What about the possibility that the models so far have always been wrong and if they wrong in the wrong direction you would never hear about them?
replyPrecisely, it's just a selection effect. There's always uncertainty, and scientists are heavily incentivized to "prune" models that show large effect sizes. The result is the observed systematic underestimations, punctuated by (suspiciously monotonic) upward revisions any time the new data comes out.
reply