Is mapmyrun accurate on hilly courses?

Hills RARELY affect distance in a meaningful way.

Let's say you are either going up or down a 5% grade for the entire 10 mile run. This would be a pretty damn steep and hilly run. Over that run, you change .5 miles in elevation.

Using Pythagorum's theorem,

a^2+b^2=c^2

a=10, b=.5

100+.25=c^2

c=100.0125 miles.

So the elevation change added a total of just over one hundredth of a mile.

Even if it was a 10% grade all over, you are only looking at .05 extra miles.

Dunno, but it's easy to be inaccurate even on flat courses with it.

I mis-measured a 10k using Mapmyrun to be 10.3k, using the 'follow roads' feature. After some discussion I remeasured it on maximum zoom, carefully following the line a runner would actually take, and got 10.05 (i.e. 250m different), within 3m of what the official course measurer had recorded - and there's a steep hill in this one. There were a lot of 90' bends where mapmyrun takes the middle of the road, and a runner cuts corners.

As a rough rule of thumb, if a route does one lap with (e.g.) 4 90 degree bends, MMR may put you 5-7m out into the road on each and therefore give you a route 30-40m too long. If it has many bends it's worse.

Of course, doing it aaccurately takes a little longer...

If you are using it for mapping training runs then it is more than accurate enough.

If you are using it to measure race courses then it is not accurate enough, get one of the measuring nerds to use a Jones counter on it.