I can show you a graph of exactly what it will look like, and there is no "fallaway at the nut end." I can think of only two ways your statement could be true: (a) if we were to define "nut end" as frets 1-19, which is not a sensible thing to say; or (b) if we lived in a world without adjustable bridges & saddles, which we don't. To sum up, sanding frets level from 1-19 with a 0.003" slant down to fret 1 is not going to create "fallaway" at the nut end. Bear in mind how crooked and humped a typical buckled neck is - imagine we simplify it to a 2D profile and draw the whole thing on a graph. Now imagine we pick a straight line that removes the crooks & the humping from this profile, whether we remove the minimum of material or some amount of material more than the minimum. Depending on the exact nature of what is being removed, if we keep the same graph as the frame of reference, the resulting straight line will inevitably have a slant when compared to the horizontal axis. You are never going to be able to draw an "unslanted" line for where your leveling beam will wind up; it will always be constrained by the profile it encounters and how this profile is progressively altered until you get to "done." To put it another way, let's say I didn't introduce the tilt at fret 20: I'd now include these frets in the progressively lowering straight line introduced by the leveling beam as I did the work. The resulting line would not be that far off from the line that you say creates "fallaway"; it too would have a "slant" of some sort if you drew the whole thing on a graph. I actually do such graphs as part of my planning; I have a whole system worked out. It's extremely simplified - I follow only 3 lines down the neck, and each line is 2D, so I don't have a 3D plane; but I find it helpful. EDIT: Re-reading what you wrote, you're not actually saying anything different than I am. The way you describe it, the 3 low frets at the end of the tongue wouldn't get leveled because they'd be shorter; the way I'm doing it, I guarantee that those frets won't get accidentally touched. But whether we do it my way or your way, the result would look the same if we graphed it - you'd have just as much of a "slant" on frets 1-19 as I would. So I think the only disagreement is that we think and work differently and plan for different contingencies; it's not that the work would end up being much different.