Great Forecasters Aren't Magic: Bare Probabilities Need Model-Based Derivation

Sam_kuyp · x · 2026-09-11

Sam Kuyp responds to David Manheim's take on judgmental forecasting. He concedes some people may be unusually good at prediction, but not through magic—their probabilities must come from models connecting assumptions to outcomes, even informal ones. Without such a model-dependent derivation, there is no reason to accept a bare probability. Manheim counters that empirical evidence shows judgmental forecasts can be predictive even when unjustifiable by concrete models.

Original post →

More from AGI Musings

AGI Musings channel →