Of course not, Duane. Never have, never will use then for Start, Finish, or Turnarounds, except for designating the general area for these points.
What I mean to say is that when the nearest fixed object for a mile mark is many yards away, GPS coordinates in an area without tall buildings, heavy overhead canopy, not or on or beneath bridges are not only a huge help, they are by themselves accurate enough for these non-certified location descriptions. I have noticed that GPS locations seem to be less inaccurate these days than say, ten or even five years ago in my experience. Part of this is that GPS software is improving, IMO.
I performed course layout two days ago for a 20K that I certified a few years ago (map is attached). The parks authority allows no paint, nails, or markings of any kind. There is little on the paved bike path part of the course that can be appropriated for fixed object reference except small mileposts on the side of the trail. As I searched for some of these posts, I concluded that two of them had gone missing in the last year. I was left with GPS coordinates and little else with which to locate a couple of the mile marks. I heard no complaints about any of the mile marks following the race.
Doubtless you know that it is often possible to locate a Start, Finish, or Turnaround on Google Earth when you have your exact ground-truthing data for these locations. It is then possible to locate an exact point on GE, then measure from it to a mile mark. This entails locating the exact fixed-object-identified point, then measuring ("Add Path" tool) at highest zoom for curves, and at lower zoom to plot a perfectly straight measuring line where feasible. Set a GE placemark at the interim timing point so located, then use the ruler to mark a small line perpendicular to the running path at this location. Then, when GE has ground level imagery available for the road in question, is is possible to locate fixed objects at ground level. It is then feasible, again using the GE ruler, to identify this object and measure accurately from it to the mile mark. IMO, this method can be perfectly fine for Mom & Pop 5Ks or any race for which +/- 10' is adequate.
I used this for a recent low-key race measurement in which I hurried through my second ride to complete it before a lightning storm arrived. I filled in the three mile marks using this method in an open area with good ground-level imagery, using my fixed-object-identified Start, Finish, and Turnaround. No complaints after the race.
I am interested to know if anyone else has used this method, or if anyone has any reservations about it other than those previously mentioned.