Research · Tyre pressure
What should you actually run?
Ten psi too high cost nine watts. Ten psi too low cost one. The penalty for guessing is not symmetrical, and almost everyone guesses upward.
For most of cycling's history the answer was printed on the sidewall and treated as a target. Then real-world testing found that rolling resistance stops falling at a certain pressure and starts climbing again — and that the cost of being over that point is roughly nine times the cost of being under it. The number you want is not the maximum; it is a break-point that moves with your weight, your tyre and the road.
Reviewed August 2026 · 2 sources
- Rolling resistance does not fall forever with pressure. SILCA's model describes a break-point: “below the Breakpoint pressure the system will be dominated by Casing Losses (though still affected by impedance) and at higher pressures the system will be dominated by Impedance Losses.” Past that point, adding air makes you slower.
- The penalty is strikingly asymmetric, and that asymmetry is the actionable finding. In SILCA's testing, “Being 10psi too high cost 9w” while “being 10psi below the breakpoint only cost 1w.” Their conclusion is unambiguous: “it is always better to be too low than too high with your tire pressure.”
- There is no single correct pressure, because the break-point moves. SILCA lists the inputs as “What surface are you riding on, what size are your tires, how much does your system weigh, what kind of bike are you riding” — and states the direction plainly: “The rougher the surface, the smaller the tire, the lower that breakpoint pressure is going to be.”
- The framing matters as much as the number: “Tire Pressure is NOT a Maximize or Minimize Variable, but Can be Optimized for your weight, tire size and course conditions.” It is a target to hit, not a quantity to push in either direction.
- The widely repeated shop-floor version — that most riders, amateur and professional alike, run more pressure than is fastest — matches the direction of SILCA's findings and the sidewall convention that produced the habit. It is consistent and plausible, but it is a characterisation of rider behaviour rather than a measured population statistic.
- Riders moving to wider tyres routinely report that dropping pressure substantially felt slower and measured faster, which is exactly what an impedance-dominated model predicts. The perception gap is real and repeatedly described; how large it is has not been quantified.
- Tyre quality changes the size of the penalty, not just its direction, and it is not published per-model. SILCA notes only that “it is going to be a lot more than 9w if your tire pressure is 10psi too high” with inferior tyres — a warning without a table behind it.
- The 9w-versus-1w figures come from specific tyres on specific surfaces in SILCA's own testing. The asymmetry is the transferable finding; the exact watt counts are not a constant you can apply to your own setup.
- Break-point pressure is calculated from a model, not measured on your road. Calculators encode surface as a category — smooth asphalt, coarse asphalt, gravel — but real roads vary continuously, and no consumer tool measures the surface you are actually on.
Why the sidewall number is the wrong target
The pressure moulded into a tyre's sidewall is a safety limit — the most that casing is rated to hold. It has never been a performance recommendation, but for decades it functioned as one, because the intuition it fed was correct as far as anyone could measure: on a steel drum in a lab, rolling resistance really does keep falling as pressure rises. The drum is smooth. Roads are not, and that difference turns out to contain the entire effect.
On a real surface, a tyre that is too hard stops absorbing small irregularities and starts transmitting them — into the bike, and into the rider, where the energy is lost as vibration rather than returned. SILCA calls these impedance losses, and their model has them overtaking casing losses above a certain pressure. That is the break-point: the pressure at which the two curves cross and further inflation starts costing you speed instead of saving it.
The asymmetry is the whole practical lesson
If the penalty curve were symmetric, precision would matter enormously and most riders would be stuck. It isn't. SILCA's numbers — nine watts for ten psi over, one watt for ten psi under — mean that a rider who cannot calculate their exact break-point still has a very good strategy available: aim low. Erring five psi under costs almost nothing measurable; erring five psi over costs real speed and a harsher ride at the same time.
This inverts the instinct most riders were trained into. Topping up hard before a fast ride feels like preparation, and it is the one adjustment that reliably makes the bike feel quicker while making it slower. Comfort and speed are not in tension here, which is unusual — the pressure that transmits less road buzz is, within the useful range, also the pressure that rolls faster.
Getting to a number you can trust
Start from the inputs that actually move the break-point: system weight (you, the bike and everything on it, not just body weight), the measured width of the tyre once mounted, and the roughness of the surface you will spend most of the ride on. A calculator will turn those into a starting pressure. Treat it as a starting pressure — the model is good, but it does not know your road.
Then adjust in one direction. Because the cost of being under is roughly a ninth of the cost of being over, the sane search is downward from a modest starting point until the bike feels vague in corners or you start bottoming the casing on impacts, then back up slightly. What defeats this in practice is not the method but the record-keeping: pressure interacts with tyre width and casing, so a number you validated on one set of tyres quietly stops applying the moment you change them, and almost nobody remembers which number belonged to which tyre.
Sources
- Part 4B: Rolling Resistance and ImpedanceSILCAEstablishes the break-point model — below it casing losses dominate, above it “Impedance does in fact dominate the rolling resistance” — and advises setting pressure “a few psi below the BreakPoint Pressure” rather than above. Also that rough or soft surfaces lower optimal pressure.
- SILCA Tire Pressure Calculator ExplainedSILCAThe asymmetry, quantified: “Being 10psi too high cost 9w” against “being 10psi below the breakpoint only cost 1w,” concluding “it is always better to be too low than too high.” Lists surface, tyre size, system weight and bike type as the inputs, and warns the penalty is larger on poorer tyres.