Reproduce “Towards Robust Reinforcement Learning Market Making: Lessons from Loss Function Experimentation” 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

Siddharth Nand, Bora Guney

The authors compare reinforcement learning market-making agents trained with and without penalties for low cash and excessive inventory. In a single-stock simulation spanning 400 episodes, the penalty-based agent generally earns higher rewards and posts narrower spreads. However, the experiment does not measure inventory risk or establish live-market performance, and shared sequential training complicates the comparison.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.