🤖 AI Summary
This study addresses the lack of reproducibility, systematic validation, and efficient resource allocation in traditional alpha factor mining. It proposes an autonomous alpha discovery system that performs end-to-end search within a space of frozen research artifacts—comprising hypotheses, expressions, evidence, logical reasoning, and review status—to generate and screen high-confidence alphas. Key innovations include an evidence-integrated search mechanism, an adversarial review process with re-execution and veto capabilities, pending-aware parallel budget allocation, and a large language model–driven prompt economy architecture with full-chain provenance. Deployed on the WorldQuant BRAIN platform, the system enabled multiple users and models to produce SPECTACULAR-grade alphas, achieving a peak fitness of 9.50 and Sharpe ratio of 3.48, with complete traceability throughout the discovery pipeline.
📝 Abstract
Language models can propose many plausible trading factors, but an autonomous research system must also allocate its evaluation budget, verify its own evidence, and preserve how each candidate was produced. We present AgonAlpha, an architecture that searches over frozen research artifacts---hypotheses, executable expressions, platform evidence, rationales, and review status---rather than formulas alone. To our knowledge, AgonAlpha is the first alpha-mining system to combine verified artifact search, a fresh-context adversarial reviewer with re-execution and veto authority, and pending-aware parallel budget allocation, together with a complete public evidence trail. Independent deployments on WorldQuant BRAIN produced SPECTACULAR-grade alphas across five users and six model backends, with Fitness reaching 9.50 and Sharpe reaching 3.48, while retaining prompt-to-expression provenance for every submission.