Signal Research & Discovery

This pillar covers the front end of the alpha pipeline: the work of turning raw and alternative data into candidate signals, and deciding which of them are real. The emphasis is on the discovery process itself — research methodology, feature construction, validation, and the lifecycle of a signal once it's found — rather than on portfolio assembly or execution, which live in their own pillars.

Start with the research methodology and validation pieces, which set the discipline for everything else: how to structure a search across many candidates, and how to separate a genuine weak signal from noise that happens to fit the sample. From there, the feature engineering and alternative-data articles widen the input space, while the market microstructure piece narrows it to alpha that lives in order flow and price formation. The signal decay article closes the loop, treating any discovered edge as a wasting asset with a finite shelf life.

Read together, these articles trace one path: where signals come from, how you construct them, how you test them honestly, where they hide in the data, and why the work is never finished. The micro alphas primer is the lightest entry point if you want the framing before the methods.