Part I · Chapters 1 to 2
Why We Look: Origins of Prediction
- 014 demonstrations
The Market Discovers the Future
A price is a forecast built from many people's information, and reading it well means knowing what it keeps and what it hides.
- 024 demonstrations
The Mathematics of Belief
Belief is a quantity that evidence should move, by the right amount, and a forecaster should be scored on it.
Part II · Chapters 3 to 7
The Engineers: Building the Methods
- 034 demonstrations
The Weather That Could Be Computed
Even exact equations forecast only as far as their starting point allows, and every series has structure to understand before forecasting it.
- 044 demonstrations
The Smoother
One number, alpha, decides how fast a forecast forgets, and the family built on it survives because it is simple and hard to beat.
- 054 demonstrations
The Filter
How much should new information change your mind? In proportion to how uncertain your prediction was and how noisy the measurement is.
- 064 demonstrations
The Unexpected Route
Box and Jenkins turned forecasting into a cycle: identify, estimate, check the residuals, and look again.
- 074 demonstrations
The Casino at Los Alamos
When you cannot calculate the answer, generate the future many times and count what comes out.
Part III · Chapters 8 to 11
The Crowds and the Experts: When Judgment Works
- 084 demonstrations
The Delphi Room
Delphi removes the room's social pressure, but convergence is not the same as being right.
- 094 demonstrations
The Oracle Problem
Expertise and calibration are different properties, and only scoring forecasts against outcomes can tell them apart.
- 104 demonstrations
The Superforecaster
Good forecasting is a set of habits that can be written down, checked and scored.
- 114 demonstrations
The Crowd and the Ox
A crowd's aggregate can beat its members, but only when their errors are independent and the aggregation is done with care.
Part IV · Chapter 12
The Competitions: How Methods Get Tested
- 124 demonstrations
The Competition
A forecasting method earns its place only by beating simple benchmarks on the same future observations.
Part V · Chapters 13 to 16
The Machine Learning Era: Deep Learning and Foundation Models
- 134 demonstrations
The Walmart War Room
A global gradient-boosted model wins when the features carry real, timely information, and only an honest backtest shows whether they do.
- 144 demonstrations
The Globalizer
One model trained across many series can forecast a series with almost no history, and its forecast should be a distribution, checked like one.
- 154 demonstrations
The Foundation
A pretrained model forecasts series it has never seen; whether it helps is settled the old way, against a simple baseline on data it could not have seen.
- 164 demonstrations
The Prophet
A forecast built from named pieces lets the analyst add what she knows, and obliges her to test it.
Part VI · Chapters 17 to 18
Uncertainty as a Feature: Probabilistic Forecasting
- 174 demonstrations
Living with Probability
A probabilistic forecast is judged on honesty and usefulness: calibrated coverage first, then as much sharpness as the evidence allows.
- 184 demonstrations
The Hierarchy
A total and its parts cannot both be right if they do not add up, and the fix should use every level's information.
Part VII · Chapters 19 to 24
Applied Forecasting: Methods in the Wild
- 194 demonstrations
Frank Bass and the Television
New products spread through two forces, outside influence and imitation, toward a ceiling the early data cannot reveal.
- 204 demonstrations
The Marketing Mix
A marketing mix model can forecast sales well and still be unsure who earned them; experiments supply the anchor.
- 214 demonstrations
The Bullwhip
A supply chain can manufacture its own volatility, and the cure starts with forecasting the right signal for the right decision.
- 224 demonstrations
The Causal Forecaster
A causal effect is the gap between what happened and a counterfactual that never happened, and every estimate is only as good as that counterfactual.
- 234 demonstrations
The Epidemiologist's Dilemma
Knowing what an epidemic is doing now and forecasting what it will do next are different problems, and each needs its own honest check.
- 244 demonstrations
The Anomaly
A model built on the old process cannot see a new one; the work is noticing the break and deciding how fast to forget.
Part VIII · Chapters 25 to 26
How to Live with Forecasts
- 254 demonstrations
Forecasting Your Own Life
Start from what happened to comparable efforts, keep the class honest, and choose the risk level of a promise on purpose.
- 264 demonstrations
How to Be Ready Without Being Certain
A forecast earns its keep through the decision it informs and the score it keeps.
Part IX · Chapter 27
Forecasting with Machines: Direction in the Age of AI
- 274 demonstrations
Giving the Oracle Direction
Model what the data can support, estimate what it cannot, and let a second route test the number.
Notebooks and skills on GitHub
Every chapter has a Jupyter notebook that reproduces its calculations and figures, and a skill an AI assistant can use to apply the chapter's method to your own data.