display vs. actual mpg section for green hybrid database
 

I've seen alot of discussion around this in the forums and was wondering if there is a place to enter in both your display mpg and actual calculated mpg.

I just threw together a spread sheet that will calculate the actual mpg based on the mileage and the amount of gas pumped at the fillup. My first tank was reported at 42.3mpg by the display over 380 miles. That comes to ~9 gallons of gas. At fillup I was only able to get 8.66 into the tank. That means that I actually got about 43.9mpg (380mi/8.66gal) which means the display is 3.6% off.

I'm not knocking the display at all, thats pretty accurate, but it would be nice to have the ability to enter in your calculated vs. display reported mpg.

Maybe we could have a little section for all hybrids that have a mpg display where people could enter in the amount of gas at fillup and the site could calculate the actual mpg and show a % difference between the actual and the display.

What do you guys think?

Re: display vs. actual mpg section for green hybrid database
 

Hi,
This is an area where folks hold strong opinions. Near as I can tell, the issues are:

• gas pump is calibrated and state inspected for accuracy
• Prius has variable size 'bladder' tank
• gas nozzle shutoff is uncalibrated

I think the consensus is use the pumps for long term averages and performance and the display for shorter range performance. My personal preference is the pump because "I pay the pump." Over long periods, 9,000+ miles, the averages look pretty good. Althought individual tanks vary, I prefer to average a set of tanks . . . until we get a handle on the variable sized, Prius tank.

Bob Wilson

Re: display vs. actual mpg section for green hybrid database
 

The Prius apparantely has some kind of 'bladder' in the fuel tank and can affect how much fuel goes into it, so it is not a valid reading.

I have also heard that the computer display is not quite accurate, that it errs on the high side.

I guess there is no real way to know your true mileage other than to know that you are doing great regardless, and the numbers you calculate are probably close.

Re: display vs. actual mpg section for green hybrid database
 

I guess the bladder type gas tank does prevent you from accurately calculating your mpg.

Toyota needs to implement an accurate gas tank gauge. That would put a stop to all our problems with calculating mpg.

Re: display vs. actual mpg section for green hybrid database
 

The variable bladder can affect your MPG rating from tank to tank. However, the weak law of large numbers proves this is irrelevant, as you add more tanks. As the number of tanks in the sample increases, the average mileage will approach the real mileage, despite the varying tank size.

Compute your MPG by miles driven divided by gallons of gas you put in the tank. With a sufficient sample size, gas tank size, weather, traffic conditions, and all the other tank-to-tank variables will average out and you will approach your true lifetime MPG.

That's the math of it. On an anecdotal note, I believe the display has a sampling error, which expresses itself more in stop and go traffic, where you're constantly changing speed. On my usual suburban/city commute, with errands thrown in, my display is always 4-6% higher than the computed value. When I drive to the beach, going 300 miles at a time without constant speed changes, the computed and displayed MPG are less than 1% apart.

Re: display vs. actual mpg section for green hybrid database
 

There really is no way to know exactly what you mileage is on any given tank. Fill differences cause the calculated figures to be off. Displayed figures tend to be a hair optimistic. Even if you filled your tank to exactly the same level and even if the pump was perfect, your trip meter/OD could be off an muck the whole thing up anyway.

Re: display vs. actual mpg section for green hybrid database
 

Well, I pick up my new 06 HF3 this afternoon, and I already have an excel spreadsheet on my phone to track mileage, gallons, computer MPG, and MFD-reported MPG. Once my GreenHybrid signature starts working, I'll supplement it with the MFD-reported data. I'd like to see a section for that online, as more data gives a better snapshot of the overall picture.

Re: display vs. actual mpg section for green hybrid database
 

I've been posting my MFD readings in the notes along with my tanks for a while now. At first they seemed to be pretty close, but over time the discrepancy caused me to start tracking it.

Re: display vs. actual mpg section for green hybrid database
 

The imroved accuracy of mpg based on multiple tanks. One is comparing "full" to another "full", which due to the bladder and pump shut-off variability may differ by 2 gallons (or some similar number). As you buy enough intervening gallons, the % error from that variabilty will shrink. For example buy 100 gallons and that uncertainty is 2%. It was already noted that the "fullness" attained in the intervening tanks matters not at all. Just the gallonage. This is somewhat different from the concept of sample size. Consider that hair to be split.

If there is a separate variability due to pump shut-off, this ideally will do a central-limit-theorem thing over time (refills) and converge to zero.

The screen mpg generally appears to be slightly optimistic, but not the same for everybody. Apparently no one has checked if that differential changes on a single vehicle through time, or vs. some plausible external variable (like temperature). I don't think a satisfying explanation for that optimism has been advanced.

Foo monkey suggests sampling error, which may be it, or close. I notice that the indicated fuel injector milliseconds does not always go to zero when proceeding "ICE off", and wonder if that might be involved also.

If gas stations lie to you about gallonage, you are screwed. I believe that this has happened to me on at least 2 ocassions out of 213 full or partial refuelings.

DAS

Re: display vs. actual mpg section for green hybrid database
 

MFD / Hand calculated (used for GreenHybrid database)
46.2 / 45.8
45.9 / 44.7
43.5 / 42.5
45.1 / 48.8 (MFD was lower! Very unusual)
45.6 / 45.5 = Total so far