Explore how a factor family moves from an initial idea to evaluation and continued refinement. This architecture demonstration follows the public project design; it does not display measured factor performance.
01 / SEED
Start with a factor family.
A seed family defines the economic idea, its initial factors, and the boundary of the search. The family retains its search history so each round can build on what is already known.
Research artifact
Seed definition + family state
02 / EXPLORE
Generate candidates with context.
Broad search opens the space of possible expressions. Anchor selection identifies promising parents for focused refinement, while multiple branches preserve alternatives beyond the current highest-scoring candidate.
Research artifact
Candidate expressions + parent branches
03 / EVALUATE
Let evidence guide the next step.
Evaluation feedback informs candidate selection and continuation. The chosen objective changes the emphasis: standalone alpha, practical deployability, or complementarity with existing factors.
Research artifact
Evaluation feedback + family report
04 / REFINE
Decide how the family should evolve.
The continuation policy can deepen focused search, reopen broad exploration, switch objectives, confirm results, or stop. Selected candidates can be promoted into the formal factor library, with reports preserving the research trail.
Research artifact
Continuation decision + selected factors