Reproduce “Adaptive Multi-Strategy Market-Making Agent For Volatile Markets” 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

Ali Raheman, Anton Kolonin, Alexey Glushchenko, Arseniy Fokin, Ikram Ansari

An Adaptive Multi-Strategy Agent (AMSA) is developed to conduct market making in volatile cryptocurrency markets by orchestrating diverse subordinate strategies. The agent continuously evaluates and selects the best-performing strategies based on past returns, showing consistent positive alpha, particularly with Hummingbot strategies, across various bull and bear market conditions.

Key insights

  • A meta-strategy approach that adaptively switches between diverse underlying market-making strategies can deliver robust positive alpha in highly volatile and unpredictable cryptocurrency markets, outperforming single-strategy methods.
  • The concept of "Narrow AGI," incorporating principles of purposeful activity and experiential learning, provides a viable framework for developing intelligent agents that adapt to complex, non-stationary financial environments without relying on explicit price prediction.
  • Strategic diversity is crucial, with specific agent families (like Hummingbot) proving more versatile across various market conditions (bull highly volatile, bear), while others (Base, NIOX) excelled mainly in bear markets when managed by an adaptive orchestrator.

Summary by alphaXiv.

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