Many are picking up on NYT article about unreliable GPS receivers:
http://www.nytimes.com/2011/12/20/he...ccurately.html

The AMC is not surprised:
http://equipped.outdoors.org/2011/12...gps-watch.html

Ray Maker strongly disagrees:
http://www.dcrainmaker.com/2011/12/w...g-article.html

The author of the article, Ms Kolata, does not say where the route was, the density of trees, the surrounding high rises of NYC, or weather conditions. These can cause major distraction to the GPS signal.

I use the Edge 705 on my bicycle. The altimeter is more random than the stock market. But the distance has very small error. For example, here are a few rides I charted and recorded:

122.28km GPS vs 123.7km map (98.9%)
76.6km GPS vs 76.8km map (99.7%)
113.2km GPS vs 111.5km map (101.5%)
36.8km GPS vs 36.9km map (99.7%)

This works for me.