Originally Posted by ericgu
You're reading the graph wrong. The grade percentage is the really jagged line, and what you're seeing is the noise that is inherent in GPS altitude measurements.
I think you are right. If you eyeball the mean of the noisy line, it's about 5%.
As for the calculations it seems to be a question of using the same units to get the right number - as Machka has done. i.e. feet/feet or miles/miles (X100, to be pedantic).
It should have been obvious that 39% was impossible; I used to ride a couple of 1 in 4s (25%) when I was growing up in England and it felt like my nose was almost touching the road!! I did zig-zag!