Incorrect mpg
 

Hiya, if you look at my fill ups I have to average mpg 25.2 and 24.6, yet my dashboard claims I am getting 31.6mpg ....

What gives?

Thanks

26 is the US MPG average. 31.6 is the UK MPG average. Make sure you click the 'UK' link under the Fuelly logo and reload the page.

Hi,

Still doesn't make sense, seems more that something is calculated wrong.

I have been using the UK link thanks, but on my fuel up stats under Log Book it says "UK MPG" and I have two results, 25.2 and 24.6. My car has never managed 30+mpg that I know of.

Anyway my badge shows 31.6 UK mpg. I believe my partial fuel up hasn't been accounted for properly.

If I do the calculations myself I get 21.16mpg and 24.57mpg (both UK), so the log book is wrong, but the badge is way out.

I think there is something wrong with your script, like it's doing a US to UK conversion on my UK mpg data?

Anyway it's making my car club think I drive like an old lady :)

Thank you for the quick response, and great site, really like it thank you :)

Whilst I'm testing, I go to the Signature page, and it says 31.6 UK MPG, if I change it to US MPG it still says 26.3 which is too high again.

It seems more that the partial fuel up I did has been used for the average calculation which wouldn't work.

Hope that helps,

Russ

Just re-read my first post - I meant "the log book is right", not wrong lol my bad, sorry.

It's a good idea to try to avoid partial tanks since they are useless for calculating mpg (it takes fuelly 2 proper tanks to recover). Just fill it to the top each time.

Thanks for the heads up on it taking time to recover, but I'm not going to change my fuelling to suit Fuelly.

Fuelly just needs to be programmed better to not go wrong when it comes to a partial fuel up and just ignore it when calculates averages :)

Thanks

Just saw you're getting 75mpg! Crazy, good for you! I'll pay more attention to A2s when someone is looking to change their car :)

Fiat Punto and Kia Rio do 97 and 94 MPG if youre trying to maximise economy btw!

It looks like you have some fuel-ups out of sequence. Take a look at your View All Fuel-ups page here. Your fuel-up on September 24 has an odometer reading that is lower than your previous fill.

The average MPG calculation uses total distance driven and total fuel volume filled. So that could be why you're seeing a calculation that looks too high.

Thanks pb, I'll look into it in a bit :)

Draig, if I wanted to get good economy I'd probably aim more towards a 500, they look fun :)

Pb, my odometer reading is fine, it went to 45k kilometres so is in sequence, so it seems to be down to bad calculating again. Hopefully given another ten fuel ups it'll be right again. Just about to fill it up again so I'll see what that does. Thanks

oh yeah, I see that now. huh, misread it.

So it looks like you've gone 1518.8 km on 135.7 litres. That gives us 31.6 MPG.

Ok I just put another fuel up in, and we're back to 24.9 which is right :) Seems that the partial fuel up confused it. There is no way in the world my car will do over 30mpg unless you're pushing it. I'll look at the numbers later to see how it's getting 1518.8km and 135.7 litres to see where the discrepency is, thanks.

Great site, and thanks for all your help :) BTW not sure it's anything to do with you but the mobile site is no longer allowing me to enter decimal places on fuel ups, it's not available on my Galaxy S3. I don't know if that's an S3 fault or something changed on the mobile version, but thought I'd say.

Thanks again

Ok I did the math and I can see that the last partial fuel up was what confused it :)

So ideally if your last fuel up is partial then the average calculation should ignore it, but otherwise wait till the next full fuel up and it's good to go.

Also if I was being a real pain my last full fuel up could get it's mpg from the full fuel up + the last partial and the odometer difference since then. I didn't get a mpg reading for the last one.

Anyway not criticising, lots of great features on the site, I just found the 300ZX comparison page which was quite interesting.

Thanks again

On the Galaxy S3, yeah there isn't anything we can do about that on our end. You can bring up this page on your phone:

https://m.fuelly.com/disable-numeric-keypad.gas?id=1

This will set a cookie that tells Fuelly Mobile to use the standard keypad for entering numbers. That's about the best we can do on this end.

When you mark a fuel-up as partial you won't get an individual MPG for that fuel-up or the next one. Once you have two full fuel-ups in a row you'll start getting that number again. The reason is that Fuelly calculates that number using the difference from full. When you add a partial, there's no full tank volume to calculate from. And if we just use the partial tank, it makes the tank look bigger than it is in reality—you'd get a misleading number.

Here's more info: Why do I have to fill the tank all the way up every time I buy fuel?

Pb, thanks for the fix, much appreciated.

I was just thinking why not aggregate partial fuel ups with the next full tank, along with mileage? You could run on half a tank for ten fill ups, then do a whole tank and find out how that changes economy.

Anyway I won't take any more of your time. Thanks again, thanks for the site.

I already suggested that, but it was rejected (they want to keep things simple, one tank = one mpg figure).

Hmmmm, well they've built a decent site I can use for free so I'm not in a strong position to argue, and they know what development they have on the books already so maybe they don't have resources for it, but I think it's a good idea as well :)

Fuelly measures distance using the difference from full. If you do a bunch of partials in a row, how do you know the distance traveled on that fuel? There's no way to do it accurately.

Paul, Russ is referring to a series of partials followed by a full (and in this scenario fuelly does have enough info to work out the total fuel volume. It is only when the full tank is done that this can be calculated).

Yeah, that's my point, ignore them until you get a full, then just count it as one large tank. This site looks good, I'm sure by using some nice bordering or something you could show that the fill ups are counted as one.

I think this would be a very neat feature, and not something I've seen before. When trying to be economical in my last car I would only ever carry half a tank, saved about 40kg weight all the time dragging that up and down hills and made a difference. I was unable to use an app/website to get my mpg though and had to rely on doing it myself :)

I hear what you're saying MMUK and RussTaylor. But several partials, even followed by a full, will give you incorrect distance information. My point is that you can't count them as one big tank because the size of the tank matters for calculating distance. The difference from full is going to be off if you include partial fuel-ups. The only point we have to measure from is a full tank.

What people need to realize is that when you use a partial fill after a full tank, that amount isn't how much gas you used to go the distance on your odometer, it is how much gas you put in your tank. It will be less than the amount you used.

When you do two partials in a row, you will never know how much gas you did actually use and you have zero accuracy in determining how much gas was used to go the distance you are accounting for.

"Difference from full" gives a point of reference for calculating fuel economy. With partial fill ups, there's no point of reference to use for measuring the amount of fuel consumed.

The only way around this would be for the user to manually determine how much fuel was consumed without filling the tank. This isn't a matter of adding a feature to Fuelly; any fuel economy calculations require good data, and partial fill ups don't allow for this on standard vehicles. The only somewhat practical way around this would be to install a very accurate, digital fuel gauge in your vehicle. Doing so would allow manual calculation of the amount of fuel consumed. In my opinion, it's better to just fill up the tank if you're really interested in maintaining a fuel economy log.

"Yeah, that's my point, ignore them until you get a full, then just count it as one large tank. This site looks good, I'm sure by using some nice bordering or something you could show that the fill ups are counted as one.



I think this would be a very neat feature, and not something I've seen before." - RussTaylor

Sorry, but the reason you haven't seen it before is because it can't work for the reasons Ciderbarrel and I mentioned.

I'm sure you're right, and I'm missing some subtlety, but can you explain where this doesn't work...

I have filled my car up, it took 50ltrs (doesn't matter). On my odo I have 100 miles. So starting point = full tank and 100 miles on odo.

I drive around, and do a partial fill up, so 10 litres, my odo reads 110 miles.

I do the same again, this time I fill up 20 litres, and my odo reads 140 miles.

I then drive around, and perform a full fill up which takes me 30 litres, and on my odo I have 200 miles.

Since the last full tank (our reference, no different than my very first fuel up on Fuelly) I have travelled 100 miles, and used 10 litres + 20 litres + 30 litres.

This is the amount of fuel I have put into the tank to return to our reference point (full tank). So 100 miles and 60 litres...

Sorry to keep bashing the point but I disagree with both of you. You say you have no point of reference, but you do from your last full fill up. You have how much you put in the tank in between, and how much it took to fill it up again.

Just because you put it in the tank, it doesn't mean that is how much fuel you burn unless you let to go to empty (the engine dies from lack of fuel in the tank).

You are confusing fuel purchased with fuel used.

Starting point is good 50L and 100 miles.

You drove 110 miles total and used 50L plus some unknown amount, it was more than 10L because you stopped at 10L. You have used 50L + X, the unknown amount

You drove 140 miles total and used 60L + X and now + Y, another unknown amount. You have no idea how much was used because you did a partial fill twice in a row. Did you use 20L? 30L? You will never know for sure.

You drove 200 miles total and used 60L + X + Y and now + Z, another unknown amount.

This is why a partial fill gives you ZERO accuracy for fuel economy.

Picture it this way. You have a 50L Coke bottle (huge, I know). You can not see through the wrapper, nor shake it to test the amount in it, nor weigh it.

You top it off, so it has 50L. You drink some and fill it back to full.

You then drink some amount, not keeping track, in a week and put only 10L back in and it is not full. You then drink SOME more, but you only put 20L back in. It is still not full, and then you drunk more and put 30L back in, it is still not full.

Tell me, how much Coke did you drink? You have NO way of knowing because you have no reference point for your measurements.

Oh man, we're going to go around and around in circles.

"You drove 110 miles total and used 50L plus some unknown amount, it was more than 10L because you stopped at 10L. You have used 50L + X, the unknown amount"

It doesn't matter, you know how much it takes to return it to full, therefore you know how much you used. You drove and used it, and you filled it back up, you know how much you used by how much it took to fill up.

X + Y + Z is the sum of all the partial fill ups + the amount required to return it to full.

So we have a huge coke bottle (say 75ltrs), and it took an amount (doesn't matter) to fill it up. We now have a full coke bottle, this is our reference.

You drink some, a bit more, you don't know how much but to help this discussion we'll say 10ltrs. So you have 40ltrs in there, you don't know that. You do the same two more times, you use another 20ltrs and now you're down to 30ltrs. But you don't know any of these measurements, because it doesn't matter.

So you have your full - x - y - z.

You then put 5ltrs in.

You drive around some more, so Full - x - y - z + 5ltrs - a.

You drive some more, then partial fill up (10 ltrs) then drive some more. So Full - x - y - z + 5 - a - b + 10 - c.

By now you've no idea what you have, apart from whatever remains of your original tank + 15 ltrs.

You fill it up (returning to our reference), this tanks 55ltrs.

You now know x + y + z + a + b + c = 55ltrs + 15ltrs = 70ltrs.

I have a way of knowing, because I have TWO references, both times the tank was FULL.

I will know how much coke I have drunk as soon as I fill the coke bottle right to the top. The amount I drank is 10L + 20L + 30L + whatever it takes to return it to the top.

Sorry Ciderbarrel, but I have to agree with RussTaylor on this one. In his example he wrote "...and perform a **FULL** fill up..." (emphasis added). In your rewriting of his example, this time with the giant coke bottle, you wrote "...and put 30L back in, it is still **NOT** full." (emphasis added) It looks like you weren't trying to deceive but just didn't read RussTaylor's example correctly.

In the examples above you don't have to know what the individual numbers are for x, y, and z, you just have to know that x + y + z = 60L.

I would say don't worry about adding partial fill-ups just aggregate the amounts on your own. Sure, you might go 800mi and add 80L to a 60L tank (or some other made up numbers) but you would get an accurate mileage for that time frame. This would really mess up trending though so I almost never do it even before I started using this site.

Oh crap, yeah I totally missed that last full fill up in his post. That changes EVERYTHING. Ignore my last post because that is way off since I missed that last fill statement. I'm sorry I missed that.

Russ's example still doesn't jive because a 10L partial and a 20L partial won't equal a 30L because of the used fuel between 1st full tank and the last full tank.

The 50L for the first tank and the 30L for the last are valid. 10L of fuel was used because that 1st partial will ALWAYS be less than what was actually used, so 90L out of some X number of fuel was used.

Some more driving was done and some more fuel was burned and 20L was put back into the tank. Was 20L of fuel actually used? That is the point I'm trying to make. A partial after a partial doesn't tell you how much fuel was used, only bought. It isn't going to be as simple as Russ's example where his math somehow works out where his next full tank EXACTLY adds up to the last two partials.

Simple logic will tell you that putting in 10L and not hitting full and then 20L more and not hitting full means that the next full tank will take MORE than 30L because between the two partial fills he used MORE than 30L.

I hope I'm not over complicating things but here is another stupid example from me: a 100L tank (to keep the math simple.

Tank was full after the 50L fill, so there is 100L in it.

Some driving done, 10L put back in, not full. For arguments sake, lets say it has 99L in it. That would be a best case scenario. 10L bought, 11L used

Some more driving done, 20L put back into it. Again, let us assume that it is back up to 99L. 30L bought, 32L used.

Buying 30L can not get it back up to 100L. That is my point.

Also, if we want to avoid all this guessing, just top off your tank. I know in the UK the gas is way more expensive than here in the States, but it is the best way to be as accurate as possible.

Ok I'm glad we're getting somewhere.

Fuel prices are different everywhere, I'll stick £10 in before I go home tonight and then do a full tank then (£80). Also as this is a website about fuel economies you can increase your economy by not driving around with a full tank. I think this is a fairly key point for this site.

Anyway Cider your new example isn't right either, sorry :)

"Tank was full after the 50L fill, so there is 100L in it.

Some driving done, 10L put back in, not full. For arguments sake, lets say it has 99L in it. That would be a best case scenario. 10L bought, 11L used

Some more driving done, 20L put back into it. Again, let us assume that it is back up to 99L. 30L bought, 32L used.

Buying 30L can not get it back up to 100L. That is my point."

No, but you do not have a full tank. So do a FULL tank fill up at the end and tally it all up :)

You don't specify how much you used in the second bit of driving, but as you're now up to 99ltr having put 20ltr back in, it must have been down to 79ltr. So you used 20ltr. So 30ltr of fill ups, you've used 31ltr, you don't know how much is in the tank but you said it was 99ltr. So top it up to a FULL tank (our reference) = 1 extra litre. So 31ltrs.

Basically there is no magic, whatever you put in is either in the tank still or has been used. I suppose simply put

If I have a full tank, drive 100 miles, and put 50 litres of fuel in to get back to full

is that any different to

I have a full tank, drive 50 miles, put in 25 litres, drive fifty miles, put 25 litres of fuel in

Assuming both examples return the same economy either way I've driven 100 miles and consumed 50ltrs. One is one full fill up, one is one partial fill up followed by a full fill up.

Anyway sorry to write so much and it gets very wordy, back to my original point I'd love partial fill ups to count as most of the time people here don't do full fill ups. If we did it once in a while to get an average that'd would suit us well :)

Thanks again

Once you have a LOT of tanks in the system, the running average will be reasonably accurate regardless of whether there are partials present or not (since the 50 litres error would become less significant overall). It's the individual tank MPGs which would be missing.

Personally I never do partials. I know my range and which are the cheaper stations in the areas I normally drive.

I climb 1200 feet to get to work each day, so if I run a half tank I save myself 50kg in weight dragging it up the hill, and seems to make a difference :) I'd only ever do part tanks if I had the choice.