A rider drops five psi out of both tyres before a rough-surfaced club ride, following advice he's read a dozen times: lower pressure means more contact patch, more comfort, and — on anything but glass-smooth tarmac — supposedly even less rolling resistance. He feels the difference immediately. Smoother. Quieter over the seams in the road. He assumes he's also faster, because that's what the advice promised. He's dropped pressure before every rough ride for years, the same five or six psi, without ever once measuring what it actually cost or gained him.

New on-road testing says he probably isn't faster. Not by that margin, and not for free — and the same testing shows two other equipment choices sitting on his bike moving the number by just as much, that he's never once thought to question.

Section 01What the field data actually shows

Most rolling-resistance numbers riders quote come from drum tests — a tyre spun against a smooth roller in a lab, which is repeatable but doesn't reflect a real road surface, a real rider's weight shifting in the saddle, or real ambient conditions. Sellier and Bellenoue built a field method instead: a coasting, energy-imbalance approach that uses a rider's own speed and elevation data over short on-road test sections under 100 metres, isolating rolling resistance from aerodynamic drag with better than 2% repeatability (Sellier & Bellenoue, 2025).

Testing real tyres on a real road, dropping pressure from 6 bar to 3 bar increased measured rolling resistance by roughly 26% (Sellier & Bellenoue, 2025). That's not a rough-surface edge case or a drum-test artefact — it's a rider's actual power going somewhere other than forward motion, on a normal sealed road, well within the pressure range most riders actually run.

Pressure wasn't the only lever that moved the number. Time-trial-specific tyre casings from the same manufacturer showed a 33% difference in rolling resistance against that manufacturer's standard version, even with the TT tyre tested at a different temperature to its stablemate (Sellier & Bellenoue, 2025). Tyres from different manufacturers varied too, by a 9% average across the three brands tested — with two of the three statistically indistinguishable from each other and the third measuring meaningfully lower, 7 to 11% depending on pressure (Sellier & Bellenoue, 2025).

Tyre temperature mattered by a similar-sized margin: comparing tyres at a cooler temperature (14°C) against warmer ones (18°C), the study measured a 26% increase in rolling resistance for the colder tyres (Sellier & Bellenoue, 2025). Pressure, tyre choice, and temperature all move the number by a comparable order of magnitude — none of them is a rounding error next to the others.

Section 02How they actually tested it on the road

The setup was deliberately simple and real-world: one experienced rider on a standard road race bike, running commercially available 28 mm tubeless tyres from three manufacturers, on a stretch of road sheltered from side winds to keep conditions consistent between runs (Sellier & Bellenoue, 2025). An initial set of calibration passes at varying speeds establishes the rider and bike's aerodynamic drag for that day; low-speed repeat passes then isolate the rolling-resistance term for each tyre, pressure, and temperature condition tested (Sellier & Bellenoue, 2025).

That's what makes the result harder to wave away than a lab figure. This isn't a drum spinning a tyre in isolation — it's a real rider coasting on a real road, with speed and elevation change measured over a short, repeated test section (Sellier & Bellenoue, 2025). The tighter the method, the harder it is to dismiss a 26% swing as noise.

Section 03Why softer costs you something real

Rolling resistance comes largely from how much a tyre deforms as it rolls. Drop the pressure and the tyre's contact patch flattens and widens against the road — good for grip and comfort, since more rubber is touching the ground and absorbing more of each bump. But that same deformation flexes the tyre's casing more with every single wheel rotation, and every bit of that flex is energy the rider paid for that never turned into forward motion. It's dissipated as heat in the rubber instead.

This is exactly why "softer is basically free speed on rough roads" oversells itself. That framing comes from a real, separate effect — the "impedance" phenomenon, where extremely high pressure on genuinely broken pavement causes the whole bike-and-rider system to bounce, wasting energy a different way. But the pressure range in this study — 3 to 6 bar — sits well inside normal road-tyre territory, nowhere near the very-high-pressure extreme where impedance takes over. Inside that normal range, the deformation cost of going soft is the dominant effect, and it's a straight, measurable tax on your legs.

Most riders have internalised half of this trade-off and never questioned the other half. The comfort argument for dropping pressure is real and well understood — nobody disputes that a rock-hard tyre on a rough road is miserable and slow in its own way. What gets skipped is the reminder that the relationship isn't linear all the way down. Comfort keeps improving as you drop pressure further. Rolling resistance, past a point well short of "as soft as it'll go," starts working against you instead of for you.

Section 04How to actually apply it

1. Don't chase "as soft as possible" for comfort

   The comfort gain from dropping pressure further keeps costing you rolling resistance well before you hit genuinely broken road surfaces where the trade flips.
2. Match pressure to the surface, not a fixed habit

   Smooth tarmac rewards higher pressure inside your comfortable range. Save the lower end for surfaces rough enough that impedance, not deformation, becomes the dominant cost.
3. Treat tyre choice as a real lever, not an afterthought

   The gap between a race-oriented casing and a standard casing from the same manufacturer showed up as a measurable, multi-watt difference — comparable in scale to a real pressure change.
4. Warm your tyres before a hard effort if you can

   The temperature effect measured up alongside pressure and tyre choice, not below them. A few easy minutes before a hard effort or a test moves the number in your favour for free.
5. Test your own setup, don't just copy a number

   Rider weight, tyre width, and surface all shift where your personal sweet spot sits. Use this data to question extremes, not to adopt one universal pressure.

Section 05Where riders get this wrong

The most common misread is generalising "lower pressure helps on rough roads" into "lower pressure is basically always better." This data argues against the second claim specifically. Inside a normal road-pressure range, dropping pressure has a real, measurable cost — it doesn't become free just because the road has some texture to it.

The second misread is fixating on pressure alone and ignoring tyre choice. A race-oriented casing from the same manufacturer moved the number by a margin in the same ballpark as a real pressure swing — that's a lever most riders never touch because they've never been told it's comparably significant.

The third is assuming this data settles the debate on very rough or broken surfaces. It doesn't — this test sits inside normal road-tyre pressure territory, not the very-high-pressure extreme where the impedance effect takes over on genuinely broken pavement. This is evidence against "lower is always better," not evidence for maximum pressure everywhere.

The fourth is treating one of these three levers as the whole story. Riders who obsess over pressure while running a heavy training tyre cold out of the garage are optimising one input and ignoring two others of comparable size. All three move together in the real world — pressure, casing, and temperature — and none of them alone tells the whole story of why a ride felt harder than the numbers suggested it should.

Section 06Applying it with HELIOS

THRIVE logs your ride data the same way regardless of which wheels or tyres you're running that day — which is exactly why equipment-driven noise like this matters. If your estimated effort and training load look consistent ride to ride but your average speed at a similar effort keeps drifting, tyre setup — pressure, casing, even how warmed-up the tyres were — is a real, physical explanation worth ruling in before you start questioning your form.

It's not a number the app measures directly. It's context for reading the numbers it does show you: a training-load trend that holds steady while your average speed at that effort quietly slides is a prompt to check the bike, not just the legs.

Section 07Bottom line

Dropping tyre pressure from 6 bar to 3 bar cost this field test roughly 26% more measured rolling resistance — a real, on-road number, not a lab artefact. Tyre casing choice and tyre temperature moved the number by a comparable order of magnitude. None of this says run maximum pressure everywhere; it says the "softer is basically free" version of the comfort argument has a real floor, and pressure, tyre choice, and temperature are three separate, measurable levers — not one.