I think the point is that all these traits are multiplicative, not summative. Your maximum running speed at the age of 40 is maybe 0.8 times what it was when you were 20, not your running speed at the age of 20 minus some. And when you have several multiplicative traits like age, gender, training history, all different aspects of genetics, etc., you get a lognormal distribution instead of a normal distribution, and with it a much larger variance between individuals.