It’s interesting that in your example, the pps of y to x was zero when visually there’s clearly…

It’s interesting that in your example, the PPS of y to x was zero when visually there’s clearly some predictive power. I can’t guess the sign but I can guess te magnitude very accurately.

It’s easy to see that y can predict very well abs(x) which is a very simple transformation. I was a little disappointed when I found this. Although I recognize that PPS is successful in

finding relationships that correlation alone can’t. Just food for thought.