Regular readers of DTW’s reviews will note we like to report on fuel consumption.
For as many years as I have been driving I have strived to maintain a fuel consumption log. The longest continuous period ran from 2006 to about 2010. At some point I was unable to note the details in a handy notebook and scribbled them on a piece of A4. The plan was that I’d transfer the figures to the notebook as soon as it turned up from wherever it had got to. During a long drive my daughter got a hold of this page and was allowed to tear it up on the grounds that this would provide a few moments peace while she was distracted. At that point I lost interest in the project as the continuity was now ruptured.
What I still do is methodically record fuel consumption from re-fill to re-fill. I presume you have all done this. You tank the car, note the miles and quickly find out…ah, 34 mpg. Interesting. And that’s it. It almost seems to be the point of the drive. It is terribly satisfying to be able to put a number on the journey.
Yet the data never accumulates. What have I learned from my five years of continuous fuel consumption figures? I do know that I did once get 34 mpg and that was the record for the car. It never again achieved that and when I hammered the car I saw the figures drop to 28 mpg over long trips. The all-time worst was achieved in a month of pure town driving: 22 mpg from the 4-cylinder two-litre petrol unit. On the one hand I was satisfied my car could get 34 mpg, as promised in the brochure, and also disappointed that the figure had a certain exceptional quality to it.
Some cars for me are defined by their MPG: the Mk1 Land Rover Freelander V6 did 20 mpg. Roll-Royce Silver Shadows manage 12 mpg. Jensen Interceptors get about the same. My Peugeot 205 managed 45 mpg and never a bit more. Bristol claimed the Blenheim could get 34 mpg. The Lancia Beta’s 22 mpg was unremarked in the review I read. And the Citroen CX Turbo used so much petrol it had a 220 mile range. Odd stuff to remember.)
From 2012 to 2013 I put the fuel figures into the notebook feature of my iPhone. I didn’t learn to e-mail the data to myself. And in a small moment of carelessness my small daughter (the same one) deleted the notes and I gave up and have not resumed. Apparently as far as my psychology has it, what is important is the data set not the individual instances. I am longing to add up as many MPG figures as possible so as to get a truly definitive long-term view of my car’s real-world consumption. This hasn’t happened.
I won’t have rolling averages or a nice, single figure to set down somewhere. I have yielded control – this then is the real drive to monitor MPG. It’s a wish to control through knowledge. Now that I see that I wonder what I could do with the information other than bemoan my lead foot or my old car’s dipsomaniac character.
Driven To Write 
Hmm. I’ve kept fuel logs for all of my cars, have rarely calculated MPG.
I’m skeptical of MPG estimates based on one tank of fuel. I can’t be sure how full the tank was after the last fill up or that when I refilled it I matched the initial level.
However, especially poor and good single tank MPGs stand out. The really good ones tell me that I started with the tank overfull or didn’t refill it completely. The poor ones make me worry about a fault.
I recently drove cross country. 2011 Accord EX (US model), manual transmission. 79,487 miles on the clock at trip’s end. Trip mileage 6011.9 miles (car’s odometer) or 6089.7 (GPS’ odometer). I’m inclined to believe the GPS because the trend of car/GPS has been falling as the tires have worn. 173.2 US gallons consumed and the last fill up may not have been complete. Call it 35 MP(US)G. Rolling average according to the GPS, 70 MPH.
I’ve done the same calculation for other long drives in the Accord. The trips’ MPGs suggest that it wasn’t fully broken in until around 30,000 miles. That’s when long trip MPG stopped rising.
This is where I have to salute the N. American attention to detail when doing such calculations.
Although a single trip can’t be statistically significant, can we allow it to be indicative? I would feel a single tank tends to be pessimistic; a good result can taken as more convincing than a bad one.
For my personal car I always brim before attempting a reading.
I’ve managed to keep track of all my consumption figures ever since I got my first own car back in 1999, with very few exceptions (lost receipts, fuelling up by someone else). First it was on small notepads like I learned from my dad, then soon it got transferred to Excel tables, and nowadays, I can even put my figures on spritmonitor.de from the smartphone, right after filling up.
While, as Fred mentions, a single tank can never give an exact and reliable figure, over time this is averaging out, and a long time average consumption should be quite accurate (minus the error of the odometer which is again influenced by the tyre diameter).
By the way, I drove my CX turbo rather calmly most of the time and was at roughly 10 l/100 km (28 mpg) which normally gave me a 600 km (370 mile) range. But once let loose on a German Autobahn or on empty alpine roads, my experience was very similar to Richard’s.
My “experience” was vicarious, a road test from Autocar (against a Saab and a Volvo). My lifetime’s experience steering a CX is 10km in Karlsruhe.
Oh, that’s a pity.
Simon: it was enough to make me realise the XM’s controls were inferior, especially the steering.
True. I’ve owned both and totally agree. I first owned a 1979 CX2400 Pallas after thirteen years upgraded to a 1989 XM 3litre manual for ten years but went back to a 1988 CX2500 GTi when I found the opportunity. The CX is a much nicer car to pilot. I have since moved to a C6 but the CX is superior to that also! Mind you the C6 is a totally wonderful car as well and I’m very happy with it. 🙂
Richard, my post reported a long term average, properly weighted. (total gallons)/(total miles).
sum(single tank MPG)/(number of tanks) isn’t quite right, is often more than good enough.
I’m skeptical about single tank MPGs because I’ve had the occasional one that was, well, odd. As much as 10 MPG below or five over the long term average. On long trips a somewhat high single tank MPG (yes, when I look at what happened on a long trip I also calculate single tanks) is often, not always, followed by somewhat low single tank MPG. Incomplete fill followed by a more complete on, I think.
My Mazda helpfully keeps a log of the last 8 tankfuls and plots them on a graph, which is kind of a lazy bloke’s way of doing it. It is enough to make me realise that “high performance diesel” actually worsens the fuel economy of this car – whereas it improves the performance (in every way) of the C6. Go figure …
One wonders how much to trust the MPG rating given by the car itself? I once read that readouts can be optimistic by around 10%. I would start working it out myself, but 1. the arithmetic would negate the point of driving a “sporty” car somewhat, and 2. I don’t think I would like the answer.
I always write down the calculated figure (from refill volume and distance) vs. the car’s computer reading. In case of the C6, the latter is optimistic in the 5% range. But i’ve had tanks with 10% deviation as well as even some pessimistic ones. Most people I know report an optimistic tendency.
Did I mention that I like numbers and useless statistics?
I tend to fill the car with half a tank (25 litres) at a time and monitor the mileage I get from 2 refills. It’s normally pretty consistent at around 370-380 miles on my usual mix of town and motorway driving, which means around 34 mpg on average. That’s all I really need to know, and any variation I can generally attribute to getting stuck in traffic jams, or travelling longer distance at optimum pace.
How do you know it is half a tank unless you have a baseline of full or empty? The fuel gauge is not much of a guide, is it?
Half capacity. 50-litre tank, divided by 2 = 25
That´s the theory but in practice? How do you know when the tank is half full? I can´t see the start of the process to give you an absolute figure. You don´t drain the system and add 25 litres plus the piping volume. I don´t think you drive about and fill it at about half and, by inspection, wait until coincidentally you get 25 litres and no more into the tank: 23.3, 26.5, 24.5, 29, 24.8, 25.1, 21,34, 25.01 ….bingo,start calculation. If the tank is brimmed you have a quite good reference that you are in control of. Half-way is not such a controllable reference, not by ordinary means anyway.
I don’t need that level of detail, it’s no use to me – simple as that.
Sam: I would never have guessed. You use a theory to arrive at a reliable approximation. I figured you´d have used a much more accurate and controlled approach to this pressing and vital matter. For me it´s really bipolar. I either don´t care or go at it meticulously. That´s very indicative of my personality, by the way.
I just feel all I really need is a rough idea of how fuel consumption is affected by traffic conditions and my driving style, and I like to think that, over the years, it’s helped me become a more efficient driver. It also makes it easier to spot when something isn’t right like e.g. a failing oxygen sensor causing fuel consumption to increase dramatically. Beyond that, I can’t think of any way I would benefit from more precise or granular readings, so I’m not going to waste any time on it.
Sam: you get to the point. It probably is a huge waste of time. I recognise that mysef. I just don´t understand how you can resist obsessing about the figures. I wonder if my interest is derived from a scientific-education?
Or taking a scientific education was an outcome of your interest in numbers. This was my case.
So yes, it’s a waste of time, but a very delightful and satisfying one for me.
Don’t get me wrong, the trip computer only allows me to read one measurement at a time and it’s always on average mpg, and I’m constantly checking it….
Simon: there is a problem the causality, isn´t there? From my reading of scientific phillosophy the whole causality question is unresolved. I should have done an arts degree but my oafish, Anglo-Saxon background led me to believe arts could be done on one´s private time while science required training.
From a scientific-philosophical standpoint you might be right about the unresolved causality, but for practical purposes it’s often quite clear. My numerical interest arose as soon as I discovered that I had more than one finger. This preceded my education by several years. (Let’s not discuss the concept of time here.)
Being a reasonable, safe-playing Germanic man, it was thus obvious to go into some fruitful matter instead of studying unprofitable arts just out of pleasure.
I’ve worked on Richard’s point of oscillating between obsession and complete disinterest, with nothing in-between. Or possibly my stopping logging fuel consumption on my various guzzlers is the fear that my partner might come across my calculations and ask me how I could possibly be so profligate.