CarvicoThe formula, and the number people get wrong
Fuel economy is one division:
The trap is in the second number. "Gallons used" is not the fuel in your tank and it is not the fuel you bought at the start of the trip. It is the fuel you bought at the end — the amount it took to put the tank back exactly where it was.
Think of it as a bookend problem. Two full tanks are the bookends. Everything between them is the distance. The gallons that refill the tank at the second bookend are, by definition, exactly the gallons you burned getting there. You do not need to know the tank's capacity, and you do not need to know how full it was to begin with — you only need both ends to be at the same level.
The one rule
Your first fill-up after starting to track never produces an MPG figure. It only sets the baseline. Every fill-up after that produces one. Apps that show you a number on the very first entry are guessing.
The full-tank method, step by step
- Fill completely, and note the odometer. Fill until the pump's first automatic click-off. Write the odometer reading on the receipt, or log it. This tank is your baseline — no MPG yet.
- Drive normally. Do not baby it for the sake of the number, unless what you want to measure is how you drive when you are being careful.
- Fill completely again — same way. First click-off again. Same station and same pump speed if you can manage it; topping-off habits are the single biggest source of noise in hand-calculated MPG.
- Subtract the odometer readings. Current minus previous equals miles driven on that tank. If you use the trip meter instead, reset it at every fill-up and never trust it after a battery disconnect.
- Divide by the gallons on the second receipt. That is your MPG for that tank.
Worked example. You fill up at 42,180 miles. You drive for a week or two, and fill up again at 42,533 miles, taking 11.4 gallons.
353 ÷ 11.4 = 30.96 MPG
Notice that the gallons from the first fill-up never entered the calculation. They belong to the tank before it.
MPG calculator
Two odometer readings and one receipt
Everything is computed in your browser. Nothing is sent anywhere — this page has no analytics and no trackers.
Doing this once is arithmetic. Doing it at every fill-up for four years is a record — and a record is what tells you that your MPG has slipped 12% since February, or what a set of tires really cost you per mile.
Partial fill-ups: the trap
You stop for $20 of gas because the line is long. That fill-up cannot produce an MPG figure — you do not know how much fuel is still in the tank, so you have no bookend.
There are three things people do about this, and only one of them is right.
| Approach | What it does to your numbers |
|---|---|
| Divide anyway | Garbage. You are dividing a full tank's worth of miles by a partial tank's worth of gallons. This is what produces the impossible 61 MPG entries people post screenshots of. |
| Delete the partial | Wrong in a subtler way. The miles stay, the gallons vanish, and the next full tank looks unnaturally efficient. Your fuel-cost totals are also now understated. |
| Carry it forward | Correct. Mark it as a partial. Add its gallons and its miles to the next full tank, then divide the combined miles by the combined gallons. One honest figure covering both stops. |
Carried forward, the arithmetic looks like this: partial fill of 5.2 gallons over 90 miles, then a full fill of 9.8 gallons over another 240 miles. You do not get two figures. You get one: 330 miles ÷ 15.0 gallons = 22.0 MPG.
Why this matters more than it sounds
Some popular trackers credit a fill-up's fuel to the miles driven before it rather than after — that is, they use the gallons you just bought against the distance you drove on the previous tank. On a steady routine the error mostly cancels out. On a road trip, where one tank is highway and the next is city, it can be off by 15% in either direction, and it never converges. If your app produces a number from a single first entry, this is likely what it is doing.
Why your dashboard reads high
Your trip computer does not measure fuel. It estimates it, by adding up how long the injectors were commanded open and applying a model of what that means in volume. It then divides by distance derived from wheel-speed sensors. Two estimates, one division, and every source of error compounds:
- Injector drift. The model assumes nominal flow rates. Real injectors age, and fuel temperature changes density.
- Odometer calibration. Speedometers are generally calibrated to read at or slightly above true speed rather than below, and tire wear and non-stock tire sizes shift the number further.
- Rounding and smoothing. Many displays smooth aggressively, so the number you see is a comfortable average rather than the last tank.
- The reset you forgot. Lifetime-average displays include the trip where you sat in traffic for two hours with the engine on.
Consumer testing has repeatedly found trip computers to read optimistic — commonly on the order of a few percent, occasionally much more. The direction is consistent enough that it is worth knowing your own car's offset. Track by hand for five tanks, compare with what the dash claimed over the same stretch, and you will have a personal correction factor that stays useful for as long as you own the car.
The useful reframe
The dashboard number is a driving-feedback instrument: good for comparing this morning's commute to yesterday's. The pump number is an accounting instrument: good for budgets, for spotting a failing sensor, and for proving to a buyer what the car actually does. Use each for what it is.
How many tanks before the number is real
Single-tank MPG is noisy, and most of the noise has nothing to do with your car. Pump auto-shutoff varies by nozzle and by how the vehicle is parked. Temperature changes fuel density and warm-up time. A single long highway run inside an otherwise urban tank moves the figure several MPG.
- One tank: an anecdote. Fine for curiosity, useless for conclusions.
- Three to five tanks: a usable average for the current season and driving pattern.
- Ten-plus tanks, rolling: the level at which a real change stands out from noise. This is when "my mileage dropped" becomes a diagnosable claim rather than a feeling.
If the average does drop and stays down, the cause is usually mundane and cheap to check — see the fuel-economy drop checklist for the order to check things in, cheapest first.
MPG, L/100km, and the UK gallon
Two conversions worth memorizing:
US MPG = 235.215 ÷ L/100km
The relationship is inverse, which is why the two scales feel so different: going from 15 to 20 MPG saves far more fuel per mile than going from 40 to 45 MPG, even though both are a 5 MPG gain. L/100km makes that obvious, which is one reason most of the world uses it.
And the classic mix-up: an imperial (UK) gallon is 4.546 litres against the US gallon's 3.785. A car rated 40 MPG in a UK publication is doing about 33 US MPG. Multiply UK MPG by 0.833 to get US MPG.
Questions
Should I reset the trip meter or use the odometer?
Either works, but the odometer is more robust: it survives an accidental reset, a battery disconnect, and the day you forget. Record the absolute odometer reading at every fill-up and the differences are always recoverable.
Does the pump's gallon figure have its own error?
Yes, but a small and regulated one. Retail pumps in the US are inspected by state weights-and-measures programs against a tolerance in the fraction-of-a-percent range. It is far more accurate than your trip computer's estimate.
What about an EV — is there an equivalent?
Same shape, different units: log kWh added and miles driven, and you get miles per kWh (or kWh per 100 miles). Charging losses mean the energy the charger bills you for is a few percent more than what reaches the battery, so a from-the-charger figure is the honest one to budget against.
Do I need an app for this?
No. A notebook in the glovebox works, and so does a spreadsheet. The reason to use an app is that partial fill-ups, unit handling, and rolling averages are exactly the parts people get wrong by hand — and that a log you keep for a decade should not depend on one laptop's hard drive.
Is hypermiling worth it?
Gentle throttle, correct tire pressure, and less idling are free and measurable. The exotic stuff (drafting trucks, coasting in neutral) trades safety and drivetrain wear for a couple of MPG. Measure first: you cannot tell whether a technique works if your baseline is one noisy tank.
Log the fill-up, skip the math
Carvico does the full-tank method properly — partials carried forward, units handled, rolling averages, and a receipt scan that fills the entry for you. Everything stays on your iPhone and your own iCloud.
Download on theApp Store