Optimal Stop-Loss and Take-Profit Parameterization for Autonomous Trading Agent Swarm

A new arXiv paper by Nathan Li, Aikins Laryea, and Yigit Ihlamur (arXiv:2604.27150) examines how better exit policies can improve autonomous crypto trading agent swarms. Using over 900 historical trades, they replayed each trade under multiple alternative stop-loss and take-profit configurations, comparing results against the existing production setup. The study found that exit design matters meaningfully: stronger configurations favored tighter loss limits, earlier profit capture, and closer trailing protection, leading to better risk-adjusted returns.

The authors also highlight a key evaluation challenge. A purely chronological train-test split initially placed the newest trades in an unusual war-driven market period, sharply distorting results. To mitigate this, the main comparison was run on randomized data, with the drawbacks acknowledged explicitly. Overall, the paper presents a practical, disciplined framework for tuning exit logic in autonomous trading systems, emphasizing transparency and systematic testing over fixed rules.