Rotational Eular maths

Alan.raceQs said:

and is beyond the capability of watch sensors.

Apologies, that was misleading. It's not the accuracy of the watch that's the problem. The problem, as I mentioned above,  is knowing the local value of G with sufficient accuracy and precision at your location. 

Despite Garmin's use of the term "milli G units" (not milliG), there's no such thing.

I suggest you do a simple experiment with the watch stationary on a table, using the returned x,y,z values (not Euler angles). Within a very short period of time, you will find yourself drifting metres away from your starting position.

The accelerometer is certainly useful for detecting gestures - it's great for differentiating between walking and running,  different swimming strokes, or waking or sleeping etc., but the cumulative errors from having to deduct the value of the G-vector in each dimension at each time delta will put you well off course very quickly.

Yeah, I did, that was how I observed the noise of about + - 15 milli G, and then calculated that to be less than a 5% error ?

I only need to measure in the vicinity of 4 seconds to 6 seconds, and the forces I expect are up to 4000 to 6000 milli GSs, so the errors is of significantly less effect given the size of the expected acceleration vector.

I hope you're wearing a g-suit. 4000 milli G for 4 secs takes  you to 350 mph.

Yeah, it's quite violent that is for certain, the G is more of an instantaneous nature, like over 1/2 or 1 second.  Just picture yourself flying a kite, that gets caught on an aircraft flying in the opposite direction at 20 kts.... and the yank you would feel... that's pretty much it.  The Landings can be 2 - 6G also, but it depends what part of the person your measuring as to the actual G. Feet get high G, body tends to be less as it crumbles