Experts suck

Simple linear models outperform experts in making predictions. Notice that the examples are mostly from the social sciences; it seems they have yet to learn a lesson that hard sciences learned over a hundred years ago. Potential root causes include misguided ethical ideals (a cultural aversion to discrimination by race/gender/income/etc), cognitive biases of the experts, and a fear of irrelevance on the part of the experts. Trust the numbers; science!

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.