Skip to tool
ecech.
📏 Productivity & Units

Your Trip Computer Is Probably Flattering You by Around 5%

It estimates from injector timing rather than measuring. The pump measures — but one tank is noisy, because auto-shutoff never stops in the same place twice.

L/100km

across every tank

MPG (US)

MPG (imperial)

Computer error

Tank by tank

How much of the spread is measurement noise

Advertisement

How the calculation works

612 km, and the pump clicked off a litre early 44.8 L recorded 7.32 L/100km 43.8 L recorded 7.16 L/100km 45.8 L recorded 7.48 L/100km Same driving. A 2.2% error in the fill is a 2.2% error in the answer — which is why one tank proves nothing.

How to Use This Tool

Record the distance since the previous fill and the fuel it took to fill up again, one line per tank. Several tanks is not a nicety — it is the whole method.

How to measure it properly

  1. Fill to the same point every time. First auto-shutoff click, no topping up. Topping up is what makes the fill level inconsistent, and consistency matters far more than fullness.
  2. Reset the trip meter at the pump, or note the odometer.
  3. Drive normally and fill up again the same way.
  4. Record distance and litres — that pair is one measurement.
  5. Repeat for several tanks and divide total distance by total fuel.

The fuel figure is always for the previous tank's driving. Getting that off by one is the usual mistake, and it produces a figure that looks plausible and is wrong.

Why one tank tells you nothing

The pump's auto-shutoff triggers on backpressure from the filler neck, which depends on the nozzle angle, how fast the fuel is flowing, how the car is parked and the temperature. A litre either way is entirely normal.

On a 44.8 L fill that is 2.2% of the measurement, and it lands directly on the result: the same 612 km comes out at 7.16, 7.32 or 7.48 L/100km depending only on where the nozzle clicked. Two tanks that differ by 3% may well be identical driving.

Averaging across tanks fixes it, because the errors are random in sign. Total distance divided by total fuel is the right calculation — not the average of the per-tank figures, which weights a short tank the same as a long one.

Same 612 km, two sources trip computer 6.9 L/100km measured at the pump 7.32 L/100km 6.1% optimistic. The computer estimates from injector timing; the pump measures volume, and is the one with a legal accuracy requirement.
Consistent optimism is common and mostly harmless — until you use it to compare two cars.

Why the trip computer disagrees

It does not measure fuel. It calculates it from how long the injectors were open, multiplied by an assumed flow rate, and divides by distance from the wheel speed sensor. Both inputs carry error, and the errors do not cancel.

In practice these displays read optimistically more often than not, typically by a few per cent. A computer showing 6.9 L/100km against a measured 7.32 is 6.1% out.

It matters less than it sounds for day-to-day driving, because a consistent bias still shows you whether today was better than yesterday. It matters a lot for anything absolute: comparing two cars, checking a manufacturer's claim, or working out a genuine cost per kilometre.

The pump is the better reference, and not a perfect one — pumps are periodically verified for accuracy, and fuel volume changes slightly with temperature, so a cold morning fill contains marginally more fuel per litre than a hot afternoon one.

Reading the spread

Once you have several tanks, the variation between them is the interesting part. Perhaps 2% of it is fill-level noise. The rest is real, and it is usually the mix of journeys: short cold trips are dramatically worse than long warm ones, and motorway speed costs more than most drivers expect because drag rises with the square of speed.

If one tank is far off the others, look at what was different about that fortnight before concluding something is wrong with the car.

Advertisement

Frequently Asked Questions

How do I calculate my real fuel economy?
Divide the distance since the last fill by the fuel it took to fill up again, and repeat over several tanks. Then divide total distance by total fuel — not the average of the individual figures, which weights a short tank the same as a long one.
Why does my trip computer show better fuel economy than reality?
Because it estimates rather than measures. It calculates fuel from injector open time and an assumed flow rate, and distance from wheel speed, and both carry error. These displays usually read optimistically, often by a few per cent.
Why do my fuel economy figures vary so much between tanks?
Partly measurement noise: the pump's auto-shutoff never stops in the same place twice, and a litre either way on a 45 litre fill is over 2% of the answer. The rest is real, and it is mostly the mix of journeys — short cold trips are far worse than long warm ones.
How many tanks do I need to measure?
At least three or four, and more is better. A single tank carries a couple of per cent of fill-level error, which is the same size as many of the differences people are trying to detect. Errors are random in sign, so they cancel across several tanks.
Should I fill the tank right up to measure economy?
No — fill to the same point every time, which means the first auto-shutoff click with no topping up. Consistency is what matters, and topping up by hand is the least consistent part of the process.
Is the pump reading accurate?
More accurate than a trip computer, and not perfect. Pumps are subject to periodic verification, and fuel volume varies slightly with temperature, so a cold morning fill holds marginally more fuel per litre than a hot afternoon one. It is still the better reference.

Related tools in Productivity & Units

Browse all Productivity & Units tools
The Mac mini the ecech. site is built on, beside a handwritten note reading ecech.com.

Made by one person

ecech. is not a content farm. Every tool here is written and checked by hand, one at a time, by someone who wanted the tool to exist and could not find a version that showed its working.

No accounts and no sign-in, and nothing you type reaches a server — every calculation on this page runs inside your browser. The ads are served by Google and do set their own cookies, which is set out in full on the privacy page. More about the site.