I've thought about that, as I always had one spot right at the crest of my 2.5% hill that could be reverse-elevated and it was behind a low hill - so you really couldn't see it. Never did it though. My gut tells me that would work, but would work against you on the downhill slack run-in from a sudden stop. But if you had a dedicated up track and down track, it 'should' work. Just theory though. I do know that I really checked the curve geometry on mine with a tiny string level and a flexible straightedge to make sure there were no dips, humps, or anything else going on.
Another thing we haven't gotten into is coupler pull-aparts under maximum load (uphill, curves, grades). This was a nightmare for me until I read a Jim Fitzgerald Ntrak article on trimming MT coupler heads to what is now known as the Reverse Draft Angle (RDA) modification that has become manufacturing standard on MT couplers. But if you have old MT couplers (body mount or truck mount) and you're either popping apart or pushing one up or down to the point the trip pins snag, this is the solution, and it will drive you nuts on heavy trains way before the stringlining issues ever get you. There was absolutely no way to drag 20+ 89' flats uphill without a vertical uncouple until I did that. Now I'd break a coupler completely off before it would do that.
I often drop my eyes down to track level and study the couplers as they go by; I want to see perfect vertical centering under load. Because if it isn't there, it not only will uncouple, but the trip pins will snag somewhere. I still think that's one of the major reasons people cut the pins off, and it's unnecessary if you fix the head problem. You may still do that, but don't do it because they mysteriously snag.
