I spent a lot of time last year exploring the curve easement topic, consulting a turn-of-the-prior-century British civil engineering paper on the matter and far more recent work by some Greek civil engineering professors, as well as the empirical procedure outlined by the late John Armstrong. The result was a spreadsheet that calculates the x-y coordinates of points on an eased curve, as well as templates for curve entry and exit, with emphasis on situations where sectionalized track is used. I hope to post the write-up before summer. As for superelevation, my layout-under-construction uses N gauge cork roadbed strips at the outer edge of curves, between easement points and tapered down within such easements by careful application of a belt sander. The area between the edge of the N half-roadbed and the O inner track is tapered using lightweight spackling compound, to provide a uniform surface for the foam O gauge roadbed atop. The 1/8" superelevation not only looks good, as Walter notes above, but is prototypical, from what I've read.
Yep, I found that the vertical easements are probably more important for reliable running of large locomotives. I had to do a little "tuning" to make sure that certain locomotives made a smooth transition onto and off of the grade. The horizontal easements aren't always critical for reliability as long as curves are wide enough, they do enhance the appearance running through the curves. However, the vertical easements will really bite you if you don't consider them. Oddly enough, some of the significant offenders for vertical easement issues are things like the Lionel scale GG-1 and the steam with lots of fixed drivers like the UP 9000 4-12-2. I found most of my articulated stuff much more forgiving on the grades.
