I did a little experiment.
I put a cartridge case which measured about 0.748" long into the chamber and pushed the bolt onto it, and the extractor just barely grabbed the rim (the bolt was in contact with the barrel).
By "just barely", I mean it didn't immediately go over the rim; rather, there was a slight delay - apparently it was right on the edge and therefore unstable - and about half a second after the bolt made contact with the barrel, I heard the extractor slip over the rim.
Then, to simulate a "worst case", i.e., a cartride case that's a little shorter than 0.748", I took a piece of printer paper (with a measured thickness of about 0.0035"), and I cut a round hole in it just big enough so that the cartridge case could fit through it.
When I tried to close the bolt with the paper in between the bolt and the barrel faces, the extractor did not fully grab the rim.
This doesn't seem like a desirable situation to me.
I can envision tolerances stacking up the wrong way and adversely affecting the reliability of the gun. Maybe the chamber is cut is a little on the deep side, and the bolt face is recessed to the max, the case length is toward the short end of the tolerance, and firing pin is a little short (or the relevant bolt dimension a little long), etc., resulting in an occasional FTF or FTE; especially when the gun gets dirty or something.