Suunto Ambit 3 GPS Watch

Which App is the best one for windsurfing

Not so surprising really in this case of distance. Both devices use the Locational data (not 'Doppler') to measure the distance. With good reception for both devices they would probably not have too many errors and both would most likely have some kind of filter to discard major spikes and similar errors. The further you go, the better the accuracy as small discrepancies in location average/cancel out in the direction of travel. Both would probably overestimate the actual distance as sideways errors accumulate and add to total distance (and therefore add speed when using locational data path, which is one of the very good reasons we use the Doppler measurement for speed).

That said, one test, or even a number of examples, does not prove much. They both could be, and probably are 'inaccurate' (over actual distance). But it is good for our confidence to see that multiple devices regularly agree pretty closely in this parameter.

For sailing, we measure the 'path' traveled, which is why slight sideways errors accumulate and over read the distance (and speed). But if you measure point A to B with locational data, the error is much smaller. It is limited to the error for only the start and finish points. But if you take a non direct 'path' from A to B (as we almost always do) the A to B measurement does not reflect the actual distance covered (or actual speed).

It is possible to measure the traveled distance using Doppler calculations of heading and velocity, but small errors in the heading perpetuate, so the calculated location (using Doppler data) after a while, might be nowhere near the actual location.

This is why the Alpha 500 calculation uses Doppler data for the speed, but Locational data for the proximity circle.