In his simulation with random measurements, the so-called Dunning-Kruger effect actually becomes more visible as the measurement error increases. “We have no instance in the history of scientific discovery,” he continued, “where a finding improves by increasing measurement error. None.”