2001-03 Prius, EPA Composite, Standard Deviation

About a month ago I was challenged about my driving and there was the recent Australian driving 'race' based upon beating their local government mileage standard. These events got me thinking about a problem of terms whose definition is tied to different government metrics and does not reflect local driving conditions.

When I bought our 2003 Prius, the EPA composite MPG was 48 MPG. In 2008, new standards were announced and today is 41 MPG. But that is not the only milage standard for the 2001-03 Prius, NHW11 model:

• 41 MPG - EPA since September 2008
• 48 MPG - EPA before September 2008
• 68 MPG - Japan (reported via New Zealand)
• 48 MPG - UK (adjusted for USA gallons)

The complaint about the original EPA number was it inflated hybrid performance. Many complained that hybrid electrics could not achieve the posted EPA numbers. That same complaint is doubly so for the Japanese and UK mileage numbers. So the EPA added additional tests and in the case of the NHW11, the new number is 41 MPG:

Yet we find the 68 user reports, 2001-03, from the EPA web site shows an average of 45.2 and a standard deviation of 3.99 MPG. To understand vehicle performance we need something that reflects actual performance.

A normal distribution will have the following percentages:

• 15.85% - below one standard deviation from normal
• 68.27% - within + or - of one standard deviation from normal
• 15.85% - beyond one standard deviation above normal

So using the average and standard deviations for the NHW11:

• ~15% < 41.2 MPG - beyond one standard deviation below average
• ~50% < 45.2 MPG - below the average
• ~85% < 49.2 MPG - under the average plus one standard deviation

Notice the historical EPA standard, 48 MPG, is fairly close to the user reported MPG averages plus one standard deviation, 49.2 MPG. The term "hypermiler" has historically been defined as relative to the USA EPA standard. Before September 2008, about 1/5 th of all NHW11 drivers qualified as "hypermilers." After the change, 4/5 ths of all NHW11 driver are "hypermilers." If we used the Japanese standard, the number is even smaller, <1% of all drivers would be "hypermilers." The term "hypermiler" suffers from the absence of a universal definition and an arbitrary change in the testing standard.

A statistical definition has the advantage of being freed from government policy changes and works in Europe and Japan. In fact, it works better in Europe and Japan because it includes local driving practices and road conditions:

• Japan - high ratio of public transportation, moderate highway speeds
• Europe - high ratio of public transportation, highest highway speeds
• USA - low ratio of public transportation, high highway speeds

Folks are free to use any term they want any way they want. But math and engineering tends to precision. Terms based upon a statistical definition have the advantage of precision and avoids the problem of 'moving the goal posts.'

Bob Wilson

 

I don't go there often enough so I'm willing to take any changes. The UK number identified it as a 2003 Prius but the USA and UK model years sometimes don't quite agree.

I found this link: What MPG - Toyota MPG Comparison Figures - Fuel Economy and technical data on new and used toyota models
I believe their 57.6 imperial MPG -> 47.96 (US) Thanks.

I'll fix the charts.
​

My point is any of the government mileage numbers are at best going to be based upon some collection of lab tests. In the case of the NHW11, they went from +3 MPG over the average to -4 MPG in one day.

Mostly I'm interested in addressing "user expectations," what folks can reasonably expect to get. We've had a lot of fun playing with the fuel economy but I'm more interested in understanding the baseline and then making a reasonable improvement without losing functionality. After all, I bought the car to get somewhere. Efficiency is important but secondary.

Bob Wilson