My thought is that, if the oars are a good length for the beam and height of the oarlocks above the water (John Harris has a formula accounting for both in his piece on rowing), the grips should overlap about 2" with the oars level.  That 2" overlap is what the simpler Shaw and Tenney formula (based on beam only, but works for most ordinary rowing boats) is meant to produce.

Pete Culler didn't like to have "collars" on his oars.  His idea was that he could shift the oars in or out to vary the leverage based on the conditions.  For example, in a lightly loaded boat going fast over flat water in calm conditions, he'd shift the oars outboard some to get the effect of putting a car into a higher gear--blades move further for the same length of stroke.  He'd move the oars further inboard in the reverse situation, rowing a heavily loaded boat, maybe into a stiff breeze, to have the effect of downshifting a car into a lower gear to go up a steep grade, the hands overlapping more on the recovery.

Most of us have trouble controlling the oars that well, however.  Takes a lot of practice to have that become natural, so most of us less serious rowers will use some sort of collar positioned for "average" conditions...hopefully.  Rowing with the fraction of a horsepower which we humans can produce is a somewhat marginal thing, and it doesn't take much being off to make the difference between feeling like you are gliding along versus trying to row out of a puddle of molasses.

.....Michael

   Thx mike!! I'll look into johns formula