Hello all. I don't know where but I think that in one thread somewhere there was a short discussion about whether the movement of the arm have an impact or not on the accuracy since gpspoints is based on timeintervals. This means that sometimes the arm will be in front of and sometimes behind the body when the position is registered. This might be even out after a couple of km training, I don't really know, I just remember once being told that gpstrackers placed on the upper arm have better accuracy than those inside the watch on the wrist.
However, my suggestion is that if there is an accelerometer in the watch it should theoretically be possible to use the movement of the arm instead of time to decide when to measure position. Could this lead to better accuracy since the arm should be in the same position relatively to the body each time? I realize that there is some things that could cause trouble like how accurate the accelerometer is-the time between armpendlings have to be very precise measured and the accelerometer then have to be very accurate on this.
As a bonus the watch will also realize that when the arm drops and accelerometer stops giving signals the person is most likely standing still (I might be running with stiff hanging arms though). As it is now readings says that when I'm stopping between two intervals the watch assumes that I've been walking up to 20-30 meters due to gpsinaccuracy.
Have a good day!