This is true of GPS in general but most GPS devices have built in intelligence to "filter" out jumping around. For example, on a car GPS it will keep you on the road on the map and predicting where you should be due to speed and direction until you get far enough off the road that it finally accepts you are not there anymore. I have a GPS data logger I bought many years ago, before mobile phones were doing it or had terrible battery life when using GPS, runs off AAA batteries and lasts for hours. When looking at the tracks, same thing. So long as you are walking the track is pretty spot on, but once you stop, the longer you stay in the same spot you begin to accumulate a cloud of locations around you. Once on vacation I left it tracking while we stopped for lunch at a place (normally I would stop tracking then restart once we were on our way again) so I had about an hours worth of "GPS Cloud" on that track. Some of the locations were literally a block or two away from where we were.
You might try the same experiment. Turn on the GPS and keep moving then stop somewhere for 15 minutes then start moving again. I bet when you are moving it will be pretty much spot on while the 15 minute break you will get the "cloud".
The other might be early locations.
When I turn on the GPS, if the last position of my GPS is not where I am at, it takes a while to get readings from enough satellites to get a location fix.
If you have the camera OFF, when you turn it on, it could be that some of the readings are early calculations, with only a couple satellites.
A thought is the calculations itself.
If the camera GPS is using say 6 digits of precision, out of the 8 digits that the satellite generates, there is 2 digits of precision that is lost.
@PakkyT interesting thought.
I was just driving to the doc this morning, and the GPS wanted to take the parallel off ramp, but I decided to stay on the freeway.
The GPS did not recognize that I was still on the freeway, until the off ramp turned away from the freeway. It presumed that I was on the path that it calculated and instructed me to be on.
Right, this is why many GPS units meant for vehicles with have other modes such as pedestrian or "off-road" you can switch to when you know you won't be following the roads exactly, you can tell the GPS to actually follow the GPS signal rather than the mapping logic for a traveling car.
Is the X supporting a combined GPS/GLONASS/BeiDou location readout that I guess would mitigate such jump around issues?
Have too many threads going so just going to quickly sum up the GPS stuff.
To start with the camera uses the following GPS - Quazi-Zenith Satellite System (QZSS) and GLONASS.
All GPS devices have an accuracy built into them, like my Garmin GPSMAPS 64s which is +/- 50 feet. Garmin states that it is typically around +/- 30 feet which I have confirmed is about what I get with the unit. I tested this while in Grad School and getting my ArcGIS certification where we had access to very accurate GPS units. The EM1X seems to be in the neighborhood of around +/- 100 feet, maybe a little bit more (still need more data to really nail it down).
As for the difference between battery or location priority the best I can determine is all that does is change how often the camera updates the location. I really wish they would provide that information but I have noticed a difference in battery live using the battery priority and little difference in accuracy between the two settings. Biggest difference is fewer groups of shots taken at the same location but the spread is about the same +/- 100ish feet or so.