Pacing strategy from high-frequency field data: more evidence for neural regulation?

We demonstrate a methodology that uncovers an athlete's true pacing strategy from high-frequency (≤1 km) split field data, even if affected by high gradient variation on course. The method thus opens up the analysis of many previously opaque but popular undulating professional and amateur races to scientific scrutiny.

The method is relatively simple to use in any standard statistical package, and execution only requires the addition of the altitude-distance trace of the event to a runner's split times (e.g., as automatically collected by a modern Global Positioning System-enabled wristwatch). In addition, as opposed to assuming a pacing function (e.g., "J shaped," "U shaped," "all-out") and testing this function on the data, the method uses a preliminary discovery step to suggest the most appropriate pacing function(s) to test on the data (if any).

The method is demonstrated with two novel case studies: Gebrselassie's world-record Berlin marathon (September 2008) and a unique data set taken from several years of the Six Foot Track Ultramarathon (45 km, Sydney, Australia).

In both cases, the method reveals highly variable pacing strategies on a microscale despite remarkable symmetry on a macroscale in one case adding weight to the recent complex system perspective of the neural regulator.