Jamaica's rising star, sprinter Usain Bolt, set a new world record in
the 100-meters at the Olympics on August 16, seemingly leaving other
world-class sprinters in his dust. However, some biostatisticians
claim, "this is nowhere near what we predict for the natural limit of
the human body".

To 'predict' the next world-record times, statisticians use
mathematical models to forecast expected results. Of course these
models don't take into account factors such as advances in training or
nutrition - or even the human body's physiology - to assess the lower
limit, they simply use previous world record times, plotted on a graph
against time, and fit the data to a curve. The statisticians have used
a lower limit for the 100-meter time of about 9.45 seconds to generate
an exponential curve which - up until Bolt's recent performance - had
been fairly successful at predicting the progress of world-record
setting runs. Bolt caused researchers to have to rethink their curve a
bit - as his time of 9.69 seconds wasn't "predicted" until the year
2030!

"This trend seems to defy simple curve fitting," wrote Tatsuo Tabata, the director of the Institute for Data Evaluation and Analysis (IDEA) in Japan.

See the curve, or read comments from other 'modelers' at Wired Science.