At the Olympics, how fast is too fast? That question has dogged Chinese swimmer Ye Shiwen after the 16-year-old shattered the world record in the women's 400-metre individual medley (400 IM) on Saturday. In the wake of that race, some swimming experts wondered whether Ye’s win was aided by performance-enhancing drugs. She has never tested positive for a banned substance and the International Olympic Committee on Tuesday declared that her post-race test was clean. The resulting debate has been tinged with racial and political undertones, but little science. Natureexamines whether and how an athlete's performance history and the limits of human physiology could be used to catch dopers.

Yes. Her time in the 400 IM was more than 7 seconds faster than her time in the same event at a major meet in July. But what really raised eyebrows was her showing in the last 50 metres, which she swam faster than US swimmer Ryan Lochte did when he won gold in the men’s 400 IM on Saturday, with the second-fastest time ever for that event.

No, says Ross Tucker, an exercise physiologist at the University of Cape Town in South Africa. Athletes are much more likely to dope while in training, when drug testing tends to be less rigorous. “Everyone will pass at the Olympic games. Hardly anyone fails in competition testing,” Tucker says.

Out-of-competition tests are more likely to catch dopers, he says, but it is not feasible to test every elite athlete regularly year-round. Tracking an athlete over time and flagging anomalous performances would help anti-doping authorities to make better use of resources, says Yorck Olaf Schumacher, an exercise physiologist at the Medical University of Freiburg in Germany, who co-authored a 2009 paper proposing that performance profiling be used as an anti-doping tool¹. “I think it’s a good way and a cheap way to narrow down a large group of athletes to suspicious ones, because after all, the result of any doping is higher performance,” Schumacher says.

The ‘biological passport’, which measures characteristics of an athlete’s blood to look for physiological evidence of doping, works in a similar way to performance profiling (see 'Racing just to keep up'). After it was introduced in 2008, cycling authorities flagged irregularities in the blood characteristics of Antonio Colom, a Spanish cyclist, and targeted drug tests turned up evidence of the banned blood-boosting hormone erythropoietin (EPO) in 2009.

Anti-doping authorities need a better way of flagging anomalous performances or patterns of results, says Schumacher. To do this, sports scientists need to create databases that — sport by sport and event by event — record how athletes improve with age and experience. Longitudinal records of athletes’ performances would then be fed into statistical models to determine the likelihood that they ran or swam too fast, given their past results and the limits of human physiology.

The Olympic biathlon, a winter sport that combines cross-country skiing and target shooting, has dabbled in performance profiling. In a pilot project, scientists at the International Biathlon Union in Salzburg, Austria, and the University of Ferrara in Italy, developed a software program that retroactively analysed blood and performance data from 180 biathletes over six years to identify those most likely to have doped². The biathlon federation now uses the software to target its athletes for drug testing.

“That would be unfair,” says Tucker. “The final verdict is only ever going to be reached by testing. It has to be.” In recent years, cycling authorities have successfully prosecuted athletes for having anomalous blood profiles, even when banned substances such as EPO could not be found. But performance is too far removed from taking a banned substance and influenced by too many outside factors to convict someone of doping, Tucker says. “When we look at this young swimmer from China who breaks a world record, that’s not proof of anything. It asks a question or two.”

1. Shumacher, Y. O. & Pottgiesser, T. Int. J. Sport. Physiol. Perform. 4, 129–133 (2009).

2. Manfredini, A. F. et al. J. Sport. Med. Phys. Fit. 51, 153–159 (2011).