Some background is explained in this post. Basically this is a rough estimate of the workload and calories burned during a hike.
I personally would trust the uphill estimates pretty well, but the downhill may be underestimated.
UPDATE #1: I think generally we all feel the downhill could be quite underestimated. The study & equation I used did the research on a treadmill, where friction on downhill is quite consistent and this is not the case in reality. For now, I am going to assume that one mile downhill (at any grade) is equivalent to walking a flat mile. Even though this isn't the best assumption (probably more work descending 20% grade than flat, but less at 10%), it's probably a decent guess until I get a lot of data collected.
Summitpost discussion / debate on the topic