On 28 November 2018, Émile Servan-Schreiber reshared Scott Page’s announcement of a new book promoting many-model thinking. On 6 April 2020, he recommended a roughly ten-minute instalment of Page’s Understanding Epidemics series. Together, the shares point to a methodological interest: using several ways of representing a problem to understand what any one model leaves out.
The book behind the method
Basic Books dates The Model Thinker: What You Need to Know to Make Data Work for You to 27 November 2018. It presents a toolkit ranging from linear regression to random walks, and argues for using multiple models to understand complex systems. The claim concerns reasoning with different representations, rather than a promise that every additional model improves a forecast.
What the epidemic series made concrete
The image attached to the April share is labelled SIR Model. The University of Michigan’s official account of Page’s series explains that SIR refers to susceptible, infected and recovered populations. It identifies a video devoted to that model, including exponential growth and the reproduction-number threshold affecting whether a disease spreads.
The university’s series guide also lists fatality-rate models, network models and curve-fitting approaches. Each brings a different part of the problem into view: counts and timing of deaths, relationships through which transmission occurs, or the shape of observed trends. The guide describes the approaches as complementary. A model’s usefulness depends on the question and the assumptions it makes.
This distinction is valuable for reading forecasts. Explaining why an epidemic can grow is different from estimating how many hospital beds will be needed on a particular date. A model may clarify the mechanism without providing a validated numerical prediction for the second question.
A connection to practical forecasting education
The same university page records a pandemic prediction-market learning experiment built by Page and Servan-Schreiber. It describes eight pandemic-related questions and an exercise designed for individuals, instructors and students. That establishes a practical connection between their educational work. It does not supply a scored result for the market.
Hypermind’s separate infectious-disease forecasting work addresses the empirical question of how a crowd performs. Page’s video recommendation addresses how to understand models. Keeping those purposes distinct prevents an educational explanation from being mistaken for an outcome forecast.
