style of racing, elovation and so on.

For example, it was perhaps more common in the past that the breakaway was let go early and the peloton took it sort of easy, whilst nowadays the competition is a bit harsher and it often takes an hour for the break to form in TDF, which pushes the average speed up a lot.

 

[MartinGT]

I think the speeds for climbing is the key speed to look at.

Yes we have had advances in technology but at the moment would you say that these technological advances change the Acg speed that much?

Interesting.

 

[taiwan]

[Overall speeds] would have increased without EPO as the sport professionalised. They will continue to fluctuate in the future and we will have a lot of difficulty attributing those changes to doping.

This thread is interesting to me because it fits my bias. The quoted point is correct I think, but that may be a minor factor compared to a large chunk of the peloton jacking their power up by 5-10%. One thing to note is that the speed hasn't gone up either.

Is the peloton:
- clean but more professional (cancelling out)
-a bit cleaner but more professional
-still dirty and no more professional.

 

[the sceptic]

Did they know about warming down in 1998?

 

[cineteq]

Using average speed as an indication for anything is ridiculous. There are so many different variables that are significant which have nothing to do with the riders' endurance: weather, road surface, equipment, style of racing, elovation and so on.

Just having a different parcours every year is enough to dismiss the OP's premise.

 

[Ney the Viking]

Still with a 3500 km tour the average speed in postal era is about an hour faster than the average speed post postal.

((3500 km) / (40.41800 kph)) - ((3500 km) / (39.96800 kph)) = -58.4984787 minutes

So postal was quite a bit faster on average. Is this analysis worth anything? *shrugs*

 

[taiwan]

I'd think that looking at average speeds was a good indicator of clenliness. You're looking at the speed averaged out over 21 stages, the speed is dictated by the whole peloton as they take turns on the front, and the OP gives data for a 14 year period. Things like weather and road surface get smoothed out IMO. Only issues are long term trends in racing practise, but tbh I could see these following the power riders have at their disposal anyway.

What we need is comparable data from the eighties and nineties. If the graph goes:
________________________________________
___________________/

...then that's enough for me.

In 1980 the average speed of a flat stage was 36.8 km/hr

If that's the sort of difference we're talking about for a flat stage then case closed.

 

[Benotti69]

I find it more interesting to look only at the flat stages. The speeds should be dictated mostly by the riders and not by any technological advances (weight, aerodynamics, frame stiffness, etc).

As you would expect, I have found that the longer a stage is, the slower the speed. This is true for the data I've looked at from 1980 to 2012 and the trend lines have the same slope. Also interesting is that the average length of a flat stage hasn't changed at all in 30 years. It's roughly 205 km and can span from 160 to 250 km.

Looking at 1980 (pre-oxygen vector doping), 2000 (height of blood boosting), and 2012 (are we cleaner?), there are some surprising results.

In 1980 the average speed of a flat stage was 36.8 km/hr. In 2000, the average speed of a flat stage was 42.9 km/hr. And this year, it was 42.6 km/hr.

My conclusion is that the likeliest explanation for the differences in speed is that oxygen vector doping (and other modern methods) allows for better recovery after hard efforts. I also conclude that things aren't remarkably cleaner in 2012. In fact, the average speed of a flat stage actually increased as the race went on. In the first half of the Tour this year a flat stage was raced at 41.1 km/hr (avg 203 km length). In the second half it was 43.8 km/hr (avg 200 km length).

John Swanson

But Wiggins wons and Sky said they dont dope so it must be cleaner

 

[maltiv]

I'd think that looking at average speeds was a good indicator of clenliness. You're looking at the speed averaged out over 21 stages, the speed is dictated by the whole peloton as they take turns on the front, and the OP gives data for a 14 year period. Things like weather and road surface get smoothed out IMO. Only issues are long term trends in racing practise, but tbh I could see these following the power riders have at their disposal anyway.

Remember, the AVG speed is only the speed of the winner. You're also forgetting technology. The new wheels make a huge difference on the flat in terms of rolling resistance and aerodynamics. Combined with far more aero jerseys, helmets, pedals and so on, the energy loss is reduced by 15-20% for high speed racing.

You also have to take into account completely different ideas on how to design a stage. Lately, having 10 completely flat stages is common. That pushes up the average speed. So does having a long, flat TT or particularly a long TTT.

Climbing speed is a much, much better indication. Aerodynamics is irrelevant there, and because of the long-lived 6,8 kg rule, we can almost completely remove technology from the equation. Moreover, using only finishing climbs, one can assume that they ride as fast, or close to as fast, as possible (unlike a flat stage, when the peloton sometimes can take an early holiday and do the 200 km in 31 km/h). So, since the climbing speeds have been significantly lowered since the 90's, it's very simple to conclude that today's peloton is cleaner. I mean, Pantani would've put close to 5 minutes on nr 2 if he had ridden Alpe d'Huez in TDF 2011...

 

[cineteq]

Climbing speed is a much, much better indication. Aerodynamics is irrelevant there, and because of the long-lived 6,8 kg rule, we can almost completely remove technology from the equation. Moreover, using only finishing climbs, one can assume that they ride as fast, or close to as fast, as possible (unlike a flat stage, when the peloton sometimes can take an early holiday and do the 200 km in 31 km/h). So, since the climbing speeds have been significantly lowered since the 90's, it's very simple to conclude that today's peloton is cleaner. I mean, Pantani would've put close to 5 minutes on nr 2 if he had ridden Alpe d'Huez in TDF 2011...

Voilà.

Alpe d'Huez
1. Pantani -------- 36’50″, 1995, 22.5 kph
2. Pantani -------- 36’55″, 1997
3. Pantani -------- 37’15″, 1994
4. Armstrong ------ 37’36″, 2004 (ITT)
5. Ullrich -------- 37’40″, 1997
6. Armstrong ------ 38’01″, 2001
7. Indurain/Zulle - 38’10″, 1995
8. Riis ----------- 38’15″, 1995
9. Virenque ------- 38’20″, 1997
10. Landis -------- 38’34″, 2006, 21.5 kph

http://le-grimpeur.net/blog/archives/52

 

[skidmark]

I find it more interesting to look only at the flat stages. The speeds should be dictated mostly by the riders and not by any technological advances (weight, aerodynamics, frame stiffness, etc).

As you would expect, I have found that the longer a stage is, the slower the speed. This is true for the data I've looked at from 1980 to 2012 and the trend lines have the same slope. Also interesting is that the average length of a flat stage hasn't changed at all in 30 years. It's roughly 205 km and can span from 160 to 250 km.

Looking at 1980 (pre-oxygen vector doping), 2000 (height of blood boosting), and 2012 (are we cleaner?), there are some surprising results.

In 1980 the average speed of a flat stage was 36.8 km/hr. In 2000, the average speed of a flat stage was 42.9 km/hr. And this year, it was 42.6 km/hr.

My conclusion is that the likeliest explanation for the differences in speed is that oxygen vector doping (and other modern methods) allows for better recovery after hard efforts. I also conclude that things aren't remarkably cleaner in 2012. In fact, the average speed of a flat stage actually increased as the race went on. In the first half of the Tour this year a flat stage was raced at 41.1 km/hr (avg 203 km length). In the second half it was 43.8 km/hr (avg 200 km length).

John Swanson

That is a really dangerous conclusion to make. Any factorial analysis worth its salt (to determine even correlation, leaving causality aside) breaks the problem into a number of different possible factors, and tries to attribute how much each of those factors contribute to the outcome (in this case, average speed). To say 'the most likely explanation for differences is modern methods for recovery' and then say 'things are faster than in 1980; it is due to doping' as if all of a sudden that is contributing 100% to the average speed is fallacious on many levels. Not saying it isn't true, but it certainly isn't scientific.

Anecdotally, I immediately think of the Giro 2009, when some old guy thought the parcours in Milan was too dangerous so the peloton rode the whole stage at 30km/h. I don't know the average speed of the Giro, but that certainly would be a significant contributing factor in taking the average speed of the flat stages down. As well, if riders' subjective reactions are to be taken seriously, I have heard a number of riders say about the Vuelta 'you used to be able to take it easy there, now everyone wants to go faster every day'. There are many factors that could contribute to overall speeds. Sure, overall trends are useful to look at as 'interesting information', but associating speed with doping is just far too simplistic.

 

[taiwan]

Remember, the AVG speed is only the speed of the winner. You're also forgetting technology. The new wheels make a huge difference on the flat in terms of rolling resistance and aerodynamics. Combined with far more aero jerseys, helmets, pedals and so on, the energy loss is reduced by 15-20% for high speed racing.

You also have to take into account completely different ideas on how to design a stage. Lately, having 10 completely flat stages is common. That pushes up the average speed. So does having a long, flat TT or particularly a long TTT.

Climbing speed is a much, much better indication. Aerodynamics is irrelevant there, and because of the long-lived 6,8 kg rule, we can almost completely remove technology from the equation. Moreover, using only finishing climbs, one can assume that they ride as fast, or close to as fast, as possible (unlike a flat stage, when the peloton sometimes can take an early holiday and do the 200 km in 31 km/h). So, since the climbing speeds have been significantly lowered since the 90's, it's very simple to conclude that today's peloton is cleaner. I mean, Pantani would've put close to 5 minutes on nr 2 if he had ridden Alpe d'Huez in TDF 2011...

Have the number of TT miles been creeping up in the last decade?

The winners' times are used, but the winner will more or less always spend most of their time in the bunch. The advantages of aero equipment are lessened in the bunch. Wheels - OK, modern training - OK but from the gut doping would be a more influential factor. I think the av speed is good because you're looking at a bigger sample. Admittedly hard to pick the causes out in a credible way though.

 

[Spencer the Half Wit]

I'd think that looking at average speeds was a good indicator of clenliness. You're looking at the speed averaged out over 21 stages, the speed is dictated by the whole peloton as they take turns on the front, and the OP gives data for a 14 year period. Things like weather and road surface get smoothed out IMO. Only issues are long term trends in racing practise, but tbh I could see these following the power riders have at their disposal anyway.

What we need is comparable data from the eighties and nineties. If the graph goes:
________________________________________
___________________/

...then that's enough for me.

If that's the sort of difference we're talking about for a flat stage then case closed.

If you took the end points of the basic graph you posted and drew a gradual upward curve then that would be normal due to the factors mentioned by other people i.e. better equipment, advanced tactics, better diet, better sports science etc.

As part of my statistics degree twenty years ago we did a project on what the optimum time would be for the 100m ten, twenty and fifty years in the future based on past records. The results showed a stepped downward pattern i.e the record would be beat then a few years before the record would be beat again. You could then plot a curve based on the data points which showed a gradual decline in the times i.e. a gradual increase in speed. The reason why the 100m was chosen because it had the least external factors i.e you basically ran as fast as possible for 100m. The amount of external factors in place for a 3 week 3,500km bike race would mean you could not give any statistical confidence to such an analysis.

Having said that having an average speed slowly increasing the a huge spike then a leveling off or slow decrease could mean that there was a huge external influence at the start of the spike.

Therefore the graph you have posted could, without any statistical confidence, be said to be indicative of a period of high blood doping from the early nineties, gradually reducing from the early 2000s onwards.

 

[Dear Wiggo]

Sampling the entire TdF in particular is a good example of the curse of averages. It tells you nothing about athletic performance.

Another example of the curse of averages is put Bill Gates in a room with 10 homeless people, and suddely everyone in the room is wealthy.

Depends if you look at median or mean wealth.

I think the key variable here is how the race is raced. The point being - noone takes it easy at the TdF as it's the race that makes your career.

 

[trevim]

Obviously average speeds isn't a very good method of analysis. Still, one thing that has always amazed me is: how can they make the same sort of average speed in a 3rd week of a GT? I would expect speed to go down as the days passed...

 

[Froome19]

2:06:05 Ronaldo da Costa Brazil September 20, 1998 Berlin Marathon IAAF,[39] ARRS[40]


2:05:42 Khalid Khannouchi Morocco October 24, 1999 Chicago Marathon IAAF,[39] ARRS[40]

2:05:38 Khalid Khannouchi United States April 14, 2002 London Marathon IAAF,[39] ARRS[40] First "World's Best" recognized by the International Association of Athletics Federations.[64] The ARRS notes Khannouchi's extended time as 2:05:37.8[40]


2:04:55 Paul Tergat Kenya September 28, 2003 Berlin Marathon IAAF,[39] ARRS[40] First world record for the men's marathon ratified by the International Association of Athletics Federations.[65]


2:04:26 Haile Gebrselassie Ethiopia September 30, 2007 Berlin Marathon IAAF,[39] ARRS[40]


2:03:59 Haile Gebrselassie Ethiopia September 28, 2008 Berlin Marathon IAAF,[39] ARRS[40] The ARRS notes Gebrselassie's extended time as 2:03:58.2[40]


2:03:38 Patrick Makau Kenya September 25, 2011 Berlin Marathon IAAF,[66] IAAF[67]


Effectively the same as the WR for the marathon progression time. Most of the reasons for it have already been mentioned.