Your accurate measurements with Reach are made between the base and rover. First you have to supply Reach with an accurate base location yourself.
The best way is to start at a place of known accurate coordinates. For example, a survey marker with published coordinates.
Otherwise you could use corrections from a 3rd party. For example an NTRIP connection or by downloading historical CORS data. Hopefully there is a station within 10km of your location.
The next best is to log data for a long time and submit it to a PPP service.
The last resort is to average base coordinates, like you did above. Let us assume that you have no other option than to use average base coordinates. So:
- You average base coordinates for, say 30min to get a good average of single-mode coordinates. Now you record that average coordinate permanently. You have just declared that the spot your base station sits on is the exact coordinate you recorded. You never do an average on that spot again. You just use that same coordinate for your base every time you place your base on that spot.
Now, in the future, maybe you get access to a better method of achieving a base coordinate (like a NTRIP correction stream from a station less than 10km away). Well, you can take your base back to that spot and use NTRIP to get a new, more accurate coordinate. You update your old coordinate, and update all your previously surveyed points by shifting them by the difference between the old and new coordinate.
To make a long story short, your only mistake was averaging a coordinate two times. You should only do it once, because you’ll never get the same result twice! Single-mode is not that accurate!