I don't think pointing out Knuth is valid in this case. What happens with people way outside the curve most likely does not apply to everyone within the standard deviation (or maybe even a few times it) .

I am not saying that I agree or disagree with the conclusion, mind you.