In today’s world of big data, I would say The James-Stein estimator is one of the most important concepts (although it was 1950s). It was a shock to the world

In today’s world of big data, I would say The James-Stein estimator is one of the most important concepts (although it was 1950s). It was a shock to the world when it was shown that least squares methods (a method used since Gauss) are actually not optimal in most cases. https://en.m.wikipedia.org/wiki/James–Stein_estimator https://en.m.wikipedia.org/wiki/James–Stein_estimator