How is that not disputed by those graphs? While there is distortion at both ends (particularly the lower end), it nevertheless shows self-assessment is positively correlated with ability.