Distinguishing cause from effect using observational data (arxiv.org)

I'm very surprised that more people haven't picked up on just how revolutionary this is. Coupled with well defined experiments this can radically change science.