1. Attachments are working again! Check out this thread for more details and to report any other bugs.

Discrepancy in Average and Calculated MPG

Discussion in 'Gen 2 Prius Fuel Economy' started by perry470, Oct 17, 2016.

  1. perry470

    perry470 Junior Member

    Joined:
    Oct 9, 2016
    13
    8
    0
    Location:
    San Francisco
    Vehicle:
    2008 Prius
    Model:
    Two
    I noticed there is a slight difference between the Average MPG in the Consumption screen and my calculated MPG after a fill-up. For example, in my last tank I drove 447 miles and the Consumption screen shows 58.1 MPG. However, base on the 7.881 gallons that I filled up, my calculated MPG is actually 447 / 7.881 ~ 56.7 MPG, or 2.5% difference. It's not that big of a deal but I wonder is the average MPG from the Consumption screen always more optimistic? When you talk about MPG on these forums, are you referring to the true / calculated MPG?

    [​IMG]
     
  2. fuzzy1

    fuzzy1 Senior Member

    Joined:
    Feb 26, 2009
    17,035
    10,010
    90
    Location:
    Western Washington
    Vehicle:
    Other Hybrid
    Model:
    N/A
    Not bad at all. I wish the Gen3s were that close. Ours typically have 2-3X larger error.
     
    Mendel Leisk likes this.
  3. 05PreeUs

    05PreeUs Senior Member

    Joined:
    Oct 1, 2016
    1,097
    563
    0
    Location:
    MSP
    Vehicle:
    2005 Prius
    Model:
    N/A
    There is only ONE method to determine MPG and that is Miles driven divided by gallons used; aka: Pencil and Paper, just as you have done.

    The ECM uses miles driven divided by CALCULATED fuel used based on theoretical fuel flow (liters/hr * hours of injector PW), which inherently introduces tons of error.
     
  4. Kenny94945

    Kenny94945 Active Member

    Joined:
    Mar 24, 2016
    758
    387
    0
    Location:
    Sausalito
    Vehicle:
    2016 Prius v wagon
    Model:
    Four
    Well, speedometers/ odometers have a variance to actual.

    2.5% variance seems in line with a mileage to gallons used.

    As for your question, IMO probably both mileage calculations are presented by forum co-members using the statistic they find easiest to report on. For example '16 V average screen shows about (about) 41 mpg. I do not know the comparison calculated mpg.

    Great fun figuring fuel mileage!
     
  5. Mendel Leisk

    Mendel Leisk Senior Member

    Joined:
    Oct 17, 2010
    54,467
    38,102
    80
    Location:
    Greater Vancouver, British Columbia, Canada
    Vehicle:
    2010 Prius
    Model:
    Touring
    We watch a lot of detective shows in the aft, and "motive" is a big factor in solving cases. Now who would benefit from exaggerated display? Toyota?

    I agree with @fuzzy1 : our third gen has an unbudgeable lifetime error around 7.5%. It jumps around a bit, due to pumping vagaries, but 7.5% is where it sticks.

    upload_2016-10-17_9-29-0.png
     
  6. srellim234

    srellim234 Senior Member

    Joined:
    Jun 26, 2015
    1,192
    1,680
    0
    Location:
    Laughlin, Nevada
    Vehicle:
    2008 Prius
    Model:
    Touring
    perry - I find that the "on-screen" mpg is usually a little optimistic but the variance actually works both ways. To give you a comparison, here are the results from my last 10 fills:

    10/09/2016 - on screen 41.50, calculated 44.07
    09/26/2016 - on screen 46.20, calculated 40.76
    09/18/2016 - on screen 40.80, calculated 40.36
    09/10/2016 - on screen 42.10, calculated 44.94
    08/29/2016 - on screen 37.30, calculated 34.34
    08/14/2016 - on screen 44.90, calculated 46.73
    08/13/2016 - on screen 42.70, calculated 48.48
    08/01/2016 - on screen 43.10, calculated 37.39
    07/31/2016 - on screen 45.10, calculated 46.48
    07/27/2016 - on screen 44.10, calculated 42.02

    We don't have a regular fuel tank and the amount the tank will hold changes based on the flexibility of the tank's rubber bladder at any given time. I don't depend on the displayed mileage and instead just fill it up when it's convenient anytime after I see two bars left on the gas gauge.
     
  7. Mendel Leisk

    Mendel Leisk Senior Member

    Joined:
    Oct 17, 2010
    54,467
    38,102
    80
    Location:
    Greater Vancouver, British Columbia, Canada
    Vehicle:
    2010 Prius
    Model:
    Touring
    Man, that jumps around a lot, probably due to the bladder tank, as you say.

    upload_2016-10-17_13-26-22.png

    (bottom right number the average diff)
     
    srellim234 likes this.
  8. perry470

    perry470 Junior Member

    Joined:
    Oct 9, 2016
    13
    8
    0
    Location:
    San Francisco
    Vehicle:
    2008 Prius
    Model:
    Two
    Thank you all for the replies! I have a better idea now.
    However, this leads me to think that even calculated MPG is not accurate.

    Here's my reasoning: Assume the statement that the tank's capacity is flexible is true, then how would one know if they are filling up for the miles driven? The tank could have changed its capacity during the course when those miles are driven. For example, let say the tank becomes bigger compare to the last fill-up. That means I am dividing a bigger capacity over the miles driven, which result in a lower calculated MPG.

    I guess the ultimate questions are, how much does the tank's capacity fluctuate? How much confidence do I have in these figures?
     
  9. fuzzy1

    fuzzy1 Senior Member

    Joined:
    Feb 26, 2009
    17,035
    10,010
    90
    Location:
    Western Washington
    Vehicle:
    Other Hybrid
    Model:
    N/A
    You can erase that tank variability uncertainty by keeping track over many consecutive tanks, and computing the longer term average.
     
    srellim234, uart and Mendel Leisk like this.
  10. Mendel Leisk

    Mendel Leisk Senior Member

    Joined:
    Oct 17, 2010
    54,467
    38,102
    80
    Location:
    Greater Vancouver, British Columbia, Canada
    Vehicle:
    2010 Prius
    Model:
    Touring
    ^ This.

    Vagaries abound on single tanks, but the more the tanks you accumulate, the more accurate (and unbudgeable) the results. In other words: calculate individual tanks, say in a spreadsheet. Then off on the side do a calc that uses the total gas and total gas (typically two columns). That number will reflect the cumulative mpg, regardless of the ups and down due to varying tank capacity.

    That's why it'll take another year of sub 4.7 (liters per 100 km's) to knock my 4.9 Fuelly badge down to 4.8, lol.
     
  11. perry470

    perry470 Junior Member

    Joined:
    Oct 9, 2016
    13
    8
    0
    Location:
    San Francisco
    Vehicle:
    2008 Prius
    Model:
    Two
    It seems that's the way it goes then. I already signed up on Fuelly and am tracking my MPG over many tanks instead of individual fill-up.(y)
     
    srellim234 likes this.
  12. Mendel Leisk

    Mendel Leisk Senior Member

    Joined:
    Oct 17, 2010
    54,467
    38,102
    80
    Location:
    Greater Vancouver, British Columbia, Canada
    Vehicle:
    2010 Prius
    Model:
    Touring
    I would a spreadsheet too, if possible.
     
  13. 05PreeUs

    05PreeUs Senior Member

    Joined:
    Oct 1, 2016
    1,097
    563
    0
    Location:
    MSP
    Vehicle:
    2005 Prius
    Model:
    N/A
    Most "modern" vehicles have ODOmeters that are MORE accurate than GPS, speedometers not quite so much for reasons I do not understand - it should be simple.

    All modern vehicles are programmed to "pulses per mile", this is simply PPM = (# driveshaft teeth) x (axle ratio) x (tire revs/mile). When you start multiplying this out, you get big numbers like PPM = 16 x 4.34 x 613 =~42567.

    Let's say you "gain" 3 Revolutions Per Mile on worn tires compared to new. 3/613 sounds meaningful, but it's really ~0.5%, so you can't really attribute MPG changes to differences in rolling diameter of the tires. Inflation from the lower to upper threshold of a tire can affect rolling distance on the order of 2% or more and in fact, Ford used that to determine the potential for an underinflated tire for a number of years before TPMS sensors were affordable.

    Because the number of teeth and gear ratios are EXACT measurements that are elementary to make, the only variable is rolling tire circumference (aka: revolutions per mile).

    Because SPEEDometers also factor in time, a measurement with *some* error (more or less depending on the quality of the clock), the accuracy of the speedometer is inherently less accurate.

    If you want to verify your odometer, find a 10 mile stretch of highway with mile markers, drive it both directions after resetting the TRIP meter as you cross the starting point. If accurate, the trip meter should register "exactly" 10 miles; a one block error is 1% and typical. To check your speedometer, run the same course at 60 MPH indicated, you should cover the measured 10 miles in "exactly" 10 minutes. As always, convert to decimal and do the math for your actual speed in MPH.

    You don't KNOW you MPG unless you know your MILES and with more miles and gallons the more accuracy.
     
  14. fuzzy1

    fuzzy1 Senior Member

    Joined:
    Feb 26, 2009
    17,035
    10,010
    90
    Location:
    Western Washington
    Vehicle:
    Other Hybrid
    Model:
    N/A
    Time can be measured more accurately than anything else, and 0.005% accuracy costs just pennies. For these purposes, time measurement errors are negligible compared to other errors.

    Part of the speedometer display 'error' is intentional, to comply with laws and industry standards forbidding it to read low. Because other measures in the calculation have some uncertainty, an error distribution is present. To keep units on the low side of the error distribution in compliance with the requirements, the system is biased high. Euro-market vehicles are commonly biased higher than U.S. market vehicles, originally by industry regulations, now mostly just to keep in line with their respective histories.

    Odometers are now commonly biased to read low. If they had the same high bias as the speedometers, product liability lawyers would swoop back in with more class action lawsuits claiming warranty fraud. The newest car in my household is biased noticeably low, after its predecessor was included in such a suit and settlement. (My particular car did not display the alleged issue.)
     
    #14 fuzzy1, Oct 18, 2016
    Last edited: Oct 18, 2016
  15. Mendel Leisk

    Mendel Leisk Senior Member

    Joined:
    Oct 17, 2010
    54,467
    38,102
    80
    Location:
    Greater Vancouver, British Columbia, Canada
    Vehicle:
    2010 Prius
    Model:
    Touring
    Editorial: first word in last paragraph was meant to be Odometer?
     
    fuzzy1 likes this.
  16. fuzzy1

    fuzzy1 Senior Member

    Joined:
    Feb 26, 2009
    17,035
    10,010
    90
    Location:
    Western Washington
    Vehicle:
    Other Hybrid
    Model:
    N/A
    Yup. Fixed.

    Low blood sugar today as I prep for a common senior 'test' tomorrow.
    That's my story, and I'm sticking to it.
     
    Mendel Leisk likes this.
  17. 05PreeUs

    05PreeUs Senior Member

    Joined:
    Oct 1, 2016
    1,097
    563
    0
    Location:
    MSP
    Vehicle:
    2005 Prius
    Model:
    N/A
    Agreed, but they don't, two small errors multiplied, make a larger one and that is why SPEEDometers have greater error in most cases than ODOmeters (they can have errors in opposite directions).

    Not. There are no laws that that forbid an Odometer nor Speedometer to read low as you assert, OEMs have for YEARS worked to get greater and greater accuracy. In point of fact, both on our 2016 Fusion read low, very slightly so(<0.5 block in 10 miles), but low, based on multiple physical measurements. I will agree that it is more common to a speedometer to read high than low, but over millions upon millions of miles, any trucking industry professional will tell you that odometers tend to read low.
     
  18. 09Prius2

    09Prius2 Member

    Joined:
    Aug 20, 2016
    201
    87
    0
    Location:
    FL
    Vehicle:
    2009 Prius
    Model:
    Two
    Mine has always been almost exactly 3mpg under whatever the readout says.
     
  19. 09Prius2

    09Prius2 Member

    Joined:
    Aug 20, 2016
    201
    87
    0
    Location:
    FL
    Vehicle:
    2009 Prius
    Model:
    Two
    Speedometers typically read 1mph higher than actual speed so that you are not speeding without being aware. Meaning if you drive 60 mph for one hour, you may have only traveled 59 miles, making the odometer reading lower than the speed displayed on the speedometer. Hence the inverse correlation.
     
  20. srellim234

    srellim234 Senior Member

    Joined:
    Jun 26, 2015
    1,192
    1,680
    0
    Location:
    Laughlin, Nevada
    Vehicle:
    2008 Prius
    Model:
    Touring
    We traveled southern California-New Mexico-Colorado-southern California this summer as well as southern California-Salt Lake City, UT-southern California a couple of weeks ago. We don't care for the Toyota GPS so we had a Garmin GPS unit with us for both trips. We found there was less than 1/4 mph discrepancy at multiple speeds when we compared the speedometer to the Garmin, with the car being 1/4 mph higher than the Garmin.