That is exactly how Fuelly is calculating MPG's. If you are referring to odometer error, since it should be a constant, averaging averages will get you the same result as the harmonic mean.
Well, no. The correct average for various mpg readings would be a weighted average. An average of say 60mpg over 600 miles carries more weight than an average of, say, 80mpg over 60 miles, simply averaging the two mpg's does not give a meaningful mpg over the 660 miles. The right way to do it is to total the fuel used over the 660 miles and divide by the total distance. Simple, and it eliminates most of the random variation in distance measurements (assuming one takes the readings at the beginning and end of the run and does not total the intermediate tripmeter readings). Hopefully this is how Fuelly calculates? I don't use it. Again, there are two types of error: one is systematic errors (such as odometer or gas flow readings being consistently off in one direction by a certain percentage) and random (or statistical) errors (essentially the precision of the measurement method, vary equally in either direction and tend to average out over multiple measurements, can be less significant percentage wise when a measurement is made over a greater interval). In science, one looks for a method which will give the best accuracy possible given time and cost constraints. One then gauges its accuracy by a series of tests to measure statistical error, and develop correction factors for systematic errors (if any). Once statistical and systematic errors have been pruned away, hopefully one can see more clearly the effects of various other factors. Essentially this is what some people (me included) are trying to do here. Records of "tank mpg's" are useful because they help reveal seasonal variations in mpg or changes due to different driving patterns. But one should not average tank mpg's.
I keep a spreadsheet log, one line for each fillup volume and distance travelled, and also set up a cell to calculate accumulating liter/100 km based on the total volumes and distances since inception of my record keeping. There's maybe 1500 km after purchase I didn't track, but other'n that... After each entry I replicate on Fuelly.
Hello, Newbie here and I know this is an old thread but I just wanted to have some clarification from anyone who might want to chime in. I've been noticing an odometer discrepancy with the mile markers on the interstate in my 2016 Prius and from what I gather in this thread, I shouldn't be too concerned about it and shouldn't run off to the dealer. My odometer under reports a mile in about every 72 miles of actual miles... I guess this is similar to what others are reporting here. It seem to be very consistent.
Unless there is a sign stating "Measured Mile", or something similar, the mile markers on the interstate system are not all that accurate. I would recommend using a GPS device. That will be far more accurate.
That depends on the road, I've found some extremely good sections. Many of the markers are set by the original highway survey. While some individual posts may be offset for mounting convenience, their individual errors to not accumulate. During periods of travel boredom, watching seconds between posts or interpolating between tenths digits on old analog odometers, I've been able to spot which posts were significantly offset. Culverts and bridges were common culprits, while some signs merely re-used the nearest other handy post. Do you notice special mileposts with two figures, e.g. "123.45 Ahead, 124.09 Behind'? These are survey 'busts', commonly caused by new construction rerouting of old roads. I have seen many of these, though not all such milepost discontinuities are prominently marked. I-90 in Washington State has an unmarked bust near MP220 (about 1/2 mile missing), but is otherwise accurately marked from the Columbia River to the Idaho border. Just make sure not to use a hiking GPS in 'Energy Saver' mode. This causes mine to jump around and roll up excess trip distance, especially in narrow valleys with obstructed satellite view. In normal mode with wide-open sky, it and my automotive GPS read within 0.1% of the above mentioned section of I-90 mileposts. A number of other highway sections seem similarly accurately signed.
Just going from brand new tires to down-to-the wear-bars can throw you off, assuming you were "on" to begin with.
Anyone with a three or four touring that can provide real vs computer mileage? I'm curious about the 17" touring tires, and the effect on mileage error... thanks.
I did a thread way back in March and it covered this issue. Also there is another master thread somewhere. Bottom line the GPS validated the odometer that we were not getting cheated on a long trip. It was a little in our favor. Did a quick check of fill ups and those pumps are all over the place (97% are pump errors for the MPG difference) No way to do this accurately without a measured gallon like they use to do in consumer reports MPG tests. Buying only a gallon is problematic on getting an exact gallon from a pump. Bottom line is I am happy and I don't care to ever do these tests again. Close enough for me.
Everyone is talking about 1-5% error... My recent trip from Houston to Austin according to google maps was around 150 miles, but at the end of the trip I had more than 250+ on my odometer, so it was off for more than 100 miles on 150 miles trip... what is happening and where do I go with this issue?
That almost sounds like GOOGLE Map readings is in kilometers, the other in miles or furlongs or light years. Mine is so close that it's almost impossible to detect. I checked it against a measured 5 kilometer section of the motorway. Distances according to my GPS are similar too.
Good catch, but if Google Maps is saying 150, and @Maryna's odo is reading 250, it sounds more like Google Maps is the miles and the odo is in kilometers. But presumably a US car, that's a puzzler. Or maybe a typo in Maryna's post: the numbers are flipped??
Yeah - I think I flipped my thoughts - but then - I'm on the bottom side of our planet and easily reverse things. Particularly on a Friday.
I've always wondered: if it's morning and the sun is to your right (you're facing north), it will commence traversing the sky in front of you? That would throw me right off.
That definitely has thrown me off the several times I've been south of equator. Even an evening crescent moon points the wrong way!
I believe that cyclones or hurricanes can't even cope with crossing the equator. I'm not sure if we solved the ODOMETER ERROR though.
