measuring milleage with online maps

[Ajax Bay]

That is still more accurate the GPS. GPS data is just a series of points. . . .
That's assuming the original GPS was accurate. Garmin only claims an accuracy of 3-5 metres, so there's a lot room for error there.

GPS is 'just' a series of points, but points calculated at very frequent intervals so the distance travelled between each point is, I hypothesise, ~quite~ accurate. The claimed accuracy of 3-5m you quote is for point accuracy and will be effectively 'cancelled out' in a distance computation, given the error will not jump from 3m one way to 3m the other way (laterally cf the road/your track).

 

[Dogtrousers]

If you're calculating from mapping software GPS error doesn't come into play. You still just have a series of points, though, which may not trace a line that is quite as wiggly as the real path, and is subject to the accuracy of the map.

But really, for any real purpose, all methods are good enough. We're not measuring tracks for world record attempts.

 

[andym]

If you're calculating from mapping software GPS error doesn't come into play. You still just have a series of points, though, which may not trace a line that is quite as wiggly as the real path, and is subject to the accuracy of the map.

But really, for any real purpose, all methods are good enough. We're not measuring tracks for world record attempts.

All true.

Depending on the amount of route points the software uses there is going to be some degree of rounding down error, which might be quite a lot or it might be very little. I'm guessing that with very clever maths that error might be pretty small.

And yes it really isn't a life-and-death question but if you want a fast, convenient and pretty accurate measure then a cycle computer is as good as it gets. There's a reason you can still see geezers trundling around those wheeled measuring jobbies.

A number of issues with this, firstly my cycle computer is set by the wheel circumference and I only have an option of exact CMs (211,212,213,214 etc) so if the true circumference was 2135 mm and i use 214 then that is a quarter of a mile out every 100 then quoted tyre circumferences are not that accurate

Doing the calculations for miles is way too complicated for me, but to take your hypothetical case, you've got an error of a bit under a quarter of one percent. Not a big deal, and good luck finding any other measure of the distance that is that accurate. And if you are concerned that the quoted circumference is not accurate enough well you can always get a piece of string and measure it.

 

[Dogtrousers]

I use a mapping website that allows you to resample a route using fewer points (for the benefit of users of GPS units limited to 500 points). Just for a laugh I took a 7,000 point 100 mile track and gradually reduced the number of points. At 700 points, a tenth of the resolution it was still 99.8% if the original length. Only when it got down to below 100 points did it seriously drop off (50 points still 95% of the original)

Which just goes to show ... Er... What was the question again?

 

[andym]

I use a mapping website that allows you to resample a route using fewer points (for the benefit of users of GPS units limited to 500 points). Just for a laugh I took a 7,000 point 100 mile track and gradually reduced the number of points. At 700 points, a tenth of the resolution it was still 99.8% if the original length. Only when it got down to below 100 points did it seriously drop off (50 points still 95% of the original)

It depends on the road - try doing that on a mountain road that climbs through a series of hairpins and you'd pretty soon start seeing some significant differences.

 

[Tin Pot]

If you're calculating from mapping software GPS error doesn't come into play. You still just have a series of points, though, which may not trace a line that is quite as wiggly as the real path, and is subject to the accuracy of the map.

But really, for any real purpose, all methods are good enough. We're not measuring tracks for world record attempts.

Speak for yourself, I take my time trials very seriously. <sniff>