Extending past a normal thickness nut a bit is required for maximum strength
I would need a white board to explain it best
You can do experiments in a tension puller to prove it
But it’s remarkably calculatable detail
There is a inverted cone of unsupported material along the shear surface in the center of the bolt face at the top of the bolt when it’s flush
This lack of material means the top most threads in the nut have less shear material so they are engaged in threads but they are not able to fully benefit from the bolt bulk which will deform and pull away in a cone shape not a flat surface shape
Flats happen from stress cracks and defects
Pulling metal samples to failure use to be a standard engineering class in like 2nd year or so and was very non intuitive
Half the point of the lab experiments is to drive home how something so simple has phds and years of research behind it and is not intuitive at all
The material failure modes are applicable to many other situations not only bolts and helps people think more critically and ask the right questions - what don’t I know about this and maybe it’s not at all the way I think it is.
-s