FinCon’s Manager–Analyst Agents and Episodic Trading Reflection
Summary
The document introduces FinCon, a proposed system that models an investment team with specialist analyst agents reporting to a manager. Analysts process different kinds of market information; the manager combines their findings, makes trading decisions, and oversees risk. Separate controls assess risk within each trading episode and compare outcomes across episodes. The system also uses working, procedural, and episodic memory to support ongoing decisions and learning.
Conceptual Verbal Reinforcement compares analysts’ insights with the manager’s decisions across episodes, then uses the comparison to recommend changes to the system’s investment beliefs and analyst focus. The article describes this framework and begins an MQL5 implementation inspired by its memory and learning concepts, but explicitly does not use the original framework’s pretrained language models. It provides no performance evidence in the excerpt: evaluation on historical market data is planned for a later installment. Its claims of improved performance therefore remain unsubstantiated here, and the MQL5 implementation is incomplete.
Key ideas
- FinCon assigns data analysis to specialist agents and reserves final trading decisions for a manager agent.
- Risk controls operate both during an episode and across completed episodes.
- Conceptual Verbal Reinforcement uses comparisons between insights and decisions to guide changes in investment beliefs.
- Working, procedural, and episodic memory serve different roles in retaining context, strategies, and past outcomes.
- The article begins an MQL5 interpretation without pretrained language models, and does not yet report historical-market evaluation.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.