Defining expected failure states is good, because it allows you to distinguish expected failures from unexpected ones. What it doesn’t allow you to do is re-cas

Defining expected failure states is good, because it allows you to distinguish expected failures from unexpected ones. What it doesn’t allow you to do is re-cast failures as successes, which is the spin that the author identifies.