Quant Radio: Seemingly Virtuous Complexity in Return Prediction

Can machine learning really predict stock market returns with just 12 months of data? This episode explores a bold claim made by a prominent academic paper using Random Fourier Features (RFF) to forecast market movements with stunning accuracy — and the fascinating rebuttal that followed.

Join us as we break down:

• The mechanics behind the KMZ RFF strategy

• Why its seemingly impressive performance might just be mathematical coincidence

• How it unintentionally mimics a simple momentum strategy with built-in volatility timing

• What this means for the limits of learning in finance…