Just my opinion, but I don't think ethanol has any reason to do with the cars computer calibration, as you'll get similar discrepancies the World over, in most cars of all makes and models using all different types of fuel.
Mileage is a pretty simple calculation: (miles driven) / (gallons consumed). The energy content of the fuel does not matter -- except that one will typically drive fewer miles per gallon for a lower energy fuel.
There are various places where errors can creep into the mileage calculation. Consider the "miles driven" part. The car should know exactly how many miles were driven from the odometer (by counting rotations of the (front) wheels) which feeds directly into the car computer. Assuming you are using standard sized tires. Using larger or smaller tires will make a difference how far one rotation will take you. The same problem can occur for seriously under inflated tires of standard size.
Then there is the issue of how many gallons of fuel were consumed. Two ways that might be used to measure that would be: 1) reading the fuel gauge; or, 2) using an inline flow meter. My guess is that Toyota uses data from the fuel gauge. Maybe the fuel gauge overestimates the fuel left in the tank? Past experience suggests that fuel gauges are non-linear. That is, the needle does not move uniformly from F to E over the course of a tank. Unless the mileage calculation takes this into account, an error will be introduced.
Finally, if the gas tank is not filled up completely during a fueling stop, the calculation for that tank will suffer. That is totally the job of the owner. Except in Oregon, where only gas station attendants are allowed to pump gas!
It seems that folks who report discrepancies observe that the car computer _overestimates_ the mpg. Perhaps this is an intentional and clever marketing ploy. While there are legitimate sources of error, it is a bit surprising Toyota did not do a better job with the mileage computer.
Miles should be the accurate part of the calculation. The thing to consider when thinking about it is that the cars that computer mpg usually also show an "instantaneous mpg", along with "average mpg"
The instant one shows 0.0 mpg when not moving, to 99.9 mpg when you let off the gas completely over like 20-30 mph.
From what I can tell, the average is calculated from the instant over time. Errors come in, with the biggest example being that zero mpg time, with no miles driven. The computer has a bunch of seconds with a zero reading(where the car went zero miles, and used a varying amount of gas per second) a bunch of seconds with 30-40 mpg, and a bunch with high mpg(coasting). Add them up, and divide by the number of total seconds, and you don't actually get the real overall mpg, even if the computer is measuring fuel use accurately, because your measurement interval is a time, not a distance.
The other issue seems to be how the fuel use is measured. It is measured by the activity of the fuel injectors, and the manufacturer doesn't bother to make it give an accurate readout, because they want customers to believe they are getting real good fuel economy. They would rather err on the side of displaying high mpg than low. It is marketing.
On filling the tank completely, all that matters is that it is filled to the same level each time, and any errors in filling will balance out if you calculate over multiple tanks of fuel.
My observations with my Versa are that it consistently reads about 2 mpg higher than actual mpg.
My old Hyundai i20 was the most untrustworthy swine when it came to reporting the mpg with the on-board computer. Often it would claim around 75 mpg, but deliver 65 mpg - recently it promised 72, but delivered 55 at the pumps! Hyundai seem to be far more inaccurate than most, but even our Honda CR-V can be as much as 4 mpg wrong.
I would have to say, as most people don't work out brim to brim mileage, it must benefit all manufacturers to have over optimistic read-outs, as these are the figures most friends and colleagues will share, over inflating a brands economy - false advertising! I believe in America, official mpg figures are pretty much what cars will do. In Europe it is a joke - my old i20 was rated at 94 mpg highway, my almost exclusively highway based lifetime mileage was 64 mpg!
Awwww Geee! I like speculating that my Toyota is more at home with Saki!
I do like the consistency of my computer. It does appear to be dead on consistent. I need to find out if there is an adjustment to calibration.
On my 2008 Ford Escape Hybrid the trip computer is usually +2 MPG high. I just look at the display and subtract 2, and that's usually right where my calculations land after taking into account my odometer offset and doing the calculations manually.
Well I just fuelled up after 722.2 miles on that tank, and the computer read 64 MPG on the nose as I pulled into the forecourt. After refilling and adding the fuel up, fuelly said I got 63.2 MPG so the computer was VERY accurate given the length of time and distance between refuelling. I worked it out to be 1.2% optimistic, I don't think I've ever had a more accurate reading on the trip computer.