Statistician Simon D. Angus of Monash University in Melbourne, Australia, writing in Medicine & Science in Sports & Exercise, says the earliest we’re likely to see a sub-2 marathon is 13 years from now. His prediction is that the chance of it happening by May 2032 are about 10 per cent (or one in 10), and that we could see a man run 1:58:05 and a woman run 2:05:31.
Angus claims this is the first study to consider when a sub-2 is likely, just how much under two hours humans can be expected to run, by what date, and at various probabilities (taking into account at what rate, and by how much, the records have been improved upon in the past). It may also be the first study to seriously consider how fast we can expect to see a woman run the marathon in the future. Angus uses statistical modelling to arrive at his predictions.
He also defends his use of the 10 per cent likelihood benchmark by pointing out that record-breaking performances in the marathon are, by nature, extraordinary and unusual, and therefore we should be asking not “when is it likely?” but “when might it be possible?”
(By this logic, Angus has calculated that the odds of a sub-2 marathon happening by 2050, rather than 2032, are one in four, or 25 per cent.)
According to Angus, speculation about how fast a human could run 42.2K goes back to at least 1925, when the British medical journal Lancet published a study by A. V. Hill entitled “The Physiological Basis of Athletic Records.” As time has passed and more data has become available as more marathons are run, it’s become possible to predict future performances with greater accuracy. For example, some years ago researcher M. J. Joyner predicted a sub-2 marathon by 2021-2022, but Angus, based on more advanced methods used by Angus, he believes that prediction was overly optimistic. (Joyner & Ruiz’s study published in the Journal of Applied Physiology in 2011 uses “linear extrapolation without accounting for variability, compared to the non-linear, stochastic modelling of the present work.”)