Discussion of 'Predictive qualities of tests of control for the error rate' by Ed Broeze