This article is from the archive of our partner .

What if runners or swimmers consistently failed to break records? Would fans care that a "virtual stasis" had overtaken their favorite individual sports? At least one researcher, Geoffroy Berthelot, believes this has already happened--twenty years ago. "Athletes' best sprints, best jumps, best throws--many of them happened years ago, sometimes a generation ago," observes Paul Kix at The Boston Globe, who profiles Berthelot's and others' academic inquiries on the subject. According to Berthelot, "the pinnacle of athletic achievement was achieved around 1988." It's all been downhill since then--or, more correctly, a plateau.

In a study published in the academic journal PLoS One, Berthelot plotted out every world record from 1896 onward. "When placed on a L-shaped graph, the record times fell consistently, as if down a gently sloped hill. They fell because of improving nutritional standards, strength and conditioning programs, and the perfection of technique," details Kix. "But once Berthelot's L-shaped graphs reached the 1980s, something strange happened: Those gently sloping hills leveled into plains. In event after event, record times began to hold." Today, 64 percent of track and field records have stood since '93.

Really? But what about all those track and field records broken by the likes of Usain Bolt? "Bolt is a very particular case," Berthelot told the Globe. "All the media focus on Usain Bolt because he's the only one who's progressing today." By his calculations, Berthelot "predicts that the end of almost all athletic improvement will occur around 2027" a year that he says the "human species' physiological frontiers will be reached."

Kix ponders this potential outcome and then poses this question in response: "What happens to the athlete who knows there are no records left to break?"

This article is from the archive of our partner The Wire.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.