Nice piece about extreme multiverse analysis! The exemplar study performed 3.6 billion regression models. The author argues against this approach and suggests that multiverse analysis should only include valid models selected based on theories. Statistical inference requires thinking before acting🤓