Research brief 005

Same model. A different target. Can the label change the signal?

Before rebuilding a return model, change what it is asked to predict. We tested that idea in earnings research.

The idea worth investigating

Cakici and Zaremba examine how transforming return targets changes machine-learning portfolio results. Ranking stocks may ask a different—and potentially more useful—question than predicting their exact returns.

The full analysis examines the study and the practical question behind it: Before rebuilding a return model, change what it is asked to predict. We tested that idea in earnings research.

Includes a house adaptation; author findings and our results are assessed separately.

Paid analysis on Substack

See what changes the investment case.

The headline is the starting point. The subscriber analysis takes you through:

  • The paper’s case for transformed targets
  • Our controlled earnings-model comparison
  • The results across broad and liquid universes
Read the paid analysis ↗

AI assists curation and drafting. The research status distinguishes paper reviews from house tests. Read our research approach.

New research briefs by email, free

Find your next research question.

Get the finding that caught our attention, the original paper, and the practical question it raises. Full reviews and house-test details are available with a paid subscription.