Last June, Amazon Web Services (AWS) revealed that F1 was moving the vast majority of its infrastructure from on-premises data centres to AWS, and standardizing on its machine learning and data analytics services to accelerate its cloud transformation.
Working with AWS, F1 was seeking to enhance its race strategies, data tracking systems, and digital broadcasts through a wide variety of AWS services – including Amazon SageMaker, a fully managed machine learning service that enables developers to build and deploy machine learning models and thereby uncover never-before-seen metrics intended to “change the way fans and teams enjoy, experience, and participate in racing”.
Using SageMaker, it was revealed that F1’s data scientists were training deep learning models with more than 65 years of historical race data, from which they could extract critical race performance statistics to make race predictions and give fans insight into the split-second decisions and strategies adopted by teams and drivers.
Speaking at the AWS re:Invent learning conference in Las Vegas, F1’s technical boss, Ross Brawn, revealed the initial fruits of their labours as he introduced a number of new graphics aimed at enhancing the fan experience.
“For next season we are expanding the ‘F1 Insights’ for our viewers, by further integrating the telemetry data such as the car position, the tyre condition, even the weather, so we can use ‘Sagemaker’ to predict car performance, pit stops and race strategy,” said the Briton. “There will be some exciting new AI integrations into next year’s F1 TV broadcast.”
Introducing the ‘car performance’ graphic, he said: “We know that somebody is in trouble: his rear tyres are overheating. We can look at the history of the tyres and how they have worked and where he is in the race, and machine learning can help us apply a proper analysis of the situation.
“We can bring that information to the fans and make them understand if the guy is in trouble or if he can manage the situation. These are insights the teams always had but we are going to bring them out to the fans and show them what is happening.”
“Wheel-to-wheel racing is the essence, a critical aspect of the sport,” he continued, referring to the ‘overtake probability’ graphic, “and now with machine learning and using live data and historical data, we can make predictions about what is going to happen.
“The graphic on the right shows what we expect is going to happen in this event. What is great about this, is that the teams don’t have all this data. We as F1 know the data from both cars and we can make this comparison and this has never been done before.”
Finally, he introduced the graphic dealing with pit stop strategy.
“Stopping at the right time and fitting the right tyre can win or lose a race,” he said. “We are going to take all the data and give the fans an insight into why they stopped and when they stopped – did the team and driver make the right call?”
Which is all very well, but other than the obvious question mark over the reality of overtaking as opposed to the “probability”, one feels that this will appeal to a narrow range of fans and unlikely to impress casual fans.
At the same time, with the sport increasingly disappearing behind paywalls…