This article was first published in Rouleur Issue 142

Paris-Roubaix has always resisted progress. Its jagged cobbles, relics of a pre-automotive France, will smash into a thousand pieces anything that’s not built to withstand the most primitive of pummellings. When roads started to become tarmacked in the 20th century, the organisers diverted the race onto long-forgotten, shockingly neglected stretches of pavé deep in the woods to preserve the race’s anachronistic, attritional character.

It favours riders who are more Classical hero than modern racing cyclist. Since 1896, the Hell of the North has crowned the strongest, the bravest, the most courageous – only those capable of enduring more than anyone else would lift the famous cobblestone trophy above their mud-caked, sweat-streaked, sometimes bloodstained heads in the centre of the Roubaix Velodrome.

That’s why Fabian Cancellara was nicknamed ‘Spartacus’. The most dominant Classics rider of his generation, he embodied that gladiatorial ideal completely. When he attacked from 40 kilometres out in 2010, or muscled to the sprint win in 2013, there were no marginal gains in evidence – it was about a rider so powerful he could literally bend the race to his will.

Okay, we already know all this. But fast forward a decade and the team that Cancellara founded and co-owns chooses not to ride Paris-Roubaix at all. Why?

Tudor Pro Cycling were invited to the race in 2024, a year after stepping up to Pro-Team level. For most teams – especially one owned by a three-time winner and Classics legend – it would have been a no-brainer. Instead, they declined, and it had everything to do with brains, as Kurt Bergin-Taylor, Tudor’s head of innovation, explains. “We said no. We weren’t ready technically to take on such a race. That was a big decision.”

That decision, and what followed, tells us everything about how Paris-Roubaix – and possibly the whole of professional cycling – is changing.

Bergin-Taylor would not object to being described as more geek bod than Greek god. He has a PhD, he is in his thirties, bespectacled, balding, very articulate. “A lot of people would say, ‘It’s Fabian, you have to do it’. But Fabian was really clear. We go there to compete, not just to take part.”

So in order to be competitive for the 2025 edition, Tudor started testing in October 2024. Matteo Trentin, one of the peloton’s most experienced Classics riders, was brought in to set a baseline or reference point. “We had a wide bucket of testing,” says Bergin-Taylor. “We were looking at wheels, tyres, pressures… quite quickly we realised there was a lot of individualisation needed – both from what the data suggests and also from rider feedback.”

It was clear the old ways of setting up bikes for Roubaix – ‘historical’ things like rider weight, tyre width, pressure charts that mechanics originally devised around FMB or Dugast tubulars – were no longer sufficient. “Riders have their own technical skillsets,” Bergin-Taylor says. “The way one rider goes around a corner is completely different from another.”

Tudor’s solution was to build their own data-collection ecosystem from scratch. “First we thought, do we make a test bike? But then you've got to take it everywhere and it only works for one rider. So we said: what about we make a system we can plug onto any bike and rider and upscale the data we can collect?”

Normally, on-bike data is recorded at one hertz – one data point per second – but if you’re riding at 15 metres per second over the cobbles (54kph), a lot happens in that 15 metres, Bergin-Taylor expains. So Tudor created a system that samples at much higher frequencies – sensors reading up to 15,000Hz that can plot acceleration, deceleration, lean angle, gyroscopic movement and rider input. Inertial measurement units (IMUs) were placed across the bike and even the rider with a central hub mounted behind the saddle – the “brain” of the system.

It was so bespoke that they built their own circuit boards, wrote the software, solved the “nightmare” problem of time-aligning a massive amount of data from multiple sensors. “The hard part isn’t collecting data,” Bergin-Taylor says. “Anyone can do that. The really big challenge is syncing it.”

This allowed them to correlate subjective rider feedback with objective, high-resolution data about how the bike-tyre-rider system behaves under stress. Corner entry speed, lean angle, braking mid-corner – all of those things that one rider does differently from the next radically alter how a tyre deforms, how a rim is loaded, how vibration travels through the system and what it does to forward motion: “Now we can actually see this as data.”

Next, how to model for the race itself. Paris-Roubaix is not just one problem to solve, but several stacked on top of each other. The first 95 kilometres of a total 259 are on smooth, flat roads. Then the first cobbled sector, Troisvilles à Inchy, with 29 more sectors that get progressively harder, with the hardest of all, the Trouée d’Arenberg, 95 kilometres from the finish.

“Most modelling is based on a time-trial,” Bergin-Taylor says. “How to get to the finish as fast as possible.” But, he explains, a better approach is to ask when the decisive moments are likely to happen. Those are invariably on the cobbles. “You might save two watts on the smooth, but if you lose 50 watts on the cobbles you’re no longer in the race. We’re looking at compliance as a whole, how you build the correct compliance level within the system. At the moment tyre pressure is the low hanging fruit.”

Wider tyres at lower pressures dramatically reduce vibration losses and improve control on cobbles while the penalty on smooth roads is surprisingly low. “There’s a negative cost but we found it wasn’t that large. Not compared to the positive effect on cobbles.” However: “The handling is different, there’s a weird sensation when you have really low pressures on smooth roads and the possibility of the tyre coming off the rim is different, so you build your mitigation strategies around that.” Bergin-Taylor describes a pressure “window” that varied from rider to rider depending on their confidence, bike handling and how aggressively they attacked the pavé.

Equipment choice is, of course, directly related to these insights, but one of the more surprising choices, at least to an onlooker, was Tudor’s use of the DT Swiss GRC 1100 DICUT at Roubaix – nominally a gravel wheel, fitted to their BMC Teammachine R aero race bikes.

“It wasn’t our intention,” says Ralf Eggert, DT Swiss’s road marketing manager. “Normally, from our positioning, we would never say the GRC is the better endurance wheel.”

On paper, the ERC – DT Swiss’s endurance road wheel – made more sense. But in real-world testing, the GRC was faster for the Tudor riders at Roubaix.

“The GRC looks a bit different,” says Eggert. “There’s a curve in the rim shape that blends with a bulky, high-volume tyre. In the wind tunnel we used 40mm tyres to optimise them for gravel racing. When you have knobs on the side, they usually destroy the aerodynamics of the rim. You don’t get any sailing effect. But with this profile we managed to get a better sailing effect and better aero for gravel racing.

And as soon as you put road tyres on, they work even better.”

The GRC’s wider 24mm internal width also pushed tyre choice towards the 30-32mm range – a sweet spot for Roubaix – and its robustness offered better protection against rim strikes at low pressures and high speeds. “We didn’t want to stand here talking about failures on the pavé,” Eggert says. “Safety is our main concern.” He confirms that in the 2025 race there were no wheel failures, no punctures and none of the spares at the feed zones were needed.

Although the UCI rules don’t allow 40mm tyres to be used, would bigger be even better for Paris-Roubaix? Current trends in road cycling would suggest they could be, but Eggert says not: “More volume means more comfort, that’s one dimension, but the other is when you increase tyre width, you increase frontal surface area, which increases drag. With our new ARC wheels launched in 2025, we increased the rim width to 22mm and many people said, ‘oh that’s not wide enough’.

“But JP Ballard of Swiss Side [DT Swiss’s aerodynamics partner] always tells us to find a balance between frontal drag area and comfort. Even though teams might now say wider might be better for riding inside the peloton where aero is not so important, it always depends where you sit in the race. If you are out in a breakaway, aero is getting more important.”

Unlike the Tour of Flanders, Paris-Roubaix is flat and, despite the cobbles, incredibly fast. In 2024, Mathieu van der Poel set a record average speed of 47.8kph. That’s why Tudor still tests wheel options in the wind tunnel – using pedalling mannequins to replicate rider-bike systems – and why robust wheels must still also be aerodynamic.

How much does Cancellara himself get involved? He was, after all, the first to win Paris-Roubaix on carbon Zipp 303 wheels in 2010. He may be Spartacus but he’s no Luddite.

“He’s there and present but never overshadowing or pushing,” says Bergin-Taylor. “He’s very much taking the lead in terms of responsibility for the team. He goes on recon rides, shares his experience with the younger riders, often calls them out of the blue to have catch-ups, but he will never tell us what to do. He trusts his staff group; he’s a good balance between supporting and giving people space to do their work.”

It seems the 2026 Spartacus doesn’t want to overpower Paris-Roubaix single-handedly. Now it’s about measuring, analysing, understanding it with an elite team of scientists and engineers – it is a major shift, both technically and ideologically. Is this now the standard approach of a WorldTour team or does Tudor see itself ahead in terms of innovation?

“Probably the top teams have similar investment in it. To put it into perspective, we have seven full time staff – three engineers full time, a data scientist, a PhD student, and an industrial designer, which is quite unique, but if we’re making 3D scans of riders, new bike parts, it’s so digitalised that someone who is world class at it is really beneficial.

“We made the strategic decision early on as a team. Most teams need to fund next year, and to fund next year you need to win races. Do you spend half a million on innovation staff, or do you pay for one more rider who can get you points? I understand that there’s a lot of teams who have to make that choice, but we can afford to not have to rush today to get the best riders in the world.

“The potential return on investment from a performance perspective is high. The teams that don’t invest in innovation will be left behind. You either go the UAE way where you can afford the best riders in the world and everyone wants to work for you. If Pogačar picks up the phone and says ‘I want a new time-trial helmet’, they will make him a new time-trial helmet. Whereas if you’re another team and you don’t have that pull, maybe you need to be the one who’s saying, let’s raise the bar for development. And I think the top teams – Visma, Red Bull, Ineos, Trek – are starting to do this. Some teams still invest in old riders, old DSs or old coaches to do it. But to move forward, you need trained engineers to bring that fresh perspective. In F1, old race car drivers are not the ones designing F1 cars. They're involved, but you need engineers to solve these really difficult problems. Fabian and Raphael [Meyer, the Tudor CEO] are really strategic about the long term development. We’re in a unique place.”

Paris-Roubaix still is and always will be brutal. There will be crashes; bikes and wheels will be broken, it will still take an extraordinary rider and a superhuman effort to win, and behind them riders will still limp into the velodrome shattered and caked in mud.

But underneath the mud and mythology, Paris-Roubaix has become one of the most technologically sophisticated races on the calendar and Fabian Cancellara knows this. Whereas once he conquered the pavé with sheer power, he understands that Paris-Roubaix in the 2020s requires brains along with the brawn – and all the signs are there that his is a team to be reckoned with in the future. “We were really happy with our first Roubaix [2025],” says Bergin-Taylor. “We were really close to having two guys in the top ten. Marius [Mayrhofer] would have been eighth but he took the wrong deviation going into the velodrome. Marco Haller was sprinting for ninth and came 12th. Matteo Trentin didn’t do the race as he got sick after Flanders. Now with Stefan Küng joining our Classics group we are looking to take a step forward.”

Read more: Classic kit: Inside Sportful’s 40-year love-affair with the hardest one-day races