Agentic AI in Automated Markets

By Irene Aldridge, Lourenco Belleau-Bhowmik, James Chan, Josh Chang, Kelly Deng, Pranav Dhingra, Xinda Ding, Haodong Fang, Satya Hanumanchi, Carter He, Shuang He, Yadong Hou, Vallab Karanam, Rithvik Kasarla, Tishya Khanna, Jihyun Kim, Sarah Konety, Abhinav Kotta, Anderson Lo, Jake Luther, Omar Mian, Ishan Patwardhan, Matteo Perona, Chenyue Qi, Shashank Ramachandran, Shubham Rateria, Jake Swain, David Thai, Michael Twersky, Susan Wu, Dean Xu, Helenna Yin, Jiachun Zheng

Rating

1538
Battle Count: 51

Relevance

8/10
Highly relevant for understanding how different behavioral archetypes (rational, manipulative, herd-following) impact market microstructure, liquidity, and volatility. Provides a framework for using RL agents as detectors for specific trading behaviors, which is valuable for both strategy development and regulatory compliance.

Implementation Complexity

7/10
Requires implementing RL environments (DQN and REINFORCE), defining complex reward functions with multiple components (stress, turnover, momentum), and integrating SHAP for interpretability. The discrete exposure control and feature engineering add to the complexity.

Reproducibility

2/5
No public code repository URL is provided in the text. While the methodology is described, specific hyperparameters for the pilot and the exact data preprocessing steps for the hourly crypto data are not fully detailed enough for immediate reproduction without the authors' code. The paper mentions a pilot on hourly cryptocurrency data (AAVE, ETC, SUI) but does not link to a dataset or code.

Abstract

Automated markets increasingly host autonomous, learning trading agents and are powered by formulaic automated market makers (AMMs). We study how heterogeneous agent behavior shapes stress, risk-adjusted performance, liquidity, and volatility in these types of markets. Three behavioral archetypes are modeled as agentic reinforcement learning (RL) systems: a rational arbitrageur, a manipulator, and a retail herd-follower. Each archetype is defined by its reward function within a common, asset-agnostic exposure-control environment, and each is trained with a value-based Deep Q-Network (DQN) and with REINFORCE using a learned baseline. A pilot on hourly cryptocurrency data (AAVE, ETC, SUI) motivates the hypotheses. Herd-followers produce the most frequent but shortest stress episodes, manipulators the fewest but most persistent, and rational agents the most stable rolling Sharpe ratios despite higher return volatility. SHAP attributions tie these patterns to drawdown and cross-asset signals, short-term volatility and position state, and momentum, respectively. The results motivate behavior-specific regulation rather than uniform limits on volatility or turnover.

Links

arXiv:2610.22522 ยท Paper rankings

The interactive Everscope explorer (charts, battles, favorites) loads below.