I’m curious whether it can be modified to find global maximum, but avoid McNamara fallacy
I guess the main issue will be the quality of input data