No matter what, my Forester always ends up calculating 23.4 MPG. I mean, I can reset the trip and it starts calculating MPG at a different number, but after a while it always "settles" on the same number. Anybody else had similar issues? Thanks.
MPG computer acting up?
('14-'18)
1 - 18 of 18 Posts
Mine always settles at 27.9. I tend to look more at the numbers at the end of the trip when shutting the car off. Those vary quite a bit. But I think it's just the law of averages as the more miles you drive, the more the fluctuations are absorbed into the average. Maybe?
The mpg calculator is a cumulative average so if you are driving the same roads under the same conditions most of the time you are probably going to end up with the same fuel economy result. You can test this theory by increasing the psi in your tires by a few pounds (but do not exceed the max psi on the tire's sidewall) and see if you get more than 23.4 mpg on your next tankful of gas.
I always have different numbers - depend on weather, traffic and other things.
I have about 29-30 MPG for my weekdays commute at Summer and 28-29 MPG during Winter months (computer). Usually I end up with 29.1 - 29.4 - 29.7 at the Summer.
Going for longer trips changes the mileage as well. It can be ~27 MPG for running fast (80 MPH) on the plains or be ~32-34 MPG when I am driving mountain roads.
If your computer always ends up with the same numbers, I guess your driving is very stable and you are going same routs day after day...
I usually calculate "real" MPG based on mileage and gas I put in (have a file on my PC and fill it with numbers each time I refill the gas). Computer usually adds ~2 MPG, so my calculated MPG for scenarios above is ~27-28 and 26-27 (Summer and Winter). Never trust the computer :crying:
BTW, my 2011 Forester had better calculation - less than 1MPG off. 2015, 2016, 2017 and 2018 are all the same - roughly 2MPG up.
I have about 29-30 MPG for my weekdays commute at Summer and 28-29 MPG during Winter months (computer). Usually I end up with 29.1 - 29.4 - 29.7 at the Summer.
Going for longer trips changes the mileage as well. It can be ~27 MPG for running fast (80 MPH) on the plains or be ~32-34 MPG when I am driving mountain roads.
If your computer always ends up with the same numbers, I guess your driving is very stable and you are going same routs day after day...
I usually calculate "real" MPG based on mileage and gas I put in (have a file on my PC and fill it with numbers each time I refill the gas). Computer usually adds ~2 MPG, so my calculated MPG for scenarios above is ~27-28 and 26-27 (Summer and Winter). Never trust the computer :crying:
BTW, my 2011 Forester had better calculation - less than 1MPG off. 2015, 2016, 2017 and 2018 are all the same - roughly 2MPG up.
Remember that most Foz have up to five different MPG calculations.
From my 2014 2.5 I Limited - the multi-function display at the top of the center stack. It shows four different MPG calculations: instant, current trip, trip odo A, trip odo B, regular odometer...
![Electronics Technology Electronic device Audio equipment Display device]()
* instant [top center] - what I'm getting right now.
* trip [left of triple meter] - what this drive is averaging from the last start.
* trip A (or B)
From my 2014 2.5 I Limited - the multi-function display at the top of the center stack. It shows four different MPG calculations: instant, current trip, trip odo A, trip odo B, regular odometer...
* instant [top center] - what I'm getting right now.
* trip [left of triple meter] - what this drive is averaging from the last start.
* trip A (or B)
- what the average is since the last reset of A or B.
* odometer
* odometer
- what the average is over the TOTAL MILEAGE since 0.
Not sure if that resets based on battery disconnect or not.
But all of these discussions about your MPG will never be perfect. Why? Because even if you are manually doing the math every time you fill up, you're never going to get a 100% accurate reading because you will never be able to have the exact same amount of gas in the tank on this fill-up versus the last versus the last versus the last.
The trip computer MPG is a bit more accurate simply by virtue of it doesn't matter how much gas was in the tank or is in the tank but it is more based on the calculations of what the CPU uses.
Not sure if that resets based on battery disconnect or not.
But all of these discussions about your MPG will never be perfect. Why? Because even if you are manually doing the math every time you fill up, you're never going to get a 100% accurate reading because you will never be able to have the exact same amount of gas in the tank on this fill-up versus the last versus the last versus the last.
The trip computer MPG is a bit more accurate simply by virtue of it doesn't matter how much gas was in the tank or is in the tank but it is more based on the calculations of what the CPU uses.
The station fuel pump lets gas flow while air is being drawn through a hole by vacuum until gasoline in the tank rises high enough to block the hole.
The mechanical linkage in the handle senses the change in suction and turns the nozzle off.
This is a very consistent process with insignificant variation between gas pumps sensing the fuel level, and evert pump made has to meet the same design standards.
Pumps are also regularly checked for accuracy by local weights and measures, so these tend to be pretty accurate.
The assertion that "The trip computer MPG is a bit more accurate simply by virtue of it doesn't matter how much gas was in the tank (that doesn't matter) or is in the tank but it is more based on the calculations of what the CPU uses" reminds me of a quote from Yogi Berra (Not Yogi Bear)… "In theory there is no difference between theory and practice. In practice there is"... because while theoretically the computer could be more accurate, in practice it isn't.
For the computer calculation, there are a large number of variables which can conspire to cause error, and they do.
If you really want to check on your CPU mpg accuracy, calculate your mpg after a refill at the same pump.
The mechanical linkage in the handle senses the change in suction and turns the nozzle off.
This is a very consistent process with insignificant variation between gas pumps sensing the fuel level, and evert pump made has to meet the same design standards.
Pumps are also regularly checked for accuracy by local weights and measures, so these tend to be pretty accurate.
The assertion that "The trip computer MPG is a bit more accurate simply by virtue of it doesn't matter how much gas was in the tank (that doesn't matter) or is in the tank but it is more based on the calculations of what the CPU uses" reminds me of a quote from Yogi Berra (Not Yogi Bear)… "In theory there is no difference between theory and practice. In practice there is"... because while theoretically the computer could be more accurate, in practice it isn't.
For the computer calculation, there are a large number of variables which can conspire to cause error, and they do.
If you really want to check on your CPU mpg accuracy, calculate your mpg after a refill at the same pump.
I guess it depends on how OFTEN the pumps are checked. My sense is that this happens once a year or even less often.
Since you know more about fuel pumps than I ever will, can you suggest why I'm seeing this up and down the East Coast since the weather turned cold:
After inserting the nozzle into the filler neck, you squeeze the handle. The price and gallon dials on the pump start spinning, but there is no sound of gas going into the tank and you can't feel fluid moving in the hose. Somewhere between .75 and 1 gallon, you suddenly hear the sound of fuel starting to enter the tank and you can feel the hose jerk slightly as the fuel starts to go in. Everything is normal from that point on.
The lag between the dials starting and the fuel flowing varies slightly between pumps at the same station and between stations, but almost every pump has some. Again, this seems to be associated with how cold it is out.
You all quite right about MPG, fuel pump accuracy and other stuff. They are not perfect.
But, speaking of MPG - this relies only on calculations of how much fuel or air passed to the engine. The system knows just a little about how many gas you put into the tank and don't use this info to calculate MPG.
Gas pumps are different story, and they are also not perfect.
But... I am doing these calculations after every time I fill the tank. I use miles between visiting gas station and how much gas I bought and let my computer calculate MPG. Yes, every single calculation can be off because of weather, pump accuracy and other factors, but average calculation on the span of over 60K miles is averagely good.
Also, comparing car calculated MPG with my numbers I always see the same difference between those two. And that means car calculates it "wrong" but accurate.
As I said - I always see car computer shows about 2 MPG better than my calculation after every trip, and I also see about the same 2 MPG difference on long time calculation (simplify this - I reset ODO A after every gas station visit and have ODO B running for thousands and thousands, so A shows me my MPG for short period of time and B average from the time I bought the car).
So this car calculation can be easily adjusted by small tweaking of the firmware. And as I mentioned above - Forester 2011 had much closer algorithm than next generation. And Foresters 2014-2018 are pretty consistent - my 2015, 2016, 2017 and 2018 all gave me about the same difference. So my assumption is - build in algorithm is bad.
As again - I don't rely on single calculation but on array of calculations over a years of use and thousands of miles.
But, speaking of MPG - this relies only on calculations of how much fuel or air passed to the engine. The system knows just a little about how many gas you put into the tank and don't use this info to calculate MPG.
Gas pumps are different story, and they are also not perfect.
But... I am doing these calculations after every time I fill the tank. I use miles between visiting gas station and how much gas I bought and let my computer calculate MPG. Yes, every single calculation can be off because of weather, pump accuracy and other factors, but average calculation on the span of over 60K miles is averagely good.
Also, comparing car calculated MPG with my numbers I always see the same difference between those two. And that means car calculates it "wrong" but accurate.
As I said - I always see car computer shows about 2 MPG better than my calculation after every trip, and I also see about the same 2 MPG difference on long time calculation (simplify this - I reset ODO A after every gas station visit and have ODO B running for thousands and thousands, so A shows me my MPG for short period of time and B average from the time I bought the car).
So this car calculation can be easily adjusted by small tweaking of the firmware. And as I mentioned above - Forester 2011 had much closer algorithm than next generation. And Foresters 2014-2018 are pretty consistent - my 2015, 2016, 2017 and 2018 all gave me about the same difference. So my assumption is - build in algorithm is bad.
As again - I don't rely on single calculation but on array of calculations over a years of use and thousands of miles.
@mb133 - The fact that you have consistent results (your noted 2 mpg variance between manual calculation and computer) says something about the process used.
If there were inaccuracies introduced by the pump, you would not get a consistent calculation compared to the computer - These would vary... but they didn't.
I'm thinking there is a good reason for a 2 mpg variance being reported by the software - The user believes they are getting better mpg... and if they never check with a manual calculation, there is no reason for them to believe otherwise, but that doesn't mean the number is right.
One of the tenents of evaluation of a process is a confirmation of the measurement system.
Your results proved that there was a difference between the two measurements, and the consistent difference proves that both measurements are repeatable, but what hasn't been proved is if either one or both is accurate.
While it is impossible to evaluate the components of the automated system without access to source software and test results of the individual components providing data to that process, it is relatively simple to ascertain the manual calculation components, as these can be independently verified.
Mileage by following a known distance and volume of fuel using a measured container. The calculation itself is immune from variation.
Many roads have speedometer/odometer checks, so it the car's odometer is correct when compared to the mileage on a known course, the only variable is the fuel volume. A couple of translucent marked volume cans could easily provide that.
Myself, it's not worth the effort to do the experiment. I know my odometer is correct, and I've added fuel just after weights and measures verified a local pump.
My car's computer says I get 2 mpg more than what I measure, and that has been as consistent as your results.
It therefore appears that the computer is incorrect, and perhaps intentionally so, as there is a good reason for the system developer to skew the results to the positive. There is no law stating that the car's internal data reporting system is accurate.
If there were inaccuracies introduced by the pump, you would not get a consistent calculation compared to the computer - These would vary... but they didn't.
I'm thinking there is a good reason for a 2 mpg variance being reported by the software - The user believes they are getting better mpg... and if they never check with a manual calculation, there is no reason for them to believe otherwise, but that doesn't mean the number is right.
One of the tenents of evaluation of a process is a confirmation of the measurement system.
Your results proved that there was a difference between the two measurements, and the consistent difference proves that both measurements are repeatable, but what hasn't been proved is if either one or both is accurate.
While it is impossible to evaluate the components of the automated system without access to source software and test results of the individual components providing data to that process, it is relatively simple to ascertain the manual calculation components, as these can be independently verified.
Mileage by following a known distance and volume of fuel using a measured container. The calculation itself is immune from variation.
Many roads have speedometer/odometer checks, so it the car's odometer is correct when compared to the mileage on a known course, the only variable is the fuel volume. A couple of translucent marked volume cans could easily provide that.
Myself, it's not worth the effort to do the experiment. I know my odometer is correct, and I've added fuel just after weights and measures verified a local pump.
My car's computer says I get 2 mpg more than what I measure, and that has been as consistent as your results.
It therefore appears that the computer is incorrect, and perhaps intentionally so, as there is a good reason for the system developer to skew the results to the positive. There is no law stating that the car's internal data reporting system is accurate.
@DragonSubie7 - Sometimes the measurements are off the balance and this is just confirms we are all right - the pump can be wrong or car computer failed. But ~95% of measurements are consistent and I consider it is good enough.
I assume these "2 MPG" are intended and help in case if your tires are wore out or you use slightly different size (ODO will give you wrong distance) or the temperature changed or any other things car system cannot adjust to. And you are right - most of the users don't bother by playing with the numbers and just trust the display. And this is fine, I guess.
I also agree it's not worth the effort to do the experiment. Reseting the odometer and keeping the receipt after I filled the tank and enter numbers into Excel file takes just a minute and become a habit for me (I know, this is odd, but as I said - takes just minute). So I don't care much about this, just keep doing (maybe I am crazy
).
I prefer speedometer to show higher speed and display better MPG than vise versa... The most important is to have consistent result even if this number is not perfect.
I assume these "2 MPG" are intended and help in case if your tires are wore out or you use slightly different size (ODO will give you wrong distance) or the temperature changed or any other things car system cannot adjust to. And you are right - most of the users don't bother by playing with the numbers and just trust the display. And this is fine, I guess.
I also agree it's not worth the effort to do the experiment. Reseting the odometer and keeping the receipt after I filled the tank and enter numbers into Excel file takes just a minute and become a habit for me (I know, this is odd, but as I said - takes just minute). So I don't care much about this, just keep doing (maybe I am crazy
).I prefer speedometer to show higher speed and display better MPG than vise versa... The most important is to have consistent result even if this number is not perfect.
I drive a 2010 Forester. I check my tire pressure weekly. I fill up at the same station (usually the same pump) each time using the same procedure (setting hold-open latch on slowest setting for the last few gallons), I drive the same route 5 days a week (mostly freeway) and my weekends usually follow similar patterns. I reset the "A" odometer at every fill-up and log my miles, gallons used, etc in a book where I also log scheduled maintenance. Everything is equal, for the most part, barring fluctuations in additives in the gas, increasing restriction through air filter, etc. My calculated mileage runs between 23.2 and 23.6 mpg. I can deal with insignificant variances there. My instantaneous mpg has been anywhere from 9.0 to 30.0, though it typically stays around 24.9. At the end of the day, it's the calculated mpg that I rely on. I'm not a speed demon, I avoid rapid acceleration/deceleration and do my best to keep the mpg as high as I can. That being said, what sort of calculated mpg do you all get with similar habits?
Our 2015 shows a pretty consistent 25.2. Wife is the main driver and most of the driving is around town. When we take it for a longer trip, like 100+ miles, It's gained a bit to 25.5. But after a few days at home to goes back to the 25.2 amt.
My Foz only has mpg calculations based on each of the two trip meters, so there is no instant mpg display possible.
The 2012 (FB25/4EAT) consistently averages 24mpg displayed with trip B (I use trip B for oil change miles).
Trip A (for mileage between fill ups) varies a lot. Unlike you, I don't do a consistent anything drive wise, but I always fill up when I get gas.
Some trips are only slow dirt road local drives (they all start out that way!) at very slow speed because the roads are terrible.
To go anywhere where there are retail establishments, it's either a 50 or 100+ mile trip.
My mileage varies a lot, primarily due to speed (75 limit on the Interstate or 65 on the rural highway) and whether I'm going up the mountain or not.
At or below 65 with few stops and lack of hills (there is never any traffic) 30 to 32 is the reported average, with about a best of 35.
Primarily high speed up hill - easily drops to ~ 20mpg indicated.
Regardless of the type of driving, when I fill up my mileage/fuel consumed calculation is ~2 mpg less than indicated.
As @mb13 said, and I agree, a consistent result is more important, because you can always add (or subtract) your own fudge factor, and this becomes important if you are trying to decide whether to buy the expensive gas here vs. can I make it to the cheap gas there...
@winklerste @DragonSubie7 I don't care about "trip MPG" or "instant MPG", they are way off because of shortness and variables. I use only average MPG between visiting gas station. This is much easier
Also, when I had Forester 2011 - similar to yours 2010 and 2012, I had ~24 MPG on my usual everyday trips to my work - looks consistent with your data. When I got 2015, MPG immediately improved to ~27 on the same routes - say thanks to CVT and bye to 4-speed automatic transmission.
Later, when I changed the job and commuted longer and used freeway, my MPG jumped up to ~30-31 MPG and that was impressive. I changed job again and see ~28 MPG now.
I am also not a speed demon and prefer smooth acceleration / deceleration and do my best to keep the mpg as high as I can. It works well as I pretty satisfied with MPG I have. I tried filled Forester with premium gas instead of mid-grade, but I didn't find any difference either in a car dynamic or MPG. So if I have time - I go to Kingsooper for mid-grade, if I am in a rush - I have Costco with the premium just block away from my house
I use calculation pattern similar to yours - Trip A for mileage between fill ups and it is pretty consistent as my routine is mostly the same. When I break a routine and drive to the mountains - MPG changes accordingly.
I also write down mileage, gallons and use Excel spreadsheet instead of notepad to automate calculation.
Trip B - I use it for longer trips to see how many miles I put during the trip and I keep it running until next trip. So Trip B can easily be 3000-5000 miles and gives me a good idea of average MPG.
Also, when I had Forester 2011 - similar to yours 2010 and 2012, I had ~24 MPG on my usual everyday trips to my work - looks consistent with your data. When I got 2015, MPG immediately improved to ~27 on the same routes - say thanks to CVT and bye to 4-speed automatic transmission.
Later, when I changed the job and commuted longer and used freeway, my MPG jumped up to ~30-31 MPG and that was impressive. I changed job again and see ~28 MPG now.
I am also not a speed demon and prefer smooth acceleration / deceleration and do my best to keep the mpg as high as I can. It works well as I pretty satisfied with MPG I have. I tried filled Forester with premium gas instead of mid-grade, but I didn't find any difference either in a car dynamic or MPG. So if I have time - I go to Kingsooper for mid-grade, if I am in a rush - I have Costco with the premium just block away from my house
I use calculation pattern similar to yours - Trip A for mileage between fill ups and it is pretty consistent as my routine is mostly the same. When I break a routine and drive to the mountains - MPG changes accordingly.
I also write down mileage, gallons and use Excel spreadsheet instead of notepad to automate calculation.
Trip B - I use it for longer trips to see how many miles I put during the trip and I keep it running until next trip. So Trip B can easily be 3000-5000 miles and gives me a good idea of average MPG.
2017 Forester Premium ES CVT (Sold and replaced with Camry Hybrid)
Joined
·
249 Posts
Excellent work I love this thread! I have also noted the MPG discrepancy all the way from my 2008 Scion xB to my 2013 Crosstrek to my 2017 Forester. It seems that all 3 vehicles would report approximately 1.5 MPG higher than reality. What drives me crazy is why they can't just put how much fuel is remaining in the tank in gallons, we have these stupid gauges, just give me a number! That being said I run a bunch of crazy calculations in my head to figure it out sometimes. 2017 Forester is 15.9 gallons. There are 12 bars on the graph. There is fuel beyond the last bar and before the first. If I assign each bar a value of 1 gallon, I can have some sort of idea of what's remaining. Also Subaru's distance to empty number is absolutely useless, as it constantly recalculates. If someone has come up with an exact gallon range for each position on the fuel bar, it would be enormously helpful!
1 - 18 of 18 Posts
Join the discussion
