Bart, it may be less unlikely than you think. Those of using wide gamut displays are seeing the image at about 98% of ARGB(98). Even our best inkjet printers and paper/ink combinations exceed that gamut "somewhat" in some areas of the spectrum, so for people using those displays (more and more common) the predictability of the saturation impact is pretty darn good.
Mark, I see two issues with the suggested approach, besides the obvious requirement for a Wide-Gamut display. First, as you mentioned, the output medium can (more so as inks improve) exceed the gamut of 98% of Adobe RGB in the RGB axes. But secondly, even more likely is that the gamut will exceed that of an RGB display in the CMY axes, due to the fact that ink/paper is in a subtractive CMY (and more) colorspace.
The difficulty you mention with OOG indicators is true, but spreading the area doesn't fully resolve that issue because as you spread or recede, while the OOG area may spread and recede accordingly, the depth of the "OOG-ness" is still not completely clarified. So those are the reasons why for this particular adjustment I prefer to trust my eyes.
The 'eyes-based' advice is more likely to fail, exactly on those very saturated colors that we'd want to keep in check (Cyan-blues are a typical case). Luckily, and that's why one can get away with it, many actual scenes do not extend to (all of) the boundaries of the gamut of the paper/ink gamut, or only small parts of the image do.
That's why I use a Photoshop Difference layer approach for all those 'harder cases', by taking the absolute difference between the original image, and the converted to output profile image. That difference is a
proportional signal for 'OOG-ness', and it can be used as a mask to locally reduce saturation (or twist Hue, or reduce Lightness) to better fit that of the paper/ink gamut, but only where it matters(!) while leaving the rest of the image untouched. All can be achieved in a single action, with one mouse click.
Cheers,
Bart