Reproduce “Plan Before You Trade: Inference-Time Optimization for RL Trading Agents” 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
Eun Go, Rohan Deb, Arindam Banerjee
FinPILOT adapts a pretrained portfolio policy when deciding, using forecasts from a separately trained forecaster. It samples price paths, updates the policy against them, then trades and repeats, assuming trades do not materially move prices. Across two markets and three years of tests, it improved returns in 28 of 30 settings and Sharpe in 27 of 30, but not universally.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.