Reproduce “AlphaLedger: Portfolio-Grounded On-Policy Distillation for Regime-Aware Alpha Mining” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xiang Li, Zikai Wei, Yiyan Qi, Wanyun Zhou, Xiang Liu, Penglei Sun, Zhenheng Tang, Yongqi Zhang and 2 more
AlphaLedger trains a local generator to build stock-ranking formulas using completed programs’ after-cost portfolio evidence. EC-OPD turns that evidence into teacher preferences over legal next actions, while EviPool selects eligible stocks and RRG adjusts factor weights from lagged evidence; the library stays fixed during testing. In Chinese A-share backtests, EviPool returned 55.4% annually, with a 2.06 Sharpe ratio and 10.8% maximum drawdown; these results do not establish future returns.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.