[ORRando] Does extreme exercise prolong life?

[Michal Young]

> ...  Tour de France
> participants show they died at an average age of 81.5 years,
> compared to 73.5 in the general populations of France, Italy
> and Belgium

The reason it is bogus to infer causality from data like this is that 
we cannot distinguish between two alternative explanations: 
   (1)  Extreme exercise prolonged the life of the participants
   (2) People who are healthy enough to live a long time are 
         more likely to be successful in a profession that requires
         extreme exercise. 

We actually can't even be sure the exercise didn't harm them. 
It's conceivable that, if these particular people did not train so 
hard, they could have lived even longer. 

My intuition, like yours, is that exercise probably prolongs life. 
I certainly feel better since I started participating in endurance
exercise.  But a statistical argument requires much more careful
reasoning than this.  Experiment design, and sound use of data from 
"natural experiments", is very challenging. 

--Michal