And from China lol
Really though, +/-.001 with calipers is about where I'd draw the line...a good set, I check mine with gage blocks periodically on the flats and tips of the jaws. The tips are the easiest to get worn or bent. I also check for light through the closed jaws...if they fail they get for reference only written on them and I go to gage services for a new set. Someday they'll get me carbide like I ask for
If you look at the Mitutoyo caliper charts, they only guarantee an accuracy of +/-0.02mm (+/-.001") for their 0>150mm (0>6") digitals.
The dials are worse = +/-0.03mm (+/-0.001")
Obviously the display resolution is better, and they are (usually) better and IME using gauge blocks, a good set will read what the gauge block says (to the 0.01 resolution).
But you can't guarantee this, and on a couple of audits, I did have customers specifically state that you cannot measure a size with a gauge, that is smaller (more accurate) than the manufacturers guarantee.
I did argue you can for instance using the Mitutoyo digital mics as a fiducual - They have a resolution of 0.001mm, but an accuracy of +/-0.003mm.
If you use an inspection gauge block (at bottom limit size) and zero the mic on it, then you only use the mic to measure across the tolerance from 0 (bottom limit) to + whatever (top limit) i.e. a very small range (with the Inspection gauge blocks calibrated back to national standards).
Which they were happy to accept (for tight tolerance shaft work).
The trouble with a lot of this, is that it falls under the "good practice guide", which is rarely written down and comes with experience and/or part of the learning we all receive.
Such as the gauge rule of thumb which is it has to be 10x more accurate than the feature it is measuring....