an age-invariant relation between experience and performance?
john "the penguin" bingham's running column raises an unusual life course hypothesis this month:
"A running guru once told me that no matter the age you start running you can expect to improve your PR [personal record] throughout the next seven to 10 years. If you want to see faster times after that, you'd better find a few downhill courses." [emphasis added]
i haven't seen any data bearing on the guru's hypothesis, but it sounds about right to me -- at least for any distance greater than a mile and for any age in the range between 16 and 106 years. it might seem obvious to suggest that a runner who first laces 'em up at 20 will likely peak around age 27 or 30, but friends who started running at 30 or 50 or 70 also seem to make marked improvements over their first decade. moreover, those who survive as runners in all of these age groups almost inevitably decline in performance over their second decade pounding the pavement.
are there any other activities that would exhibit a similar age-invariant relationship between experience and performance? i'm stumped. in some demanding physical activities, performance rises and falls very rapidly (e.g., playing linebacker, opening beer bottles with one's teeth), but the rate of improvement and decline surely varies for twentysomethings and sixtysomethings. for other activities, there is a gradual, running-like ramp-up in performance for a decade or more, but decline is hardly inevitable and likely depends more on age than duration (birdwatching, cooking, chess, karaoke).
distance running (or similar activities, such as bicycle racing or wheeling) might be unique in that almost anyone can pick it up at almost any age, it rewards long years of training and experience, and it ultimately imposes a predictable physical toll. hmmm. i guess i can think of one activity that might exhibit a similar age-invariant relationship between experience and performance, but i dasn't mention it. after all, this is a g-rated family blog.