The data vs driver conflict at the heart of F1's ground-effect cars

For all Formula 1’s image of being a gladiatorial contest between drivers, it remains very much a data-driven series. Teams readily invest millions in telemetry, data analysis, machine learning tools, the latest computer systems, simulations and simulators because at the root of car performance are cold hard numbers.

No matter how much the cars have changed over the years, the law of physics haven’t - and it is tapping into the brainpower of processors and systems that is key to providing a team with the set-up answers, performance profiles and data points needed as it heads into each race weekend.

The process feeds itself. Teams collect as much data as they can because the more they provide, the better the responses they get back.

Let’s not forget that engineers and drivers like nothing more than having all the answers in front of them to maximise car performance, and minimise the potential for nasty surprises or a bit of jeopardy getting thrown into the system. Normally, if the computer says no, then you don’t do it.

But, as the F1 pitlane begins to understand the current generation of ground-effect cars, there is a fascinating divergence opening up between what the data says and what actually works out on track. The first inkling of this came last year as teams rapidly found that the abundance of downforce that they could get in the wind tunnel by running their cars super close to the ground was not repeated in the real world.

The end result was a significant amount of porpoising that overshadowed a lot of the season for a lot of teams before they finally got a handle on just what they needed to do to find a sweet spot of peak performance without terrible bouncing.

Mercedes was one to badly suffer with porpoising as simulations didn't match reality last year

Photo by: Carl Bingham / Motorsport Images

As Mercedes chief technical officer Mike Elliott said: “If you were to go back to the old regulations, you could put the car where you wanted to put it. You had big travel in the suspension, which allows you to shape the balance better through the corners. You weren't limited by stiffness, and you could chase where the aerodynamic performance is in the regulations.

“With these cars, aerodynamically they want to run close to the ground. But, if you run them close to the ground, you have to run them stiff. And that's one of the traps we fell into last year, if we're honest.

“So, there's always going to be that balance you have on this set of regulations with cars that want to run really close to the ground: how do you get that balance right?”

"It could be that the effect of ride in simulation is not exactly matching what is the feeling of a driver on the real tracks" Andrea Stella

The FIA’s move to alleviate the risks of porpoising through the lifting of floor edges by 15mm this year appeared to change that balance a bit and left some teams a little more comfortable to push the boundaries once again.

And, indeed, a lot of the time the computers have been telling teams that the best way to run the cars is low and stiff – as it maximises the sweet spot for peak downforce and keeps the car in that ideal window.

But this theoretical ideal is still not a perfect match for reality. Super stiff cars do not like the bumps and kerbs that are part and parcel of circuits. And there can also be negative consequences from such an approach when it comes to tyre degradation and driver confidence.

Raising of the floor edges for 2023 played a part in McLaren's early-season struggles

Photo by: Erik Junius

As McLaren team boss Andrea Stella explained: “When you talk about what the data says, it's ultimately some aerodynamic maps that you produce from the wind tunnel. You run them in the aerodynamic simulations and these simulations, they tell you: this will be the right height profile over a lap, that gives the most downforce on average and the best lap time.

“This could come, for instance, with set-ups that are very stiff, because then you can control the right heights. And this level of stiffness requires a compromise of what is the right performance on bumps and so on.

“Then you need to lean on the simulation to try to take into account the effect of ride on lap time, and it starts to become quite complex. Then it could be that the effect of ride in simulation is not exactly matching what is the feeling of a driver on the real tracks.

“There's also a degree of adjustment that you need to do when you put the car on the ground, and you see how stiff you can actually run it. If conditions are very high grip, then you can afford to run slightly stiffer. But, if the grip is low, you need to make the car more compliant from a stiffness point of view.

“These are normal activities, I would say, for any team. And definitely, in simulation, you would run a stiffer car than you can actually run on track.”

There is another element to throw into the equation too: the driver. The current generation of cars are not especially nice to drive, as they do run stiff and they constantly seem to be on a knife’s-edge when it comes to being in and out of the balance window.

These characteristics mean that, sometimes, what the data says is the best approach to car set-up is at odds with what the driver wants. A super stiff low car can be more punishing on its tyres, more sensitive to changing conditions, and much more difficult to drive consistently as tyres degrade and fuel burns off.

Ferrari is now taking Sainz's feedback on board rather than solely relying on what the data suggests

Photo by: Sam Bloxham / Motorsport Images

It was this conundrum, between doing what the data said or ignoring it and focusing on driver needs, that is understood to have been at the core of Ferrari turning the corner on its race-day inconsistencies this year.

As the squad uncovered some inherent flaws in its SF-23, which was great on a single lap but struggled over race distances, the chase for answers was on. Recent form, including a podium finish in the Belgian Grand Prix, suggests that it has indeed unlocked the answers needed to address things.

Part of its solution has come from a change of aero concept, as it shifted away from its in-wash bathtub solution of sidepods. But it is also understood that some of its improvements came from a different approach to setting up the car.

While, earlier in the year, the team had followed the data in running its car low and stiff, because that was theoretically the fastest, a request by Carlos Sainz to ignore the computer and run things higher and softer to make it more compliant appeared to pay off.

"Your simulator and your simulations are not going to tell you precisely what that confidence level is giving you, and maybe that's the next development down the line" Jock Clear

The downside of less theoretical peak downforce was more than made up for by increased driver confidence, less tyre degradation thanks to improved handling under braking and in corners, plus much greater consistency as the car was more compliant.

That shift, in trusting the feedback from the driver over what the data says, is a bold step for any team to make – but one that simply cannot be ignored now.

But there is also an interesting flipside to putting the focus back on the driver rather than the numbers. It is that this is ultimately going to put an even bigger reliance on getting the data right.

Teams now need to feedback what the driver is saying into data produced from the wind tunnel

Photo by: Toyota Racing

As Ferrari’s Jock Clear admitted recently, the path to success will be in inputting driver feedback more into the data loop so that the computer’s answers are more in line with the real world.

“I think it's one of those things where you have to decide which direction you are going to go in,” he said. “Your simulator and your simulations are not going to tell you precisely what that confidence level is giving you, and maybe that's the next development down the line - to really get the driver simulators with really good fidelity, where the driver confidence is clearly evident.

“Then you can drive that development quantifiably. At the moment, you have to make a bit of a judgement call on: ‘OK, we can bring this much more downforce, but the shape of the math is probably not going to be as easy to drive.’

“So how do you evaluate the two? The driver is the only one that's going to be able to tell you that and, as such, you have to rely on a combination of the simulator, the track time, and simply anecdotal [feedback].”

That may be the case now, but don’t imagine for a second that, in the fast-moving world of F1, teams will not quickly fill in that missing gap between the drivers and the computers. After all, the data always wins in the end.

Even despite recent difficulties, data is always going to play a vital role

Photo by: Sam Bloxham / Motorsport Images