Faster till when?

In the December issue of Significance, there is an interesting article by Joseph Hilbe (who also wrote a nice review on our book Bayesian Core in the Journal of Statistical Software) on the prediction of the ultimate time in athletics. (There are plenty of interesting articles in this issue, actually! In a few years, this general audience Statistics journal from the RSS truly turned into an excellent magazine.) The questions are whether or not there is a lower bound, called the ultimate performance time, to the performance times in athletic events like the 100m or the marathon, and if yes, what is the predicted value for this ultimate performance time. Hilbe’s paper is actually a criticism of a single paper by Kruper and Sterken published this year in Statistical Thinking in Sports. The shortcomings of Kruper and Sterken’s model are clearly exposed, the most glaring one being that the ultimate performance time for women 5000m has been beaten this year by four seconds… For instance, the predicted ultimate time for the men marathon is 2:00:56, in relation to the current world record of 2:03:59 by Gebrselassie. There is no reason to believe that younger men like Nicolas Menza who won the Boulogne-Billancourt half-marathon this year in 1:00:12 cannot improve this record if training earlier for marathons. So the model of Kruper and Sterken is limited and more factors should be taken into account like training facilities and techniques, morphological improvements, alimentation and medical advances, not to mention drugs or genetic manipulations… But I do not understand the perspective taken by the paper that does not go further into Statistics than factual descriptive statistics and that concludes that “experience in the sport and a basic knowledge of human physiology and biomechanics are better (to make) predictions than mathematical analysis“. Hardly fitting for enhancing the profession for the general public and for a journal whose motto is Statistics making sense!!!

2 Responses to “Faster till when?”

  1. Joseph Hilbe Says:

    I used the Kruper and Sterken model since it was the lead article in a ppopular book on the subject. Most general readers to not read journal articles on sports statistics. The problem is that I could have used many other similar attempts to model performance and use it to predict out of sample events in the future. This is pretty risky business. The model predictions I have seen are generally common sense predictions that don’t require a model to predict with some assurance, like the 2:01 marathon — but I think Bekele has a better chance of doing it than Menza.

    I recently saw an article from the Journal of the Quantitative Analysis of Sports (I think that is the official title) that make a few better predictions than Kruper and Sterken, but with confidence intervals wide enough to include most any reasonable performance. My original conclusion remains.

    I attempted to demonstrate in my article that, for the most part, improvement in athletic performance has been very slow, if adjusted for a host of conditions, eg change in competition surfaces, implements, training technique, shoes, remuneration, and so forth. World records are the outiers, to be sure, and are many times achieved by a combination of physical preparedness and good luck. There are so many things that must be taken into consideration. In training an athlete to a word record in the men’s shot put in 1976, and others to national championships — including myself — I have experienced how random luck effects performance at the very top level. Remember Bob Beamon’s 8.90 (29’2.75″) long jump at the 1968 Mexico city olympics. He had never jumped further tham 27’2″ before – nor afterwards. The wind assisted the jump to the exact limit, his foot was barely over the board, but did not make a mark in the plasticine, he jumped at high altitude — everything just right. Carl Lewis, a far superior jumper, never officially broke Beamon’s record, but never had everything working together. I watched him jump near 30′ in 1983, but foul by less than a centimeter.

    Or consider Michael Phelps – his 8 gold medals, three of which were as a member of a relay team. He was behind in one race, but because his body was placed properly, he could reach straight and touch the timign system a 100th of a second faster than another swimmer whose body was closer to the finish, but who arched to reach the finish, costing a hair of a second slower. Think of the many truly lucky factors that went into his 8 wins, and now give me a model that could have predicted it?

    I believe that statistical modeling does have a place in assessing athletic performance, and even within limits of predicting out-of-sample future performances. But not all events can be subsumed under one model. Finally, if a model is in fact developed that predicts a future performance within its confidence intervals, but if the intervals are so wide to make the prediction superfluous, I take it that the model is superfluous as well.

    Care must be taken when attempting to explain inherently different events using the same model. Care must also be taken to develop a model that actually provides us with new information; ie tells us something that we would not otherwise be able to predict.

    Joseph Hilbe
    Chair, ISI Sports Statistics committee

    • Thanks for the comment, Joseph. I understand your point and it was already well-made in the paper but I nonetheless keep thinking that Statistics has a role to play in this analysis. Granted, all records are outliers and unique events and depending on a single individual and the result of a combination of “lucky factors” (the victory of the US natation relay team over the French team by an hair, while being amazing, is a bit of a counterexample because so many world records in natation fell in Beijing that it shows that training techniques have reached a level where they count somehow as much as individualities), but they still are random events and thus fall within the domain of Statistics… This is to me the paradox of the Statistics of extreme events: they are so far in the tails of a distribution that nothing seems to be predictable about them and, still, there are statistical models that do a better job than sheer expert opinion, as shown by the paper of Tony O’Hagan et al., almost next to yours in Significance. While being nowhere near an expert on sport Statistics (my only brush with it was an attempt at modelling of VO2 max curves), or even on extreme events, I wrote this entry because I thought it did not fit in Significance whose aim is to open laymen to the appeal of statistical modelling…

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.