Should I give up on my 830?

The satellites don't send signals to gps receivers.  They broadcast the signals and appropriate receivers can pick them up.  Nothing goes the other way.

Doesn't look like gps to fit is available yet for 830. Creating a data screen that has gps accuracy and signal strength fields and i'll try and watch it some tomorrow in the trouble spots and see what I see. I'm also using a speed sensor which in the past has been great at smoothing giant speed fluctuations in dense woods, surely that can't have anything to do with poor gps track performance 

I set my 830 to GPS+GLONASS. Unpaired my phone. No navigation. Did a 40k ride and it worked without a problem. Tomorrow I'll try navigation.

No one is stopping you to go and help ESA and GSA fix the satellites. I understand that Garmin has their fair share of software bugs. But you are barking up a wrong tree here. At this rate, people are gonna complain that they want a chemtrails detector widget from garmin.

I thought GALILEO is still not yet fully operational. This is expected in 2020 right? You can already use it but I think it can be expected there are still issues or some outages?

Most tracks I have seen going completely wrong are reported by users using GPS+GALILEO. I am not using GALILEO yet and I have never experienced such errored track.

I'm definitely getting an offset track using GPS+GLONASS 

I'm definitely getting an offset track using GPS+GLONASS

With any GNSS system the tracks you get aren't always perfectly correct even when the receiving device has no issues.  There are always errors which can be caused by many things.  The errors when you have a good view of the sky will usually be less than the standard system error.  That standard system error is likely big enough to show you off to the side of a track on the ground though.

If you want higher accuracy GPS you need access to military receivers.  If you want higher accuracy Galileo you will need special receivers and need pay for the access.  The free to public is good enough for bike riding even if the tracks are a bit off at times.

Wongataa said:

If you want higher accuracy GPS you need access to military receivers.  If you want higher accuracy Galileo you will need special receivers and need pay for the access.  The free to public is good enough for bike riding even if the tracks are a bit off at times.

At the risk of going slightly off-track here, with the US GPS system, there is no difference between the service provided to civilian or military uses. The only difference would be that military applications may choose to use dual-frequency receivers which are more accurate, but also much more expensive. They are available for civilian applications too, but are generally cost-prohibitive. You may be thinking of selective availability, but this was discontinued in 2000 and since this time civilian applications can achieve full accuracy without restriction.

GPS accuracy will always depend on quality of signal, which will vary with terrain, buildings and atmospheric conditions, but the committed signal accuracy (measured as transmitted from the satellites) is easily enough for the levels of accuracy needed to tell you that all of your friends are faster on a given segment on Strava ;-)

This statement gives some numbers to this:

The accuracy commitments do not apply to GPS devices, but rather to the signals transmitted in space. For example, the government commits to broadcasting the GPS signal in space with a global average user range error (URE) of ≤7.8 m (25.6 ft.), with 95% probability. Actual performance exceeds the specification. On May 11, 2016, the global average URE was ≤0.715 m (2.3 ft.), 95% of the time.

Broadly speaking, the public (FOC) Galileo service should offer similar, if not slightly better levels of accuracy (when it's working, anyway!), but also offers a higher-precision, 1cm-level accuracy on an encrypted stream for paying customers. I'm sure Garmin are using the public stream and it's more than good enough for this use case.

In short, both services are good enough for what is needed here and the accuracy issues reported here are likely to be due to signal limits, receiver firmware bugs or other issues specific to the Edge 830.