The physics of tyre pressure needs to be considered when we wonder about the recommended pressures for low profile tyres seeming too high (especially if like me, you're used to higher profile tyres at lower pressures previously)
Lower profile tyres have a smaller containing volume because the air cavity inside the tyre is smaller but the
quantity of air in the tyre is not too far off what you would get in a higher profile tyre - the main difference is
volume. Let's look at an example...
Tyre Size | Internal Volume | Pressure | Temperature | Quantity of Air (moles to litres) |
225/45 R17 | 12.14 litres | 38 psi | 15 C | 12.809 litres (1.327 moles) |
225/40 R18 | 11.08 litres | 42 psi | 15 C | 12.924 litres (1.339 moles) |
The Ideal Gas Law, of course, defines the relationship between volume, pressure and temperature so it's no surprise that smaller volumes at higher pressures (at the same temperature) require proportionally equivalent quantities (i.e. moles) of gas as larger volumes at lower pressures.
Simply put, one should expect to see higher pressure requirements for lower profile tyres because of the reduced internal volume of the air cavity (plus some variation for the model and performance characteristics of the car, I expect). Hell, road bicycle tyres can be ~80 psi + for the same reason: relatively small internal volume to hold the quantity (moles) of air required to retain tyre structure.
Now, comparisons with tyres and balloons or tennis balls are flawed because tyres are constructed very differently and this is important for considering under-inflation for comfort or grip reasons. Slight under-inflation is probably ok but go too low and the shape of the tyre distorts as per this well known image...
The construction of tyres cause concave deformation if they are under load but soft on pressure. Go too soft and you end up with
less contact patch and less grip, not more, with increased wear at the edges.
So, I would probably not deviate more than a few (<5) psi below the factory recommended pressures personally. Experiment and observe wear patterns (I wonder if there's a way to do a short term tyre wear test using a safe paint?
This paper seems to suggest it is possible).
I probably should double check the figures above but I'm off to watch the Malaysian F1