The pack of riders compete in heavy rain during the 8th stage of the 101st Tour de France cycling race, over 161 km from Tomblaine to Gerardmer La Mauselaine, in France, 12 July 2014. (EPA/NICOLAS BOUVY)

He’s swift, adaptable, and getting better every year. Stage by stage, he takes on the world’s most brutal bike race, complete with grueling climbs and punishing flats. But this Tour de France star isn’t a cyclist, and he doesn’t do his work in a jersey. He’s John Eric Goff, a physicist who uses Newton’s laws and a painstakingly developed model to predict how many minutes and seconds each stage of the race will take cyclists to complete — even though he’s never seen the race in person.

Descriptions of the route from fans on the ground make Goff “green with envy,” but he has to content himself with an Internet feed in an office 4,000 miles away. He may be far away, but the Lynchburg College physics department chair has quite the success rate. During this year’s Tour de France, his predictions differed from the winning times of eight consecutive stages by fewer than 1.85 percent. The secret lies in the very variables that make the race notoriously difficult for bookies who want to assign odds: things like air drag, elevation, and terrain.

Every year, Goff uses the Tour de France’s own publications on each stage of the race to create a complex profile of his own. He gathers published data on elite cyclists’ power output, information about turns and hills, and intel on the newest bike materials for his calculations, which come together with the help of Newton’s laws and inclined planes.

Eric Goff pictured with his student (left) Chad Hobson, who helped with this year's predictions. (John McCormick, Lynchburg College)

For Goff, a bike is simply an object on an inclined plane (better known by its less sexy name, “ramp”). When it’s acted on by things like gravity, friction, and force, the bike starts to move. Goff uses this simple principle as a starting point—the Tour’s twisty Alpine passes are ramps, the cyclists add power, and the road adds friction. Armed with his model, he publishes his prediction to his blog, props his feet up on his desk, and sees whether he was able to predict the peloton’s progress.

Following the Tour de France isn’t just a spectator sport for Goff: it’s been his job since 2003, when a student got him interested in the race. For over a decade, he’s published research on the race and the science of other sports, like World Cup soccer.

In recent years, Goff has had to adjust his model for bigger power outputs (read: faster cyclists). Could doping be the reason? He’s coy. “It’s quite possible that in 10 years cyclists have just improved,” he concedes, noting that better training methods, diets and gear are being developed all the time.

Goff also demurs when it comes to the issue of betting. Though he himself doesn’t bet on the race, he admits that his predictions, which model how an elite cyclist will likely perform during each stage, could be used for gambling. “It would be incredibly naive of me to think that there aren’t people who rely on my blog for a bit of help,” he says, adding that every year, his blog generates e-mails from fans worldwide who imply they do just that.

His model might help others win big, but in true physicist style, Goff prefers to focus on the science behind the cyclists. There’s no telling which real-life conditions each cyclist will face once they’re on the road, he says—or which strategies will be used by teams who want to protect their leading cyclists and who have to deal with things like urine-throwing bystanders and freak hailstorms. After all, he says, the human element can confound even the best model. “This is a three-dimensional world.”

Erin Blakemore (@heroinebook) is a freelance journalist from Boulder, CO. She is the author of The Heroine’s Bookshelf (Harper).