My surveying text (“Surveying” by Charles B. Breed) says “competent tapemen can usually obtain precisions of 1/5000 or less…”
In the layout of two calibration courses associated with the measurement of the 1996 Olympic Marathon in Atlanta, eleven teams of two experienced measurers measured the lengths.
Raw measurements of the two courses were:
West course
Minimum – 480.205 m
Maximum – 480.440 m
Range – 0.23 m
1/5000 = 0.09
East course
Minimum – 480.87 m
Maximum - 481.09 m
Range – 0.22 m
1/5000 – 0.09
In my own experience I am happy to get within 2 or 3 cm for two measurements. All it takes to produce a big difference is for the sun to come out on one measurement and be overcast on the second.
The 1/5000 level of precision would suggest that 10 cm in 300 m or 4 inches in 1000 feet is not unreasonable. Most of the time this will be achieved.