Alright...did a small control survey yesterday and observed the short lines as was suggested. The results of two of the lines, in U.S. Feet...gps(converted to ground) 426.66 Measured by the trusty Geodimeter, 426.65. Vertical differed by 0.03'. On a longer line GPS (ground) was 1045.29' Geodimeter was 1045.29' with a vertical difference of 0.04'. I can live with these kind of results. 0.04' in 1000'...not too bad. This was in a nice little tight network with at least 2 observations on each point...so maybe that answers my question of the first post...Before, with using the 2 base, 1 rover method, I tended to see a bit more than 0.01' between points.

Stumbling in the dark, Tom Bryant PLS ST Louis MO

You confirmed what most of the conversation was saying, in that the short lines are most likely to be occupied by conventional means when it's time to topo. Observing them directly 'tightens' up the line..
I would be surprised if either method showed much difference, but I always like to measure the short lines just because their the most likely candidates to be checked by other surveyors with conventional means...
Doesn't sound like your stumbling to much...
TM