PROJECTION OF WORLD RUNNING RECORDS

Abstract

Dear Editor-in-Chief: Whipp and Ward claim that their 1992 article (4) investigating world-record running speeds recorded during the 20th century has been misrepresented in our recent article (2). Whipp and Ward (4) fit linear models to world-record running speeds for both men and women (distances 200 m, 400 m, 800 m, 1500 m, and marathon) and highlight with arrows, when the men's and women's predicted world record times would intercept (based on theoretical extrapolations of the linear models). In their letter, Whipp and Ward argue that their 1992 paper (4) neither stated nor implied that women will run world-record times faster than men. While it may not have been their intention, the very title of their article "Will Women Soon Outrun Men?" implies that such a conclusion might be inferred from the evidence presented (however unlikely), a conclusion further emphasized by the readers' attention being drawn to the points when the men and women's lines intercept. It is interesting to note that we are not alone in our interpretation; see Astrand (1). We are delighted by Whipp and Ward's support for our strategy of fitting a flattened "S-shaped" logistic curve and the resultant findings. We do, however, have reservations regarding Whipp and Ward's claim that their choice of linear models was guided by the "rule of parsimony." World records in a number of events (women's 200 m, 400 m, and 800 m) have not changed since 1992. However, the flattened S-shaped logistic model was found to fit these data significantly better than linear models; that is, the F ratio (and significance) associated with the two additional logistic curve parameters for women's 800-m world records was F2, 19 = 9.4 (P < 0.001); see Nevill and Whyte (2). Clearly, even prior to 1992, there was sufficient evidence to confirm the superiority of the flattered S-shaped logistic model compared with linear models. We now recognize that Ward and Whipp (3) had previously considered a sigmoid-like relationship when modeling their data. However, given that Ward and Whipp had considered similar nonlinear models, we find it even more surprising that in their letter, Whipp and Ward suggest that linear modeling is the most appropriate/parsimonious model to describe running world-record speeds, especially given the superior fit of the logistic model, even on data that remain unchanged since 1992. We are in complete agreement with Whipp and Ward regarding the multifactorial basis of human performance. It would appear clear from our evidence (2), however, that women will not run faster than men in Olympic distance events. Indeed, in recent as yet unpublished work we have observed similar trends in swimming world records that support to the use of the flattened S-shaped logistic model. These results confirm both a deceleration in more recent world-record swimming times and the absence of convergence between men and women. Alan Nevill Research Institute of Healthcare Sciences University of Wolverhampton Walsall, United Kingdom Gregory Whyte The High Performance Centre English Institute of Sport Bisham Abbey National Sports Centre Bisham, United Kingdom