I didn't want to get into this one but since its getting nutty I thought I'd give you something to sink your teeth into. First, yes the G is the largest plain, which is important because it also makes it the least forgiving to tension changes. That saddle is usually the farthest back (besides the low E) to compensate for the pull sharp when fretting. Slight changes in tune are most noticeable on that string. Evidence for that would be to give each tuner exactly one half turn sharp, and see which string has the greatest pitch travel.
Second, and perhaps more important, is that the G seems to go out of tune more often because of it's place in the tuning scheme. Between the G and B is only a 4 step interval while the rest are 5. I don't know quite how to say it, but there's a weird relationship going on right there. I've found that if the G isn't tuned properly, it becomes the main offender in the chords that are out of tune. The G and B relationship is usually the cause, and I find that they usually sound best together when the E through G set is pushed a little sharp progressively, then the B is a little flat against the G. The high E is in tune with the B. This sounds the best to me, and gets all my chords in tune.
The compensated nut idea (having the contact points at the nut loosely resemble the stagger at the saddles) is brilliant, too, because most of the time, the G sounds sharp when fretting it at the first couple frets because you are pulling it sharp. On a jumbo fret guitar (this IS the jemsite
) you can squeeze the G sharp very easily. I think the compensated nut does 9/10ths of what the Feiten tuning system does.
So to sum it up, part of it is because it does "knock" more out of tune, all things being equal, and part of it is because the guitar is a crazy beast. I'm not ready to concede to sun and moon or what kind of underpants you have on or whatever