Ben Recht's forecasting lecture: the math of turning past frequencies into future odds
beenwrekt · x · 2026-09-03
Berkeley researcher Ben Recht published a live blog of Lecture 3 of his graduate seminar "Forecasting: A Critical Retrospective," on the mathematics of transmuting frequencies of the past into odds of the future.
The hook: since the Great Depression, US law has required mutual funds to carry the disclaimer "past performance is not indicative of future results" — yet nobody actually believes it. The core question is how to judge forecasters by track record when the past can't perfectly predict the future, and how the choice of evaluation metric leads to particular forecasting algorithms.
He opens with Edmond Halley's two examples: celestial motion (Halley's comet, pairing three observations across 76-year gaps using Newton's orbital insights) where the past strongly predicts the future, versus pricing individual insurance policies, where it barely does. A substantive read for anyone interested in forecasting theory, statistical learning, and financial modeling.
More from Research
- LLMs are bad at Puzzlescript games, PuzzleJax team tells IEEE CoG — Amidos2006 · 2026-09-03
- ICM panel with Akshay Venkatesh and top mathematicians on AI and math now out — AlexKontorovich · 2026-09-03
- Philosopher says Fable 5.1 found Arrhenius's favored population ethics impossibility theorem is false — willmacaskill · 2026-09-03
- AI Is Forcing Mathematicians to Ask What They Really Value About Math — stevenstrogatz · 2026-09-03
- Caltech and NVIDIA Show Physics-Aware AI for Extreme Weather Prediction — AnimaAnandkumar · 2026-09-03
- Agents transmit invented languages to newcomers, enabling cumulative cultural evolution — EliasEskin · 2026-09-03