Sharpe Ratio: 3.0+ - Learning-to-Rank Momentum Strategy

A new paper, Learning to Rank: Enhancing Momentum Strategies Across Asset Classes, applies machine-learning ranking models to improve traditional momentum.

The author finds that replacing simple return-based sorting with learning-to-rank algorithms—which better identify relative winners and losers, significantly boosts performance across equities, bonds, commodities, and FX.

The strategy is simple:

• Use a learning-to-rank model to score assets.

• Go long the top-ranked decile, short the bottom decile.

The intuition: smarter ranking extracts more signal from the same data—turning a classic anomaly into a higher-Sharpe edge.

papers.ssrn.com/sol3/pa…