Researchers have developed a new online learning algorithm for market making that achieves a high-probability regret bound of $\widetilde{\mathcal{O}}(\sqrt{T})$. This improves upon a previous guarantee of $\widetilde{\mathcal{O}}(T^{2/3})$ for a feedback model motivated by limit order books. The new algorithm utilizes a discretization of the bid-ask space combined with the Hedge algorithm, and it also addresses environments where both market prices and trader valuations are adversarial, demonstrating that sublinear regret is impossible in such fully adversarial settings. AI
IMPACT This research advances theoretical understanding in online learning for financial applications, potentially influencing algorithmic trading strategies.
RANK_REASON This is a research paper detailing a new algorithm and theoretical results in online learning for market making. [lever_c_demoted from research: ic=1 ai=0.4]
- 2026
- alphaXiv
- arXiv
- CatalyzeX Code Finder for Papers
- Connected Papers
- CORE Recommender
- DagsHub
- Gotit.pub
- Hedge
- Hugging Face
- Influence Flower
- Litmaps
- Maran et al.
- ScienceCast
- scite Smart Citations
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
