A car travels steadily along a level road. Its tyres turn smoothly, and no rubber seems to be scraping across the surface. Why must the powertrain still supply energy to keep those tyres rolling? A loaded tyre continually changes shape where it meets the road. Rubber does not return every bit of the mechanical energy used to deform it. Much of that unrecovered energy becomes heat, creating rolling resistance even without obvious sliding [1][4]. The loss occurs as new parts of the tyre repeatedly enter and leave the contact region.

That mechanism explains why a rolling-resistance improvement is useful. It does not, by itself, tell us the percentage improvement in the entire car's energy consumption. We will follow one hypothetical set of tyres from a force ratio to work over a kilometre, then place that work inside a larger energy budget. All numerical cases are invented teaching examples, not tyre measurements or performance promises. The result is a way to interpret an efficiency claim: identify the quantity that changed, the conditions behind it and the total against which its saving is being compared.

Rolling without gross sliding does not imply zero mechanical dissipation in a pneumatic tyre. Each rotation loads and unloads material around the contact region. Viscoelastic deformation produces a loss between mechanical energy supplied during loading and recovered during unloading; much of that loss appears as heat [1][4]. A resisting force or moment can represent this dissipation in a vehicle model, although the physical mechanism is distributed through a deforming tyre rather than located at one imaginary sliding point.

This article follows that effective loss through three distinct quantities: the rolling-resistance coefficient, the mechanical work required over a distance, and the fractional change in a complete energy account. The original examples assume level travel and explicitly stated loads, coefficients and accounting boundaries. They contain no measured product data. The key question is why reducing a component's loss by a particular percentage does not establish the same percentage reduction in whole-vehicle consumption. Answering it requires both the tyre mechanism and the denominator of the claimed improvement; a coefficient alone supplies neither a complete vehicle model nor a range prediction.

Imagine following one small region of rubber around a loaded wheel. As it approaches the road, the surrounding tyre structure bends and changes shape. After passing through the contact region, it recovers. An ideal perfectly elastic material would return all the energy temporarily stored during that change. Real tyre rubber combines elastic behaviour with time-dependent resistance to deformation. Some energy returns and some is dissipated. This repeating loss is called hysteresis [1]. Recovery of the visible shape therefore does not prove full recovery of the energy.

For a miniature energy account, suppose an imaginary deformation cycle receives 10 units of mechanical energy and returns 8. The missing 2 units have been transferred elsewhere, principally into heat in this simplified picture. That is a 20% loss within this invented cycle, not a 20% loss of the car's energy. Different tyre regions undergo different deformations, and the whole tyre combines those contributions. Material formulation, structure and geometry affect that response [1][2]. The example explains the mechanism without pretending that one small piece of rubber describes an actual tyre's measured efficiency.

Hysteresis describes the difference between work supplied during loading and recovered during unloading over a deformation cycle. In a force-displacement representation, the loop area corresponds to the mechanical energy dissipated per cycle. A purely elastic idealization has no such net loss when it returns to its original state; viscoelastic rubber need not recover the complete input work even when its macroscopic shape recovers [1]. The European Commission's explanation likewise identifies deformation-related hysteresis as a major rolling-loss mechanism [4].

For an original normalized cycle, W_in=10 units and W_return=8 units give W_loss=2 units. The cycle loss fraction is 2/10=0.20. Neither that fraction nor the drawn cycle can be substituted directly for a vehicle efficiency. The actual tyre integrates spatially varying deformation, material response and repeated cycles, while construction and compound choices influence those contributions [1][2]. A force-level representation is useful precisely because it summarizes the resulting loss under specified conditions. It must not be confused with the coefficient of sliding friction or treated as evidence that the contact patch is uniformly slipping.

Apply the main example's steady 160-newton resistance over 1 kilometre. The mechanical work needed to overcome it is 160×1,000=160,000 joules. Dividing by 3,600 converts that energy to about 44.44 watt-hours. For this one-kilometre segment, the corresponding energy per distance is 44.44 Wh/km. The distinction matters: 44.44 Wh is the energy for the segment, while 44.44 Wh/km is the rate of energy use with distance. This account covers rolling resistance at the wheels, before any powertrain or charging losses.

Now hold both the load and coefficient fixed while changing speed. At 20 metres per second, the power needed is 160×20=3,200 watts. At 30 metres per second it is 4,800 watts. The faster trip needs more power, but each kilometre takes less time, leaving the same rolling work per kilometre in this model. That result comes from the fixed-force assumption; real tyre resistance can change with operating conditions, including speed and temperature [1]. It also says nothing about the car's other energy demands, which can change substantially when speed changes.

For constant effective resistance parallel and opposite to motion, the required positive work over path length s is W_rr=F_rr s. Thus 160 N over 1,000 m requires 160,000 J=44.444… Wh; normalized over that segment, the consumption is 44.444… Wh/km. For varying conditions the general account is W_rr=∫F_rr(s)ds. The resisting force itself does negative work on vehicle motion, while the equal sustaining contribution supplies positive work. Using the magnitude avoids reversing those two sign conventions.

At speed v, the sustaining power is P_rr=F_rr v. Holding F_rr=160 N yields 3.2 kW at 20 m/s and 4.8 kW at 30 m/s. Traversal times for 1 km are 50 s and 33⅓ s, respectively; P times duration gives 160 kJ in either case. Constant Wh/km follows from constant force, not from a universal property of tyres. The source review describes operating-state and speed-related influences on rolling resistance [1]. Changing coefficient, load or path conditions requires integrating the appropriate force rather than retaining this example's constant value.

A tyre laboratory tries to isolate the property being compared. Load, inflation, temperature and the measurement procedure need suitable control; otherwise a different result may reflect a different condition rather than a better tyre. The Transportation Research Board's report explains why controlled rolling-resistance procedures improve repeatability and why laboratory results still differ from varied road experience [1]. A useful report therefore attaches conditions to the coefficient instead of presenting it as a timeless number belonging to the tyre in every situation.

Imagine two reports that both say 8 N/kN. One describes a stated new tyre under a controlled procedure; the other gives no size, condition or method. Their printed ratios match, but the second report does not establish that the underlying comparison matches. Conversely, two reported values may differ partly because the tyres or conditions differ. The questions to keep are concrete: what exact tyre was assessed, under what load and inflation conditions, with which method, and with what stated measurement uncertainty? These questions identify missing evidence without assuming that either result is dishonest or that a laboratory test predicts every journey.

A rolling-resistance measurement is conditional on the test specimen and procedure. Controlled methods support repeatability and instrumentation accuracy by specifying operating conditions, yet the source report explicitly distinguishes laboratory comparability from the diversity of field conditions [1]. Its 2006 discussion is used here for that enduring measurement principle, not as a claim about the latest edition or legal status of a particular test standard.

An interpretable record identifies tyre specification and state, imposed load, inflation treatment, speed regime, thermal conditioning, measurement method and uncertainty where available. Equality of two reported coefficients is not enough to establish equality of the measurands if those conditions differ. Likewise, a numerical difference does not establish a material advantage without the procedure's uncertainty and comparability. In our example, C_rr=0.010 and 0.008 are stipulated model inputs. They are not inferred from a label class, a current catalogue or a vehicle's displayed consumption. Preserving that distinction keeps a clean worked calculation from being mistaken for a product comparison.

A tyre's design is only one part of the rolling-loss story. Changing load or inflation changes how it deforms. Temperature affects the material response, and a rougher or more deformable road changes the interaction between tyre and surface [1]. A vehicle on another route or carrying another load is therefore not automatically repeating the same rolling-resistance experiment. The simple coefficient model remains useful, but its inputs must represent the situation being discussed.

Consider a small mathematical check. Keep the coefficient at 0.010 while increasing our car's total normal load from 16,000 to 18,000 newtons. Resistance rises from 160 to 180 newtons, or 12.5%. That follows under the declared constant-coefficient assumption. It does not prove that every real tyre responds by precisely that percentage, because the coefficient itself can vary with conditions. Nor does the model justify changing inflation to chase a calculated minimum. NHTSA directs owners to the vehicle's tyre information label or owner's manual for the specified pressure [3]. The operating recommendation and an energy-model parameter serve different purposes.

The approximation F_rr=C_rr N is useful across a declared operating range, but C_rr can itself depend on load, inflation, temperature, speed, construction and road interaction. The source review describes near-proportional force-versus-load behaviour under normal conditions while also identifying important operating variables and surface influences [1]. The coefficient is therefore not a universal constitutive constant independent of state.

For a sensitivity calculation with C_rr held at 0.010, increasing N_total from 16 kN to 18 kN gives 160 N to 180 N: a 12.5% increase. If the coefficient changes as well, the ratio is F_new/F_old=(C_new/C_old)×(N_new/N_old). Keeping both factors makes the assumption visible. Road roughness can also affect energy pathways outside a narrow isolated-tyre laboratory description, so a whole-vehicle road result requires an explicit boundary. These dependencies motivate proper model inputs; they are not instructions to optimize inflation independently of vehicle requirements. NHTSA identifies the vehicle label or manual as the pressure reference [3].

Our 5.56% reduction belongs to a specified mechanical energy account. To predict battery consumption, fuel use or electricity at a charger, we would also need the relevant conversion losses and other demands. A heater, for example, is not part of the defined wheel-work ledger. The National Academies describes different vehicle energy pathways and explains why rolling-loss changes do not translate into one universal consumption response [2]. Moving a result from one boundary to another needs additional information.

There is a second arithmetic trap even when the boundary is fixed. Imagine an ideal system with the same available 16 kWh in both cases and energy per distance exactly proportional to our wheel account. The baseline allows 100 kilometres; the revised 15.111 kWh per 100 kilometres allows about 105.88 kilometres. The energy-per-distance reduction was 5.56%, but the distance increase is 5.88%, because distance is divided by consumption. This separate hypothetical exercise is not a real range estimate. It shows why 'percent less energy per kilometre' and 'percent more kilometres' need not be numerically identical.

A change in mechanical wheel work cannot simply be relabelled as an equal fractional change in fuel, battery-terminal or wall energy. Conversion efficiencies, accessory loads and operating profiles affect the mapping, and the cited vehicle-energy discussion separates those pathways [2]. Even a constant efficiency would preserve a fractional improvement only for the terms to which it applies, with no additional unchanged energy demand entering the denominator.

For a separate ideal fixed-energy illustration, consumption changes from e_0=16 kWh/100 km to e_1=136/9 kWh/100 km. Their ratio is e_1/e_0=17/18, so the consumption reduction is 1/18≈5.56%. At equal available energy and otherwise identical conditions, distance is inversely proportional to e: D_1/D_0=18/17. Its increase is 1/17≈5.88%. With an invented 16 kWh allowance, the distances are 100 km and 1,800/17 km≈105.88 km. The different denominators explain the unequal percentages; neither number is a measured EV range benefit or an assertion that usable battery energy equals wheel work.

Suppose a driver changes tyres and later sees a lower average consumption figure. The observation can be real without showing how much of the change came from rolling resistance. Route, speed, wind, traffic, payload, temperature and accessory use may also have changed. Laboratory controls exist partly because ordinary journeys combine many influences [1]. A single before-and-after display gives a total result; it does not separately measure each contribution inside that total.

The useful distinction is between observation and attribution. 'Consumption was lower on this journey' describes what was recorded. 'The tyres alone caused the entire difference' requires evidence that rules out the other explanations. Our calculation avoids that problem by stipulating unchanged other demands, which is why its 5.56% result follows cleanly. For a real report, look for comparable conditions, a stated method and repeated evidence with uncertainty. This is a way to read a claim, not an invitation to perform road tests, coast in traffic or change tyre settings. A good comparison makes its controls visible instead of hiding them behind a precise percentage.

Whole-vehicle consumption is an aggregate observable. Estimating the causal contribution of a tyre change requires controlling or modelling competing changes in route, aerodynamic demand, vehicle loading, thermal state and accessories. The source report explicitly warns that varied road surfaces, inflation, alignment and other operating conditions affect field rolling resistance and its agreement with laboratory values [1]. A raw consumption difference does not uniquely identify ΔC_rr.

For the worked budget, attribution is possible because E_other is fixed by construction. In observational data it is an assumption to evaluate, not a fact supplied by the display. Even with matched reported averages, differing speed distributions or thermal histories can change the underlying demands. A defensible comparison needs an identified protocol, compatible tyre specifications, sufficient repeated evidence and appropriate uncertainty treatment. The article proposes no on-road experiment. Its purpose is to distinguish a measured aggregate change from a modelled component contribution, and to explain why a precise result without those controls can be less informative than a carefully bounded range.