Engineers, not racers, are the true drivers of success in motor sport

“I ALWAYS THOUGHT records were there to be broken,” Michael Schumacher, a star Formula 1 (F1) driver, said in 2013. At the time, his record of 91 career F1 victories looked safe: the closest active racer had just 32. Yet on October 11th Lewis Hamilton of Britain equalled the mark. Mr Hamilton is also on pace to tie Mr Schumacher’s record of seven F1 championships later this year.

Mr Hamilton’s ascent has ignited debate over whether he is F1’s best driver ever. Comparing athletes across eras is always hard—especially in motor sports, where a racer depends on his car. Moreover, F1 has regularly changed its scoring system and its number of races, drivers and teams.

However, statistical analysis can address many of these nuances. We have built a mathematical model, based on a study by Andrew Bell of the University of Sheffield, to measure the impact of all 745 drivers in F1 history. It finds that Mr Hamilton’s best years fall just short of those of the all-time greats—but so do Mr Schumacher’s.

The model first converts orders of finish into points, using the 1991-2002 system of ten points for a win and six for second place. It adjusts these scores for structural effects, such as the number and past performances of other drivers in the race. Then, it splits credit between drivers and their vehicles. (Today, F1 has ten teams, each using two drivers and one type of car.)

Disentangling these factors is tricky. Mr Schumacher spent most of his peak at Ferrari, as Mr Hamilton has at Mercedes, leaving scant data on their work in other cars.

However, their teammates varied. And drivers who raced alongside Mr Hamilton or Mr Schumacher tended to fare far better in those stints than they did elsewhere. If Ferrari’s and Mercedes’ engineers boosted lesser racers this much, they probably aided their stars to a similar degree. Because most drivers switch teams a few times, this method can be applied throughout history.

Between the two racers with 91 wins, the model prefers Mr Schumacher. He won 1.9 more points per race than an average driver would have done in the same events and cars, edging out Mr Hamilton’s mark of 1.8. Limited to their five best consecutive years, the gap widens, to 2.7 points per race for Mr Schumacher and 2.0 for Mr Hamilton.

This difference stems mostly from the impact of their cars. Both stars raced in the finest vehicles of their day. But 20 years ago, cars from Williams and McLaren were nearly as strong as Ferrari’s. In contrast, Mercedes now towers over its rivals, enabling Mr Hamilton and Valtteri Bottas, his teammate, to coast past lesser cars. Before joining Mercedes, Mr Bottas had never won a F1 race. He now has nine victories.

Yet on a per-race basis, the greats of yesteryear beat both modern stars. Three of the model’s top four drivers stopped racing by 1973; the leader, the Argentine Juan Manuel Fangio, won five titles in the 1950s.

These pioneers had short careers. Fangio started just 51 races, to Mr Schumacher’s 306. However, the model is impressed by them, because the impact of cars relative to drivers has grown over time. On average, it assigns drivers in the 1950s 58% of their teams’ points; today, that share is 19%. Fangio, who was a mechanic by training and won titles using cars from four different firms, was known as “the master”. The masters of modern F1 are engineers who sit behind laptops, not steering wheels. ■

Sources: Ergast.com; F1-Facts.com; “Formula for success: multilevel modelling of Formula 1 driver and constructor performance, 1950-2014”, by Andrew Bell et al., Journal of Quantitative Analysis in Sports, 2016;The Economist