>Huge data sets today render most statistical tests useless since no hypothesis is strictly true given enough precision, and the focus on eking the most out of

>Huge data sets today render most statistical tests useless since no hypothesis is strictly true given enough precision, and the focus on eking the most out of every data point is overkill. Mind, huge data sets and huge amounts of computing power also make the use of Bayesian methods feasible, which don't have the NHST problem that anything is significant with enough data-points.