There seems to be growing discontent with the "official" mpg figures that are being published by the manufacturers. (KIA and Hyundai have had to revise mpg downwards claims in the US for example.) I have noticed that the official figures are harder (impossible!) to achieve in real life in most cars.

Does this mean that the tests are effectively meaningless or that they should be used a measure of relative mpg between models ?

The tests where never to be considerd as achievable figures for real life driving i think but more for direct comparison between cars to help customers make a selction of which car has better fuel consumption.

The reason some cars get closer to the claimed figures than others is that some cars are designed specifically to get good results in the test as this may help sell cars.

This is a bit sneeky really as normal driving may not replicate the type of driving likely to achieve these higher figures. In may even mean one car with lower combined mpg could actually get better real world mpg than the car that was designed for this specific test.

Not sure if im explaining myself very well.

There are tricks manufacturers can do to almost cheat the tests. Cheat is maybe the wrong word but you know what i mean. look at hybrids, their official combined mpg results are way higher than a comparable diesle yet in the real world the diesle would probably beat it.

It is easy to design a car to ace a specific test but harder to design a car to achieve good results in all real world scenarios.

Edited 23 November, 201213 yr by Jockdooshbag

How to find " real world" mpg figures.....

http://www.honestjohn.co.uk/realmpg/

The tests where never to be considerd as achievable figures for real life driving i think but more for direct comparison between cars to help customers make a selction of which car has better fuel consumption.

The reason some cars get closer to the claimed figures than others is that some cars are designed specifically to get good results in the test as this may help sell cars.

This is a bit sneeky really as normal driving may not replicate the type of driving likely to achieve these higher figures. In may even mean one car with lower combined mpg could actually get better real world mpg than the car that was designed for this specific test.

Not sure if im explaining myself very well.

There are tricks manufacturers can do to almost cheat the tests. Cheat is maybe the wrong word but you know what i mean. look at hybrids, their official combined mpg results are way higher than a comparable diesle yet in the real world the diesle would probably beat it.

It is easy to design a car to ace a specific test but harder to design a car to achieve good results in all real world scenarios.

I understood you, but then I knew that "official fuel figures" were of no practical validity in real World driving anyway.

In the specific case of hybrids, the issue is that the tests are unrealistically short, and they can lean on their battery packs in a way they can't in real life.

I've always knocked off 10-15% in my head when looking at these figures. I've actually found the figures in the TFSI Octy to be achievable with a careful foot but not in the long term.

Hybrids will be particularly troublesome I predict. With daft punters looking at the 250mpg figures for things like the Ampera then wondering why they get 40 when it's on the generator all the time since they can't be arsed to read a manual or charge it.

The tests should really be done by getting a selection of 10-100 average punters to do a test run. Knock off the top and bottom 10% then average the rest.

Edited 23 November, 201213 yr by Aspman

I've always knocked off 10-15% in my head when looking at these figures. I've actually found the figures in the TFSI Octy to be achievable with a careful foot but not in the long term.

Hybrids will be particularly troublesome I predict. With daft punters looking at the 250mpg figures for things like the Ampera then wondering why they get 40 when it's on the generator all the time since they can't be arsed to read a manual or charge it.

The tests should really be done by getting a selection of 10-100 average punters to do a test run. Knock off the top and bottom 10% then average the rest.

I agree, well up to a point. I'd also like the tests to be run over 1 hour for the "real World" tests, and 30 minutes for any remaining "steady speed" tests. Ok, for repeatability of calls for braking and acceleration, you'd need to use a rolling road simulator and to prevent "learning the road" can't get any specific driver to carry out more than one test on any profile every 2 years.

I like claimed mpg figures, they're my favourite fairytales.

More seriously the tests that they do to get official mpg figures are screwing up the cars. A bit like the obsession with the Nurburgring is doing for suspension.

Car are tuned exactly to give the best results in the tests. That's why you get flat spots at certain point in the rev range, because that's where they do particular tests. They also don't help numpties pick the right cars. They don't realise that if they pick a 1.0l 60hp car then DON'T drive like a granny the mpg might well be worse than if they'd bought a 2.0l since the car will live in 3rd being thrashed to death.

My in-laws bought a Kia Sportage 2.0 4x4, official mpg put to them was about 44mpg. Now I thought that was in fantasy land, a big petrol 4x4 doing 44mpg. But they insisted. 2 months after getting it they averaged 22-25mpg and back it went.

Maybe one option is to dump the official figures and have a range i.e Octavia TFSI you'll get 28-37mpg depending on driving style and road conditions.

I had no idea how the tests were carried out so did a bit of research (usual disclaimers apply):

From http://www.green-car-guide.com/the-secrets-behind-official-mpg-v-real-life-mpg.html

How ‘official’ fuel tests are conducted

The first surprise to most people may be that the car makers themselves, rather than some independent body, run the tests, albeit to a brief specified by the appropriate department of transport.

The tests themselves are conducted on a ‘Rolling Road’ dynamometer and have two parts. The first part consists of 2.5 miles of accelerating, slowing down and idling in the lower three gears (unless it’s an automatic), after a cold start, at an ambient temperature that we expect to experience on a English summer day. The maximum speed is 30mph and the average 12mph.

Following on immediately from this, the second figure is derived over a further 4.3 miles driving in all gears, accelerating very gently up to a maximum of 75mph, and averaging 39mph. The total fuel consumed in both tests, divided into the total 6.8 miles, gives the combined result.

Peter views this test as being too simplistic and therefore not accurately reflecting the driving of the average motorist. “It’s not that the method of measurement is crude – in fact the laboratory testing method works on the inherently accurate basis of assessing the amount of Carbon Dioxide and other gasses emitted by the car’s exhaust. From this, the researchers are able to calculate both the CO2 figure as well as the MPG ones that you see published. It’s just that the tests themselves do not reflect enough of the vagaries of everyday driving conditions.”

Chris

Peter views this test as being too simplistic and therefore not accurately reflecting the driving of the average motorist. “It’s not that the method of measurement is crude – in fact the laboratory testing method works on the inherently accurate basis of assessing the amount of Carbon Dioxide and other gasses emitted by the car’s exhaust. From this, the researchers are able to calculate both the CO2 figure as well as the MPG ones that you see published. It’s just that the tests themselves do not reflect enough of the vagaries of everyday driving conditions.”

Chris

That's my view too; hence #6 above where I try and improve the test design.