In this episode of Forecasting Impact, hosts George Boretos and Mariana Menchero speak with Nicolas Vandeput, supply chain data scientist, author, educator, founder of SupChains, and organizer of the VN competitions. Drawing on years of experience in demand forecasting and inventory optimization, Nicolas challenges several widely held assumptions and shares a practical perspective on what truly creates value in forecasting.

We discuss the role of machine learning in demand forecasting, the importance of strong benchmarks, common pitfalls in forecast evaluation, and how inventory optimization changes the way we think about forecasting performance. Nicolas also shares lessons from working with practitioners across industries and reflects on the key insights from the VN1 Forecasting and VN2 Inventory Planning competitions.

Most forecasting goals focus solely on accuracy, but Harrison Katz reveals why understanding the true purpose of forecasts can transform how organizations make decisions. From leading Airbnb’s demand predictions to academic research, he uncovers the nuanced role forecasts play beyond numbers—and why misaligned goals and governance can distort the entire process.

In this episode, Harrison shares his nonlinear journey from dropping out of high school at 15 to leading Airbnb’s forecasting strategies across hundreds of global markets. He discusses how COVID-19 exposed the limitations of traditional models and taught invaluable lessons about system agility, regime shifts, and the importance of separating models for flow versus lead times, and not for supply versus demand. Harrison breaks down the often-overlooked concept of forecast governance, emphasizing that forecasts are not just numbers—they are political, strategic, and even behavioral tools that influence outcomes.

In this episode of Forecasting Impact, hosts Mariana Menchero and Laila Akhlaghi speak with Alex McCullough about his research on prediction markets, probabilistic forecasting, and the crypto ecosystem. He is best known for his work analyzing the accuracy of Polymarket, where he has built widely referenced dashboards to evaluate market calibration and performance. Through his analysis, Alex helps bring rigor and transparency to the discussion around how prediction markets perform as forecasting tools. Alex's work includes the following dashboards:

Initial dashboard: https://dune.com/alexmccullough/how-accurate-is-polymarket

And the widely referenced Brier Score dashboard: https://dune.com/alexmccullough/polymarket-brier-score