This Hedge Fund Left Decisions to a Computer – And Made 19%

This isn't super related to the article, but it reminds me of a chapter in Daniel Kahneman's Thinking Fast and Slow on a similar subject that I wanted to share.

In this chapter, Kahneman recalls a study performed by Paul Meehl, who discovered that statistical algorithms trumped expert intuitions roughly 60% of the time, with the other 40% of cases resulting in draws. These analysis were performed across a variety of different fields including "violations of parole, success in pilot training, and criminal recidivism."

Meehl suggests that experts form inaccurate predictions due to their overconfidence and unnecessarily complex analyzation. Another possible reason is that humans are generally inconsistent when summarizing complex information. An example being "radiologists who evaluate X-rays as normal or abnormal contradict themselves 20% of the time when they see the same picture on separate occasions."