How McLaren found the right formula for F1 success with Splunk

Chris Middleton Profile picture for user cmiddleton January 18, 2023 Audio mode
Summary:
Motor racing legend McLaren and its data analysis partner Splunk discuss how they make cars go faster – in both the real and virtual worlds

An image of a McLaren Formula 1 car on the track
(Image sourced via McLaren)

British Formula One and automotive giant McLaren is celebrating three years of its partnership with US machine data specialist, Splunk. But why is such a relationship so important in a world that, on the face of it, is all about speed, engineering excellence, and driver skill?

The answer is that F1 is no longer just about the physical realm, where data is critical, but also about its virtual and simulated counterparts. In all these environments, racing generates a phenomenal amount of data about a car’s performance on the track in all weather conditions, which can lead to race-winning decisions and improvements. 

In a market that is the definition of fast moving – not just car speeds of over 200mph, but also constantly changing regulations and technologies, plus new safety and sustainability/circularity demands – even marginal improvements can mean the difference between being first or last on the grid.

As a result, today’s physical F1 cars are packed with sensors, monitoring every aspect of these fine-tuned, multimillion-dollar machines. James Hodge is GVP and Chief Strategy Advisor for Splunk. He says:

On the real racing cars there are initially about 300 sensors, but that gets reduced down to about 120 over free practice, qualifying, and the race, because of weight. And because they won't do anything in real time with a lot of that data during the race. You only keep a sensor on if you're going to make it actionable.

Ed Green is Head of Commercial Technology at McLaren. He explains:

We were one of the first teams to run sensors on cars. We started out with, say, 20 monitoring low-level telemetry, maybe temperature or pressures. And now up to 300 measure everything from air pressure all the way through to gearbox temperatures and ride height. 

And from those 300 channels of data – some are reading at 10 hertz, and others at 60 to 80 hertz – you can synthesize data together and start getting data points that you're not physically measuring yourself.

Throughout McLaren's history, we’ve pioneered and led the way in this, and that's inspired work across other industries as well. But it's also really helped us to go fast! 

This sport has got such fine margins – a percentage-point difference between first and last positions on the grid. Sometimes that's seconds, but sometimes it’s hundredths between lap times, or milliseconds. In a world like that, you've got to use tech and you've got to use data to help you see the way forward and how you're going to improve.

But at some point, might the laws of physics intervene, and it will no longer be possible to make a car go faster on a given track? Green says:

That's a really good question. I'd say no, because there are so many regulations. Two years ago, the aerodynamics of the Formula One car were designed to push the car into the road. But this year, we had a regulation change, which means we've now got to suck the cars onto the road, in effect. So, might we one day come up against the laws of physics? Maybe, but only in a totally unregulated world!

In the 21st Century, these issues increasingly apply in simulations and in competitive esports just as much as they do on the tarmac at Grands Prix, with major F1 brands having a presence in all these realms and more. Indeed, it could be argued that physical events are almost becoming brand ambassadors for their virtual, esports counterparts, which attract millions of fans from a broad gaming demographic. 

McLaren – winner of 12 physical drivers’ championships and eight constructors’ titles since 1974, after debuting at the 1966 Monaco Grand Prix – has been a prime mover in all these spaces. For example, while McLaren Racing came fifth in the real-world Constructors Championship in 2022, the McLaren Shadow team is the reigning Formula One Esports champion, the virtual racing series that launched in 2017.

Lindsey Eckhouse is McLaren Director of Licensing and Digital Products. She says:

What is unique about F1 Esports, as opposed to other esports, is how similar it is to the real-life product, to Formula One. There are a lot of shared insights and shared experiences, plus a shared approach, or one that we can shadow. That's where the name McLaren Shadow comes from. 

And that's why a partner like Splunk, who partner with the Formula One team and offer a variety of technical solutions to them, are critical to us as an F1 Esports team as well. As with McLaren Racing, we take Splunk’s Data to Everything platform, with custom dashboards, and are able to analyze real-time data that ultimately improves our F1 Esports performance.

Blurred lines

The Woking, Surrey-based automotive company began its partnership with the San Francisco-based machine-data specialist just before the pandemic hit. COVID not only shut real-world events for an extended period, of course, but also caused a spike in popularity for F1’s esports counterpart. 

Today, Splunk is the bigger company, with revenues of $2.67 billion against the $771 million of the McLaren Group (of which McLaren Racing is a subsidiary). Proof that even in the heady, big-money world of F1 racing, data is now the real money-spinner.

So how linked are real-world F1 racing and esports from a technology perspective? 

In a sense, the F1 simulator rigs are the bridge between the real and virtual environments. F1’s star drivers increasingly train in these simulators – combinations of the virtual and the modelled physical worlds – before taking the new season’s car out on a physical track. Terabytes of data might be gathered from simulated laps on the globe’s most challenging circuits, information that could save the driver’s life in the real world, as well as win races.

McLaren’s Green adds:

What's amazing is when drivers join the team and we put them in the simulator, we've got historic data too. We can see how drivers throughout history would have raced in those conditions.

As for esports, Eckhouse explains:

What we do is analyze the laps done in the simulator rig, but from a gaming perspective, and use those insights to inform our race strategy. Just like the Formula One team analyzes their real-life laps and the data points that come through the track, using those insights to develop their race strategy in the real world.

It’s a similar approach, but in esports ours is fully focused on the rigs. Yet the blurring of the physical and the virtual is one of the nuances of Formula One and F1 Esports, in that the game is incredibly similar to the real-life racing experience.

In both areas, Splunk has been integral to our overall practice regimen. The real-time data dashboards they've created for us from an F1 Esports perspective, quite genuinely led to us winning the championship last year and helped our driver Lucas Blakely win the individual F1 Esports Series Pro Championship.

Virtual crossover

So, do people ever cross from the virtual world into the real one? In other words, could McLaren identify esports talent and train them up to be real F1 drivers? Eckhouse says:

That’s a hypothesis we would love to test more, and understand if that’s a reality we can make happen. But Lucas is a brilliant real-life driver too, and has beaten Sebastian Vettel on a real-life track.

In both realms, analyzing data is critical for success, concludes Splunk’s Hodge:

It’s about good stats and predictive modelling. And we're starting to look at more areas for improvement. So, as we’re improving lap times, what do we think is the ultimate place we could get to? Where is the best lap possible – looking at different tires and aerodynamic setups versus others? How long is this tire going to last in these conditions? 

That's the kind of the area we're getting into now, plus what is the best type of practice to get ready for a Grand Prix? On race weekend, the McLaren team have done all of their ‘What if?’ planning, so they are, as much as possible, best prepared or resilient.

But of course, there are key differences between the real and virtual worlds, he says:

There’s a lot more complexity in the physical one, because physics and practicalities are more in play. There are race locations all around the world. Every race weekend, McLaren moves in an IT rig that always follows the cars. It goes in the same plane and takes the telemetry that gets streamed off them from the track. 

It also gets streamed over the Internet back to the McLaren Technology Centre. So, the first thing we do for them is make sure that's working perfectly, because it’s mission critical: you're not allowed to start up a Formula One engine or keep it running without that telemetry, largely because of safety concerns. 

They have 32 race engineers, over a weekend, sat in a calm environment making decisions that get relayed back to the trackside. So, they're running blind without that data. Plus, the secondary benefit is financial. These engines are very expensive, and you only have a few per season.

But he adds:

“The esports team gives us a lot more freedom to rapid prototype and put things into rapid development, whereas with the Formula One team there's so much more complexity, there's cost cap considerations. There’s a lot more deliberate intent, if you like.

A great example is supercomputing capacity. You are limited per season in how much you can do. If you run, say, a computation for an hour, and it didn't write the results to the database because you’ve run out of disk space, then you've lost an hour you can never get back. So, our foundation is resilience. 

For us, it’s about the marginal gains that we can focus on to build that resilience for McLaren, to get better decision-making and mission-criticality. That’s where digital systems are so important. And when you get that right, you can move on to the next thing, and the next, and start to foster a data culture. 

Before, it was ‘We just need to get great drivers and make them practice.’ But now, it's ‘How do we use data to make a more efficient decision on how to train better – on how to do things better?’

My take

Welcome to the new world of racing: fast, accurate data means faster, better, safer cars: bit by achievable bit.

Loading
A grey colored placeholder image