Not in my experience. Convincing people to delete tests that only assert one invariant when the business changes its mind is easy. It’s the ones that have residual value after removing one invariant that trap people into spinning their wheels.