If you analyze a surface with a mix of negative and positive mean curvature, then the rapid change of the coloring may be misleading:
This is how it could look like if the range may also also show negative mean curvature:
I’m NOT saying that this would do a better job regarding what the help file says. It simply is showing the real mean curvature.
However, we know how to interpret the current implementation. But I know many users who simply do not understand the curvature analysis. I think this command needs some love…