Net Assessment Prediction is Hard, Especially about Confidence Levels
26 snips
Oct 2, 2025 Explore the overconfidence problem plaguing national security experts. Recent research highlights how biases lead to inflated predictions and false positives. The hosts debate the implications for foreign policy and ways to cultivate humility and accountability in decision-making. They critique current military gatherings and economic policies while applauding strategic moves like increased missile production. Insightful discussions on fostering better prediction practices leave listeners contemplating the complexities of national security assessments.
AI Snips
Chapters
Books
Transcript
Episode notes
Experts Are Systematically Overconfident
- National security experts are often overconfident about predictions, producing many false positives.
- Friedman found 90% confidence corresponded to only 57% accuracy, showing systemic miscalibration.
False Positives Inflate Threats
- Overconfidence produces threat inflation through false positives and alters policy choices.
- Different actors value type I and type II errors differently, creating policy tensions.
Lunch Chat About Type I And II Errors
- Christopher Preble shared a lunch story highlighting debates over type I and II errors.
- He explained type I as a false positive and type II as a false negative to illustrate differing stakes.
