Hamilton? Schumacher? Senna? A machine learning algorithm reveals the fastest F1 driver in the last 40 years. The history of F1is made up ...
Hamilton? Schumacher? Senna? A machine learning algorithm reveals the fastest F1 driver in the last 40 years.
The history of F1is made up of endless debates about the talent of individual drivers. The oldest have known the Homeric battles between Prost and Senna, Schumacher and Hakkinen, or more recently, Hamilton and Rosberg/Vettel.
Usually there are three justices of the peace: the number of victories, the number of pole positions, and of course the number of world championship titles. However, some drivers have never won anything, and yet the study of their lap times shows how fast and skilful they are, sometimes more than some reigning champions. Only if F1 is all about high speed, it's not just that, and the best are those who manage to repeat performance race after race, season after season, with metronome consistency, resisting all pressures. In F1 perhaps even more than elsewhere, for equal talent it's all a matter of mind, and finding the fastest driver of all time is not easy.
Except perhaps for an artificial intelligence.
That's the idea behind this study conducted by the official organization that runs the F1World Championship.
The aim of the study is to compare drivers over four decades in order to identify the "fastest" drivers based solely on pure qualifying speed. The sponsor used Amazon Web Services (AWS) technology to arrive at the conclusion and ranks all drivers since 1983 while removing the relative performance of the car from the equation.
AWS - which also provides real-time information on tyre condition during races - says it used machine learning to create "an objective, complex and data-based driver speed ranking.
The top ten list does not include the then World Champions, including Alain Prost, Nelson Piquet, Nigel Mansell, Mika Hakkinen and Kimi Raikkonen, who spent much of their careers in midfield teams. Based on this data, it also gives a time differential between the drivers as if they were participating in a qualifying session.
Champions (and stars) are not necessarily the fastest
Unsurprisingly, those reputed to be the greatest champions in the history of F1are trusting the podium, with the legendary Senna taking first place, confirming arithmetically 30 years on that his aura as an ultra-fast and unusual driver was not usurped. Among the drivers still active, the Dutchman Max Verstappen is fourth and the Monegasque Charles Leclerc seventh. Both started their careers recently and are only 22 years old. Alain Prost, despite his four world championship titles, does not figure in this top 10, as he was above all a strategist gifted with an exceptional racing intelligence that allowed him to win by rarely being the fastest.
Here is the ranking. It still contains 6 world champions, and it restores a bit of justice to drivers who may have been a bit underestimated during their career.
1. Ayrton Senna - 0.000
2. Michael Schumacher - +0.114
3. Lewis Hamilton - +0.275
4. Max Verstappen - +0.280
5. Fernando Alonso - +0.309
6. Nico Rosberg - +0.374
7. Charles Leclerc - +0.376
8. Heikki Kovalainen - +0.378
9. Jarno Trulli - +0.409
10. Sebastian Vettel - +0.435
F1 does not provide exact details on how the results were obtained, but the system purports to use a matrix of data to compare teammates to each other over time and then link this data to other teammates over the course of a driver's career.
"By comparing teammates in qualifying sessions, the machine learning-based tool focuses on a driver's performance, building a network of teammates over the entire period, all linked to each other and therefore comparable. By comparing lap times between teammates only, the algorithm of the fastest driver effectively normalizes the performance of the car and the team. Overall, this provides a picture of how drivers of different generations compare to each other by analysing the purest indication of raw speed - the qualifying lap," explains an F1 press release.
Rob Smedley, a former Ferrari and Williams engineer who now works for F1, said teams use similar studies to decide who to sign as a driver:
"In the team environment, this type of modelling is used to make key decisions about driver selection. As drivers are most often the team's most valuable asset, it is important that the selection process is as robust as possible. "Such a process would therefore be deployed by F1 team strategists to present the most objective and evidence-based selection possible. The fastest driver allows us to build a comparative table of drivers by analysing the purest indication of raw speed, the qualifying lap - and it is important to note that this pure speed is the only element of the vast arsenal of drivers we are analysing here, to highlight the fastest drivers ever encountered, which is very exciting".
As a motorsport website points out, there was no mention of Colombia's Juan Pablo Montoya, whose average speed of 262km/h during qualifying at Monza for Williams in 2004 was the fastest lap in F1history until Ferrari driver Kimi Raikkonen - another omission - reached an average of 263km/h in 2018.