Modelling the development of world records in running

We model the development of world records of metric running events from the 100 meter dash to the marathon for men and women. First, we review methods to fit time-series curves of world records in general. We discuss methods to estimate curves and review candidate functional forms that fit the systematic shape of the progress of world records. Next, we fit the asymmetric Gompertz-curves for 16 events and compute implied limit values. In order to assess the implied limits we use the Francis (1943)-model to relate limit records and distance in a log-log specification. We compare men and women and conclude that there is a fixed difference in record times between the two sexes. Finally, using the log-log relationship between time and distance we calculate the development of the world record of the mile in a robustness check.