Stock ranking vs. market timing: what really works

There are two families of bets: calling the market's direction (timing) or, within the market, ranking which names to overweight (cross-sectional). They don't perform the same.

In our work, almost all market timing — via macro, news, alternative data — came out null. Meanwhile, ranking stocks at a weeks-to-months horizon did show a robust edge.

It makes sense: predicting the aggregate means competing against everyone at once; ranking names exploits more persistent relative differences. That's why our focus is ranking, not the crystal ball.