Man vs. Machine: What Makes a Formula 1 Champion?

Taking a look at former Formula 1 champs, joining in on the everlasting debate with a brief analysis of historical race results.

Photo by Carl Jorgensen on Unsplash

If you’re a fan of Formula 1, you‘re probably aware of the classic Car vs. Driver debate. What makes a champion? A driver who outclasses their peers with superior skill and craft, or the carefully engineered multi-million dollar car?

Many have thrown their hat into this debate. Former Mercedes Engineer Paddy Lowe contends that current champion Lewis Hamilton’s success in recent years has more to do with the driver’s talent than he is given credit for; Daniel Ricciardo, an F1 driver himself, believes that 75% of F1 achievements can be accredited to the car.

Researchers have applied complex statistical models in attempts to separate driver performance from their team and their car, in order to compare driver performance across teams and across eras.

Here, I propose a metric which measures how much a champion driver contributes to their championship win.

Proposed Metric

The problem with attempting to tackle the Car vs. Driver debate using numbers is that the driver and the car are inseparable. Points scored, lap times, or any other statistic are calculated with the results achieved by the car with the driver at the wheel. So how do we separate the Man from the Machine?

We compare him to the other guy sitting in an identical car. Every F1 champion has had a teammate who drives a car designed and created by the exact same team. As such, any difference in performance can be attributed to the difference between the people behind the wheel.

The formula for the proposed metric.

The above is the formula for the “Driver’s Influence” or DI; P_D is the number of points the driver scored in a given race, P_T is the number of points the teammate scored, and P_Max is the number of points the winner of the race scored. Where the driver had multiple teammates, P_T is calculated using the average.


For one race, this metric ranges from -1 to 1; however, barring a few exceptions, the average DI for the winner of the Driver’s championship over all races over the course of one season is expected to be positive (otherwise the teammate would be the champion).

Calculating the champion’s DI over 1 season, a DI near 1 would mean that the driver consistently finished near the top every race, while their teammate consistently finished near the bottom. In other words, their success is 100% attributable to their own skill, since their teammate, who had the same car, was uncompetitive. On the other hand, a DI near 0 means that the champion and their teammate performed similarly over the season, suggesting that their success is more attributable to the car.

More discussion around the suitability of this metric is discussed later.

Historical Data

Using data from the 1950–2020 F1 seasons, I calculated the average DI for each year’s champion, averaging over all the races of the season. DNFs are counted as 0 points. The races in which the champion did not participate are ignored.

For each season, the season’s championship winner’s DI is calculated and averaged over all races. We observe a downward trend in the amount of the driver’s contribution to the championship win as the years progress.

We can see from the plot above that there is a very slight downtrend in the mean DI (least squares estimate plotted in red).

Each of the points in the plot above is the average of all races in that year; and as such, each point can be thought to have its own level of confidence which is inverse to the variability of the DIs of each individual race in that particular year. Using this, we can create a weighted least squares estimate (in blue).


Suitability of DI as a metric

DI is a summary statistic of a champion’s performance against someone with an equal car in the race. As such, it provides a control treatment against the performance of the driver. However, an obvious shortcoming of this metric is that it is sensitive to the teammate’s performance; a poor teammate might inflate the DI, while a competitive teammate might deflate the DI. As such, using this metric requires the assumption that the teammate’s achievements in the car are representative of an average driver on the grid, if they had the same car. The validity of this assumption is probably quite controversial.

DI is also sensitive to the scoring system used in a particular season, limiting its comparability across seasons. On the other hand, we are trying to determine the importance of a driver’s contribution in winning a championship, and a championship is won by scoring points in a particular season, under the rules of that season. As such, arguments could be made both for and against some adjustment to standardize the DI across different seasons.

There are two years where the champion’s average DI was actually negative; this is because of fierce competition between teammates Niki Lauda and Alain Prost in 1984; and teammates Ayrton Senna and Alain Prost in 1988. In these seasons, not every race counted towards the championship. These three drivers are known as some of the greatest in the history of F1; was it because of their McLaren, or simply because they were great drivers? DI would suggest it is because of their car.

Photo by Sven Brandsma on Unsplash

Lastly, holding the numerator (points scored over teammate) constant, DI decreases as the relative finishing position falls. That is, if the driver scored 10 points more than their teammate in 1st place, their DI is higher than if they scored 10 points more than their teammate in 4th place. Again, the desirability of this behaviour is debatable. For example, if mechanical problems occur during a race, affecting both cars, the DI will decrease. On one hand, this highlights the importance of having a good car, and the decrease in DI is justified. On the other hand, such a situation does not diminish the driver’s importance, in principle.


It is quite apparent that in winning Formula 1 championships, a good driver is important, but a good car is indispensable. The driver’s role seems to be increasing overshadowed by the flashy cars that come out every year.

Is this surprising? Perhaps, perhaps not. On one hand, as teams develop more advanced technologies, a well developed car can truly stand out. After all, it is no secret that F1 is a two-tier sport. Of the top teams’ payments from F1, only about 30% or less go towards the drivers’ salaries; it makes sense that they are paid in proportion to their importance.

Does this detract from the enjoyment of the sport? I don’t think so. It’s perfectly fine to enjoy the sport for the amazing innovation and feats of engineering without paying much attention to the personalities of the individual drivers. F1 continues to enjoy a rapidly growing fanbase, in light of our findings today.

Code used to calculate and visualize the DI over time can be found here.

mostly for fun