World record progression for men's 100 m. Data, from Ref. 1; dotted line, least-squares fit of exponential function to data through 2005; red line, fit to data through 2008; and purple line, fit to data through 2009. (You can see the real size image by clicking on the image).

Until the year of 2005, the plot of the world record for men's 100 m sprint as a function of year allowed a good fit by an exponential function with an additive constant, value of which meant the possible limiting value of the record. The data through 2005 gave the limiting value of 9.66 seconds. The data through 2008, however, predicted an improved limiting value of 9.43 seconds owing much to Usain Bolt's record, in the Beijing Olympic Games, of 9.69 seconds, which was one of outliers with respect to the fitted curve (Ref. 2).

Now his record of 9.58 seconds in Berlin on August 16, 2009, again changed the limiting value down to 9.09 seconds (with errors of plus or minus 0.49 seconds). When a single new record affects the possible limiting value this much, it would mean the following two facts: (1) The application of exponential fitting to data on men's 100 m sprint is completely inadequate. (2) Bolt's running ability is extraordinarily and exceptionally wonderful.

Notes added later:
Originally the title of this article was "Bolt's World Record Changes Statistical Prediction Again," but I replaced the word "Statistical" by "Empirical." The reason is this: The method used is often called statistical, but I think it better to be called empirical. Further, I'm not a statistician but a maker of empirical formulas.

A related blog article appeared: Michael Banks,

Bolt out of the blue, physicsworld. com (August 17, 2009).

Read also

Bolt's New 200-m Record and the Limiting Value.

- "World record progression 100 metres men," Wikipedia, The Free Encyclopedia (18 August 2009 at 00:44).
- "World Records for Men's 100 m Defy Simple Curve Fitting (2)" Femto-Essays (September 14, 2008).

Last revised: August 24, 2009.