Tyre pressure calculator

The starting point is Frank Berto’s 15 % tyre drop method: the correct pressure is the one at which the tyre sinks 15 % of its height when you sit on the bike. Below that, the tyre deforms too much and you lose energy; above it, it bounces over irregularities and you lose energy too, as well as grip. On top of his data, the fit published by Dave Adams is used, which expresses pressure as a function of the load on each wheel and the tyre width. Berto's table was published in Bicycle Quarterly, in the article written by Jan Heine (2007) and reviewed by Berto himself; that is the source to check any number on this page against.

That rolling resistance depends on both pressure and the weight each wheel carries is not the method's doing: it has been measured. Grappe and colleagues (1999) crossed five pressures with four vertical loads and measured the coefficient of rolling resistance for each combination, because the loss comes from the hysteresis with which the tyre deforms and both things change it. Hence the calculation here enters through the load on each axle and not through your total weight: that is the variable that was measured.

Load is distributed according to discipline, because posture moves the centre of gravity. The more stretched out you are, the more weight reaches your hands and the front wheel: that is why road splits 40 / 60 here — the figure Berto used — and mountain, more upright, 36 / 64. Then two corrections are applied: less pressure on irregular surfaces, and less with a tubeless setup, where the pinch flat that forces you to keep a margin does not exist.

Of those splits, road's 40 / 60 is the only published one: it sits in the same Heine and Berto table, next to 45 / 55 for a randonneur and 35 / 65 for a city bike. Gravel and mountain are interpolated between those two ends, and the two corrections are workshop convention, not measurements of your bike. The real split depends on your position, where your hands are and the frame geometry, and it moves several points between the tops and the hoods. They order the result between axles; they are not an exact figure.

What was refined to sharpen the number

Three things go into the calculation that almost no calculator asks about. The real width, measured with a caliper and not read off the sidewall: the same 28 tyre on a wide rim measures 30 or 31, and in the formula width is raised to the power of 1.58, so that extra millimetre drops the pressure more than it looks. The casing, because a supple one sinks more at the same pressure and asks for a little more air for the same drop. And the surface, which is what separates riding tarmac from riding dirt: Bertucci and colleagues (2013) measured aerodynamic and rolling resistances for a mountain bike in field conditions and made it clear that the surface is a first-order variable, not a nuance. The selector's three coefficients are still workshop figures; what is measured is that the surface weighs enough to belong in the calculation.

The result comes as a tuning band, not an exact figure: ±3 %, which is roughly what your own gauge may be lying to you by.

Gravel and mountain bike

The calculation is the same and the selector now carries each discipline's widths, up to 2.6 inches. Worth knowing: extrapolating above 40 mm is the model's doing, not Berto's original data, which was measured on far narrower tyres. For a 29 × 2.25 tubeless it gives 18 to 22 psi depending on weight, which is where practice sits, but down there the safety floor matters more than the formula: that is the zone where you strike the rim and unseat the tyre.

Where it stops working

• Berto’s data is from the nineties. It was measured on tyres far narrower than today’s. From 32 mm upwards the reference becomes less and less reliable, and it is better treated as a ceiling to come down from than as a target.
• The front comes out considerably lower than the rear.That is not a bug: the front wheel carries much less weight. Even so, many riders prefer to narrow that difference for cornering grip and for margin against a pothole. Both readings are shown below.
• Casing matters and is not in the formula. A supple tyre deforms more than a stiff one at the same pressure.
• The top half of the curve is measured worse than the bottom half. That overshooting the pressure makes you lose energy to bouncing on real surfaces is the observation the method rests on, but there is no field measurement for it as solid as the one for rolling resistance. If you are torn between two figures, the lower one has more behind it than the higher one.
• The sidewall and the rim always win. No calculation can exceed the manufacturer’s maximum or a hookless rim limit.