Winner's avg. speed trends in the TdF

[StyrbjornSterki]

I've been throwing some numbers in Excel, looking for trends in average speeds of the GC at the Tour de France. To moderate the impact of changes in route and other temporary factors, I used a rolling average of 10 consecutive years. My stats begin with the first season after WWII, 1947, so the first season listed is 1956.

Speeds are in KPH, beginning the TdF of the year indicated, averaged with the previous nine

Before I get to the analysis of the latter years, I want to get to the bottom of the biggest anomaly on this graph, the 1966 to 1976 dip. I'm asking here at CN because that's a little before my time so I don't have the institutional knowledge to answer this.

What could be the reason for that decade-long decline? Were the routes that much harder? The teams that much weaker? Is it because they began trying to control performance-enhancing drugs in 1966?

From a statistical standpoint, this really strikes me odd because the dip basically represents the meat of the career of Eddy Merckx.

What have I overlooked? What are your thoughts?

 

[Jeanne]

from a purely statistical standpoint, I don't think what happend between 1966 and 1976 is significant.

For possible reasons for the slight decline in average speed there, I think there are many people on this forum that can give us information...

 

[Anonymous]

I think what would be interesting to do is overlay the total mileage for each tour, as well as the total climbing mileage for each one.

Im not old enough to remember the 70's but maybe more climbing etc. Also dont know if there were any regulation changes over equipment.

Also bear in mind massive changes in technology over the last 20 years.

the graph is a little bit decieving. We are only talking a 4kph average increase over the last fourty years.

 

[VeloCity]

I find that plateau beginning around 2006 to be interesting.

Just out of curiosity, I wonder how big a role timing systems have played over the years in the calculation of average speeds? I'd imagine they've become far more accurate than they were in the pre-digital age.

 

[hfer07]

the chart itself is wrong: 2005 is the fastest TDF ever ridden at 41.68 kph....

 

[hrotha]

the chart itself is wrong: 2005 is the fastest TDF ever ridden at 41.68 kph....

If I understand it correctly, 2005 in the chart represents the average speed of 1996-2005.

 

[icefire]

the chart itself is wrong: 2005 is the fastest TDF ever ridden at 41.68 kph....

Don't go so fast. Each point in the chart represents a 10-year average value. Read the OP.

 

[spalco]

Interesting that the coming of EPO doesn't seem to have caused a disturbance in the evolution of the average speed.

 

[2wheels]

That high reading in 1967 was just before Tom Simpson croaked on the Ventoux. The subsequent dip for the next decade was the other riders saying "Holy sh**!"

 

[Anonymous]

That high reading in 1967 was just before Tom Simpson croaked on the Ventoux. The subsequent dip for the next decade was the other riders saying "Holy sh**!"

its not year by year figures, its ten year average figures.

I think to see the reasons that ten year period 67-77 would need to be broken down into individual years. Could just be one very slow, mountainous year scewing the figures.

 

[palmerq]

Interesting that the coming of EPO doesn't seem to have caused a disturbance in the evolution of the average speed.

That high reading in 1967 was just before Tom Simpson croaked on the Ventoux. The subsequent dip for the next decade was the other riders saying "Holy sh**!"

I have moved this thread to the clinic, because it's certainly going to go down that way..... and you can have a freer discussion here.

 

[Duartista]

Oddly, the 1967-76 dip coincides with a shortening of the route, including the first 3 sub-4000 km TDFs since 1905.

The biggest change in distance was between 1987 and 1988, when around 1000km were chopped off the route.

http://bikeraceinfo.com/tdf/tdfstats.html

 

[2wheels]

its not year by year figures, its ten year average figures.

Yes, I understood the graphic. And I bet in 4,900-odd posts, you've encountered this 'wink' icon before? It means I'm making a funny.

 

[StyrbjornSterki]

The reason I didn't post in The Clinic to begin with is I was giving them the benefit of the doubt ...and hoping to avoid getting embroiled in PEDs stuff. Anyway, the by-year results graph is pretty chaotic, which is why I went to the rolling average. And the dip does not appear to be the result of one over-hard tour:

 

[Anonymous]

The reason I didn't post in The Clinic to begin with is I was giving them the benefit of the doubt

To be honest, I think a mod was being a little overenthusiastic. This can easily be discussed without doping.

Havnt looked at all the years in depth, but 1973 for instance featured 3 more mountains than 72 and was considered a tough tour.

stages classed as mountain stages.
1973 - 2048.5km
1972 - 1606.0km
1971 - 1653.1km but the mountains were easier than 72

400km is a fair difference.

I would say thats enough to account for the 71-73 figures

I think what is interesting from the chart is firstly the EPO era is pretty noticable, but also the "lance years" are pretty noticable (although the two could arguably be called the same). Then we see a steady drop once lance went walkies.

 

[StyrbjornSterki]

Especially since the images now are on different pages, this might help a bit. The annual chart begins with the first TdF after WWII. The years don't line up because the "average" graph doesn't show the years prior to 1956.

 

[StyrbjornSterki]

'73 also was Merckx-less. Dunno that had any impact but I suspect the level of competition is one of the reasons accounting for the brief plateau at 2002 -- no Ullrich or Pantani, and a 7 minute+ margin of GC victory.

In the 10-yr Avg chart, the trend beginning ~1976 continues pretty smoothly right up until 2006, then the plateau. So maybe the riders all felt the need to be more cautious post-FLandis.

 

[cat6cx]

Interesting that the coming of EPO doesn't seem to have caused a disturbance in the evolution of the average speed.

I agree, why isn't there a change in slope starting around 1990. The assumption is always that EPO drove up the race speeds.

Maybe dip is causing some statistical issues. Perhaps the terain and route pay a big role from 62 to 89 and if you could factor those out (which would be very hard to do), you wold see a flatter slope from 62 to 89 then a increase after.

Very interesting graph.

 

[Escarabajo]

Especially since the images now are on different pages, this might help a bit. The annual chart begins with the first TdF after WWII. The years don't line up because the "average" graph doesn't show the years prior to 1956.

Nice charts.

At first view I'd recommend cutting the moving average to maybe 3 years, max 5. 10 years is too much. It smoothes the curve too much.

The year by year is not so good either. Is chaotic because is specific of how hard the mountain stages are, how it was raced, wind factors etc. With the moving average it kinds of decreases those effects to certain point.

Good work though.

 

[ultimobici]

Have you taken into account the time bonuses given up to the 90s? Some sources for the average speed may be skewed if they have used the winner's official time rather than the actual time.

 

[roundabout]

On a somewhat related note, I've recently come across some interesting data regarding a MTT over what is very likely to be the same course in the TdS. The years are 1969/72/78/81/85/90/93.

The 1993 results are quite telling.

 

[frenchfry]

I think what would be interesting to do is overlay the total mileage for each tour, as well as the total climbing mileage for each one.

Im not old enough to remember the 70's but maybe more climbing etc. Also dont know if there were any regulation changes over equipment.

Also bear in mind massive changes in technology over the last 20 years.

the graph is a little bit decieving. We are only talking a 4kph average increase over the last fourty years.

I don't know about you, but if I do the same 100km ride at 30kph it isn't the same as at 34kph, the difference between feeling good and on the limit.

 

[Anonymous]

I don't know about you, but if I do the same 100km ride at 30kph it isn't the same as at 34kph, the difference between feeling good and on the limit.

but this is where its difficult to read too much into these things.

What if for one of the 100km i put a hill in the middle, or made your bike 2lb lighter, or moved your gears from the downtube to the handlebars. this is why its so difficult to compare.

That for me though is why the lance period stands out. Technology wasnt that difference, the courses were similar, but something in those 5 years gave the winner a bit of a lift.

 

[D-Queued]

but this is where its difficult to read too much into these things.

What if for one of the 100km i put a hill in the middle, or made your bike 2lb lighter, or moved your gears from the downtube to the handlebars. this is why its so difficult to compare.

That for me though is why the lance period stands out. Technology wasnt that difference, the courses were similar, but something in those 5 years gave the winner a bit of a lift.

You clearly are new to the sport. Not something, three things:

Will power, high cadence and hard training.

Dave.

 

[rgmerk]

There was a paper in one of the sports medicine journals doing this (in a more mathematically sophisticated way using time series analysis) which showed similar things, IIRC.

Can't recall the reference, though.