Last car I owned actually was pessimistic on the highway by around 1 - 2% depending on speed. In the city depending on how many short trips I drove it was optimistic by up to 5%.

Surprisingly most younger people do not know how to calculate fuel usage especially L/100KM

Have not had a chance to check out present car on a long trip.

I put fuel in the Honda yesterday, trip computer read 66.9 MPG, actual was 68.24 so the actual figure was 2.3% higher than the computer reading. Like Bluerover says, not all trip computers over read.

But very rarely! The Prius has been both very close, and way off. The last 2 tanks were just under 70 mpg, it guessed just over 70 mpg. However, it has also guessed 70 when 62 was the real figure. I must have had 50 fill ups in the i20, only a couple of them I remember stating a lower guess on the computer. I have to say, when it comes to L/100KM though it may as well be pints per furlong - my head only computes in miles per (imp.) gallon...

Generation II (2004-2009) Prius owners have another issue that really affects onboard computer versus "real life" calculations. Our cars are fitted with flexible fuel bladders inside the fuel tanks. Depending on the flexibility of the bladder at any given fill the tank capacity may vary by as much as 3 gallons, even when filling at the same station, same pump and same rate of fill. My onboard mileage will usually be off +/- 2 or 3 mpg but it has been off by as much as 6 mpg.

I glance occasionally at the onboard readings to get a sense of how I'm driving at the time but I don't expect specific accuracy. The same with actual calculations at the pump. I depend on averages over multiple tanks to check the health of the car. I don't get excited about any individual fill.

I just take the fuel fill receipt and record the miles traveled. You can travel miles without consuming fuel if you EOC. You can consume fuel without travelling any distance if you allow your engine to idle. I love EOC and hate idling. Fuelly just allows me to publicly display my mileage and keep track of that mileage over time, while confirming my math is correct, a skill I acquired when calculators weighed 20 pounds and weren't allowed in school.

To check your difference between two different tires if you replace them, or even if you just want to know.

Mark a point on the tire and the pavement, roll the car until the mark covers one revolution precisely, then mark the pavement again. Measure the distance between the two marks with your preferred measurement type.

A 25 inch diameter tire will roll right at 3.1416X25 inches which is 78.54 inches (per revolution). A miles is 63,360 inches. 63360/78.54=715.6 revolutions per mile.

Lets say you put slightly larger diameter tires on your car. They roll 705 revolutions per mile. Since they roll fewer revolutions per mile your speedo (and odometer) will register lower at identical speeds. To calculate the difference divide 715.6/705=1.015. Multiple the distance recorded on your odometer by 1.015 and you get the true distance.

Works fine or everything but ice, metric, US, Imperial, just require adjustments.

Just found something that may very well effect the computer being off.
Been using my dash readout of MPG. It only goes to 99.9. That's all. The bar readout on the computer screen only goes to 60mpg. I take it the computer is using the numbers same as the dash readout.
For example; It reads 99.9 whether I'm coasting in drive, or coasting in neutral. We all know the car is doing better mileage in neutral.
Still, mine is very close. It's also very consistent. Very accurate as far as working with it. Great for motivating people to get better mileage.

Just to contribute a data point on computer-indicated vs actual fuel economy, for my latest fill-up, yesterday (664 miles, combined city/highway driving):

• Computer says: 6.0 L/100 km = 39.2 MPG US = 47.1 MPG IMP
• Actual: 6.38 L/100km = 36.87 MPG US = 44.28 MPG IMP

The computer was about 6% optimistic in this example. I'm driving a 2015 Audi Q5 3.0 TDI.

The car doing better in neutral depends upon the deceleration fuel cut off behavior of the car.

As for my dash display, the computer ignores engine off coasting miles for the MPG calculation. So it is lower than Fuelly.

Keep in mind that you're assuming the display's value is the number being used in the computer's calculations. It might not be. It might just be a display limitation. When you're using no fuel and your vehicle is moving, you are effectively getting "infinity miles per gallon". Most people would not recognize the "infinity" symbol -- the number "8" on its side -- or understand "infinite". The 99.9 could well be just a programmer's decision to show the highest possible value on a 2.1 format display.

From a physics perspective, car manufacturers simply do not want to put in an accurate and expensive fuel flow meter, when "pert near" is good enough for most folks. Ditto with fuel gauges, in which you can drive 200 km (120 miles) and still show full, drive 1/2 of your known range while the gauge shows 2/3 full, etc. Again, it's an inexpensive "pert near" solution that serves the purpose of giving a relative reading (e.g., my tank is less than full, and more than empty) rather than an absolute (and accurate, and correct) one.

Yup Steve... Agree with your putting things in perspective. Like the term "pert near" good enough.

I'm just rationalizing the difference, somehow.
The tank I have now. Even though the calculation came up a little bit lower than my best. That's 0.7 mpg lower. My computer said it's the best I've gotten. It makes me feel good, if nothing else. Reference to my fuel thread makes me feel even better.