out of curiosity I took my forerunner 220 to the track for the first time today and I didn't expect any positive results, oddly enough it was spot on every 400m (counter clockwise, watch on the left arm + running close to the outer marking, so that my left arm would swing in the middle).
No, this is typical. I have read that other dude's accuracy and precision comparisons, and something does not pass the sniff test. I run the exact same route 3x week with two other guys who have different GPS watches. Over hundreds of runs, my watches have had a standard deviation of <0.02 miles on the 5 mile run. Their watches are similarly precise. That is pretty freaking excellent, and does not reconcile with fellrnr's results. For example, he has the iPhone ranked pretty high relative to other GPS watches, but in my everyday use, they are crap. When I have done the same statistical comparisons with an iPhone, the SD was a little more than 0.1 miles. And further contrast to fellrnr's results, I have observed GPS watches getting more precise as newer versions are released. My 735XT is definitely more precise than the 910XT it replaced.
So, I suspect something in his protocol may amplify errors or issues that do not occur in real life use.
The other mystery in his testing and analysis (which he does not address) is the "calibrated" pods get their calibration from GPS. So, to argue that a pod is highly accurate while GPS is not is a bit of a conflict.
Something is off that needs a credible, statistics-based defense.